GEORGIA DOT RESEARCH PROJECT 1224 FINAL REPORT MICRO-DYNAMICS OF BUSINESS LOCATION AND GROWTH AND ITS EFFECTS ON THE TRANSPORTATION NETWORK AND CONGESTION IN GEORGIA AND THE SOUTHEAST REGION OFFICE OF RESEARCH GDOT Research Project No. 12-24 Final Report MICRO-DYNAMICS OF BUSINESS LOCATION AND GROWTH AND ITS EFFECTS ON THE TRANSPORTATION NETWORK AND CONGESTION IN GEORGIA AND THE SOUTHEAST REGION By Dr. Vivek Ghosal Dr. Frank Southworth Georgia Institute of Technology Contract with Georgia Department of Transportation In cooperation with U.S. Department of Transportation Federal Highway Administration August 2014 The contents of this report reflect the views of the author(s) who is (are) responsible for the facts and the accuracy of the data presented herein. The contents do not necessarily reflect the official views or policies of the Georgia Department of Transportation or the Federal Highway Administration. This report does not constitute a standard, specification, or regulation. 1 1.Report No.: F FHWA-GA-14-1224 2. Government Accession No.: 3. Recipient's Catalog No.: 4. Title and Subtitle: Micro-dynamics of business location and growth and its effects on the transportation network and congestion in Georgia and the Southeast Region 5. Report Date: August 2014 6. Performing Organization Code: 7. Author(s): V Vivek Ghosal and Frank Southworth 8. Performing Organ. Report No.: 9. Performing Organization Name and Address: Georgia Institute of Technology 790, Atlantic Drive, Atlanta, GA 30332 12. Sponsoring Agency Name and Address: Georgia Department of Transportation Office of Research 15 Kennedy Drive Forest Park, GA 30297-2534 10. Work Unit No.: 11. Contract or Grant No.: 0010766 (RP 12-24; UTC Sub-project) under RP 11-24) 13. Type of Report and Period Covered: Final; June 2012 August 2014 14. Sponsoring Agency Code: 15. Supplementary Notes: Prepared in cooperation with the U.S. Department of Transportation, Federal Highway Administration. 16. Abstract: The project explored the linkages between industry dynamics and economic activity, and the macro-congestion aspects of freight transport. The Kia Motors manufacturing plant near West Point, Georgia was selected for case study. The principal study effort went into collecting data elements in sufficient detail to allow for in-depth empirical analysis. This included collecting economic activity and supply chain data associated with the growth in activities both within the plant and among the many parts suppliers that have moved into the area to serve it. Impact variables examined include those related to employment in a wide range of occupations, schooling, educational attainment, and population and migration patterns. A series of economic multipliers are reported, showing some significant differences between counties with component supplier activity and other counties in the region. The project also developed a detailed spatial mapping and explored the derivation of supply chain cost estimates associated with both domestic and international movements of auto industry inputs and outputs, involving individual and multimodal highway, rail, and waterway shipments. 17. Key Words: Plant Location, Economic Impacts, Economic Multipliers, Freight Transportation Flows and Costs, Automobile Manufacturing Supply Chain, Multimodal Analysis, Global Shipments Modeling 18. Distribution Statement: 19. Security Classification (of this report): U nclassified 20. Security Classification (of this page): Unclassified 21. Number of Pages: 92 22. Price: 2 Micro-Dynamics of Business Location and Growth and Its Effects on The Transportation Network and Congestion in Georgia and The Southeast Region Table of Contents List of Tables ..............................................................................................................................3 List of Figures .............................................................................................................................4 List of Abbreviations Used in This Report...................................................................................5 Executive Summary ....................................................................................................................6 1. Introduction.............................................................................................................................8 2. Framework for Analysis and Data ...........................................................................................8 2.1. Component Suppliers Data .................................................................................................9 2.2. Core versus Non-Core County Classification....................................................................11 2.3. Economic and Business Development Data ......................................................................12 3. Examining the Economic and Business Development Effects ................................................14 3.1. Changes in Key Economic Variables ................................................................................14 3.2. Calculating Multipliers .....................................................................................................19 4. Supply Chains: Components and Final Product Flows ...........................................................34 4.1 Introduction.......................................................................................................................34 4.2 Measuring the Effects of Congestion-Induced Delays on Firm Transaction Costs ..............36 4.3 Product Flows in the Automotive Industry Supply Chain...................................................44 4.4 Estimation of Modal and Intermodal Transportation Costs.................................................55 4.5 Global Supply Chain Modeling: Putting Freight Costs on Intermodal Networks ................70 5. Study Summary and Conclusions ..........................................................................................76 Appendix A. Component Suppliers ...........................................................................................78 Appendix B. Economic and Business Effects ............................................................................85 List of Tables Table 1.1 Local Components (Auto Parts) Suppliers by County ................................................12 Table 2.1 Percent Change in Employment by Industry ..............................................................15 Table 2.2 Percentage Change in Migration ................................................................................17 Table 2.3 Percent Change in Education .....................................................................................20 Table 2.4 Percent Change in Schooling .....................................................................................21 Table 2.5 Percent Change in Household Income........................................................................22 3 Table 3.1 Multipliers of Employment by Industry .....................................................................27 Table 3.2 Multipliers of Migration ............................................................................................29 Table 3.3 Multipliers of Education ............................................................................................30 Table 3.4 Multipliers of Schooling ............................................................................................32 Table 3.5 Multipliers of Household Income...............................................................................33 Table 4.1 Automobile Parts (Component) Supplier Locations Data File ....................................48 Table 4.2 Geocoded Foreign Import Shipments Data File..........................................................50 Table 4.3 KMMG Production Statistics .....................................................................................53 Table 4.4 KMMG Sales Statistics..............................................................................................53 Table 4.5 Privately-Owned Railcars Cost Breakdown by Selected O-D Rail Distances..............64 Table 4.6 Network Link Attributes File .....................................................................................75 Table A.1 List of Kia Motors Manufacturing Georgia (KMMG) Suppliers in Georgia and Alabama ................................................................................................................................... 78 Table A.2 Employment and Investments of KMMG Suppliers in Georgia .................................81 Table A.3 Employment and Investments of KMMG Suppliers in Alabama ...............................82 Table B.1 Actual Change in Employment by Industry ...............................................................85 Table B.2 Actual Change in Migration ......................................................................................87 Table B.3 Actual Change in Education ......................................................................................88 Table B.4 Actual Change in Schooling ......................................................................................90 Table B.5 Actual Change in Household Income ....................................................................... 91 List of Figures Figure 4.1 Regional Highway Congestion Forecast for 2040 .................................................... 38 Figure 4.2 National Highway Congestion Forecast for 2040......................................................39 Figure 4.3 Major Rail Lines and Current Congestion Levels in Georgia ....................................40 Figure 4.4 Business Costs of Traffic Congestion .......................................................................45 Figure 4.5 Principal Components of Automobile Manufacturing Supply Chains........................46 Figure 4.6 Transportation Links in an Automotive Industry Supply Chain (Generic) ................47 Figure 4.7 Map of Georgia and Alabama Automotive Parts Supplier Locations.........................48 Figure 4.8 Rate of Growth in Imported Auto Parts from 2008-2012 by Selected US Ports of Unlading ...................................................................................................................................51 Figure 4.9 Example Trans-Pacific Parts Shipment Routes .........................................................52 Figure 4.10 Elements of Truck-Trip Based Freight Transportation and Logistics Costs ............58 Figure 4.11 Example Initial Input Screen for Rail Costing Program ..........................................63 Figure 4.12 Results from Rail Costing Scenario Using Railroad-Owned Railcars ......................64 Figure 4.13 Results from Rail Costing Scenario Using Privately-Owned Railcars .....................65 Figure 4.14 Effect of Number of Railcars Per Train on Shipment Rate per Automobile .............65 4 Figure 4.15 Freight Data and Modeling Components of Automobile Manufacturing Supply Chains: Flows and Costs Modeling ...........................................................................................71 Figure 4.16 Prototype Supply Chain Routing Model Interface ...................................................72 Figure 4.17 The ORNL Multi-Modal/ Inter-Modal Freight Network Data Model ......................74 Figure 4.18 Simplified Foreign Seaport Link-Node Representation ...........................................76 List of Abbreviations Used in This Report 3PL Third Party Logistics operator/agency ACS American Community Survey ATRI American Transportation Research Institute BNSF Burlington Northern Santa Fe Railroad CSCMP Council of Supply Chain Management Professionals CSX, CSXT CSXT Railroad FAF3 Freight Analysis Framework, Version 3 FHWA Federal Highway Administration GC Georgia Central Railroad GDOT Georgia Department of Transportation GDP Gross Domestic Product HOG Heart of Georgia Railroad JIT Just-In-Time (delivery service) KMMG Kia Motors Manufacturing MSA Metropolitan Statistical Area NODUS, SMILE, STAN Freight network models NS Norfolk Southern OEM Original Equipment Manufacturer OOIDA Owner-Operator Independent Drivers Association ORNL Oak Ridge National Laboratory STB Surface Transportation Board TEU Twenty-foot equivalent unit (shipping container) URCS Uniform Rail Costing System US DOT United States Department of Transportation VHSS The Hamburg Shipbrokers' Association 5 Executive Summary The research documented in this report set out to explore the little understood linkages between the micro-foundations of industry dynamics and economic activity, and the macrocongestion aspects of freight transportation. A major barrier to such understanding has been the difficulty of obtaining the necessary data for analysis purposes. Recognizing this, the principal study effort went into collecting and merging the necessary data elements, in sufficient detail to allow for in-depth empirical analysis. For our case study, we used the location and rapid expansion of the Kia Motors Manufacturing (KMMG) plant in West Point, Georgia, as a natural experiment, to study the resulting economic and transportation effects of placing a large industrial plant in a little developed, semi-rural location. The KMMG plant started operations in 2008, with an initial production capacity of 250,000 vehicles per year. After subsequent expansions, the current production capacity of the plant stands at some 360,000 vehicles per year. Large automobile assembly plants like KMMG are known to generate significant economic multiplier effects on neighboring areas. The plant has also had significant effects on transportation flows, which arise from both the inward movement of automobile components, and the outward movement of finished automobiles. First, to get a clear understanding of the industrial processes involved, we developed a taxonomy of the automobile supply chain, identifying its major component categories (Section 2). We identified the locations of the many component suppliers that have located in Georgia and in neighboring Alabama Counties, following the decision of Kia Motors to locate in West Point. For these component suppliers, we obtained information on the types of components they manufacture and supply to KMMG, as well as data on their employment and investment levels. The location of these numerous component suppliers in areas close to the KMMG plant is found to provide a substantial boost to the overall economic activity in the region. Using the American Community Survey (ACS) database, we studied the economic impacts on counties surrounding the plant, by dividing them into core and non-core counties (Section 3). Core counties are defined as those where a meaningful number of component suppliers are located, whereas a non-core designation refers to neighboring counties with a lack of meaningful component suppliers. We examined a comprehensive set of variables, including those related to employment in a wide range of occupations, schooling, educational attainment, and population 6 and migration patterns. In our examination of these data and computation of multipliers, we found that in some categories of economic and business development the core counties show substantial differences compared to non-core counties, while in other areas differences are less clear. (The report includes numerous tables containing these multiplier effects, in Section 3 and in two Appendices). To understand the inflow of components to the KMMG plant, and the outflow of finished automobiles, we identified the types of, and origination points for, those components obtained from suppliers outside the southeast, including outside the United States. We obtained detailed data on many of the individual shipments associated with the automobile manufacturing supply chain, and on its uses of local, regional, and national highway, rail and waterway (including seaport) networks and cargo transfer facilities (Section 4). This includes data on the freight flows associated with international, multimodal land-sea shipments, originating in both Asia and Europe. Data sources and software tools for costing the transport of both components and finished vehicles were then identified, and are described. In particular, we explored the nature of just-in-time transport costs and the value of on-time reliability, and assessed the availability of existing data sources for doing so. An interview and tour of the Kia facility indicated the considerable importance of reliable, on-time components delivery to its production process. To allow the routing and mapping of product shipments, and to better model the door-todoor costs involved in individual freight movements, we enhanced an existing global truck-railtrans-oceanic freight network database. A description of this model-supporting database is followed with a brief discussion of the potential for significant freight movement bottlenecks within the southeast region, based on an interpretation of recent highway and rail traffic forecasts for the next three decades. Brief reference is also made to the potential impacts of a capacityexpanded Panama Canal on the competitive advantage of different intercontinental land-sea routes, and the markets currently serving the KMMG plant and region. This is seen as a topic worth further exploration, using the sort of multimodal network-based analysis tool and data described in this report. While more detailed analytical and econometric modeling is needed to better understand the exact magnitudes of the effects of the KMMG plant's decision to locate in central Georgia, the database constructed during the project represents an excellent starting point for such an effort. The project also demonstrates the level of effort needed to construct similar datasets for other manufacturing plant-based studies. 7 1. Introduction The State of Georgia offered approximately $500 million in incentives to attract Kia Motors to locate its manufacturing assembly plant in Georgia. In March 2006, Kia President E.S. Chung and Georgia Governor Sonny Perdue signed the contracts for Kia to build its first North American automobile manufacturing facility on over 2,200 acres in West Point, Georgia. In November 2007, Kia Motors Manufacturing of Georgia (KMMG) announced that its first production vehicle would be the next generation Kia Sorento, and in February 2008, KMMG's first employment application process closed with an automotive industry benchmark-setting 43,013 applications received. By September 2010, KMMG had produced over 100,000 vehicles and in January 2012, they completed an expansion by increasing the plant's full production capacity to 360,000 vehicles annually. In July 2013, KMMG reached a landmark of 1,000,000th vehicle produced.1 The scale of operations at KMMG is expansive, with significant growth from its inception to the current stage. The objective of our study is to examine the impact of the KMMG plant's location on the region's economic and business development, and the resulting transportation flows and logistics. In sections 2 and 3, we lay out the framework for our study, describe the data we compiled, and present a broad overview of the economic and business development effects that can be attributed to the KMMG plant. In section 4, we examine the effects and implications for a broad range of transportation issues. These include demand for transportation, logistics, and traffic congestion effects on business performance, among others. 2. Framework for Analysis and Data The town of West Point is located in Troup County, Georgia. In the 2000 Census, West Point had a population of about 3,400. There was no major industry located in West Point before Kia Motors started operations. Troup County as a whole had a population of about 58,800 in the 2000 Census. Troup County has no recent history of major manufacturing. In much earlier periods in the early-to-mid 1900s, Troup County had some presence in the textile industry along with agriculture. But these industries faded as the US lost much of its competitive advantage in textiles. Based on our examination of Troup County's history and the 2000 US Census and 1 KMMG history and operations information are available at: http://www.kmmgusa.com/about-kmmg/our-history/ 8 American Community Survey data before Kia located, Troup County is best described as semirural. In an important respect, this allows for a relatively clean experiment for us to study. The KMMG advanced manufacturing plant is located in an area with no significant manufacturing, or other industry, present. This potentially allows us to cleanly track the economic and business development in Troup County and related areas in Georgia and Alabama before and after KMMG. To compile data to examine these effects, we considered several sources. For the economic and business development effects, we first considered the US Census. Given that the KMMG plant investment and activities by component makers started in 2008, the relevant Censuses we could use would be 2000 and 2010. This choice posed several problems. The main one being that the year 2000 was considerably before 2008 and one could argue that much could have changed in the intervening 7-8 years before the KMMG plant startup. Given this, we explored using the American Community Survey (ACS) data, which provides Countylevel data at a higher frequency for many Counties. Our examination of the ACS data showed that we could use the 2005-2007 ACS for the pre-KMMG data, and the 2009-2011 for the postKMMG data. Being able to use the 2005-2007 and 2009-2011 ACS data solved a very important empirical identification problem as there was no overlapping period. The pre-KMMG and postKMMG data were cleanly separated, allowing us to examine the effects of the KMMG location on a range of economic and business variables. Overall, the combination of factors related to the location of the KMMG plant in Troup County and having access to economic data that can be separated as pre-and-post, allows us to study a clean natural experiment. Below, we describe these and other data in detail. 2.1. Component Suppliers Data The ACS data are at the County level. After the KMMG plant location decision, many automotive component suppliers decided to locate in Troup County and other nearby areas. We compiled the list of component suppliers who located pursuant to the KMMG location decision. Since Kia doesn't disclose their supplier list and no existing studies have provided a comprehensive list of Kia suppliers, we collected the component suppliers' information ourselves. We focused on collecting information of Kia's suppliers in Georgia and Alabama as 9 the intention of the research is to study the local economic impact of Kia Motors Manufacturing Georgia (KMMG) assembly plant in West Point, GA, which is more related to the state of Georgia and Alabama. For the Alabama part, the Alabama Department of Commerce published a list of Kia's Suppliers in Alabama in 2013. Based on the company names on the list, we turned to the Alabama New & Expanding Industry Report composed and published by Alabama Development Office for detailed information about year, jobs, investment, and industry code. Then, we went to the companies' websites for their nationality and address information. For the companies that didn't give factory addresses on their websites, we obtained information on them by searching Google Maps and www.manta.com. Regarding Georgia, because no State government departments appear to publish a specific list of Kia's suppliers in Georgia, we collected suppliers' names using different methods. First, in the report "2011 Automotive Manufacturing in Georgia" by Georgia Power Co., information on some of Kia's component suppliers in Georgia was listed. Then, we used key words searching (like "Kia supplier") to find news, reports and articles that mentioned Kia's suppliers on the Atlanta Journal-Constitution and Georgia Chamber of Commerce. Next, similar to the approach we took with Alabama, the companies' websites, Google Maps and www.manta.com were used to obtain the companies' nationality and address information. However, different from the Alabama list, which only includes existing suppliers, news articles that reported incoming future suppliers were also scanned and the information checked to verify that it was still up to date. As an example, an earlier article reported that DangNam Tech, a supplier of Kia, was about to come to Columbus, GA, investing $29 million and creating 350 jobs. However, later articles said that as the financial situation changed in 2009, the company would no longer come to Georgia. Table A.1 in Appendix A provides a list of 117 component suppliers of KMMG West Point assembly plant (25 in Georgia, 92 in Alabama) with company names, supplying components and location information. As noted above, only information of suppliers in Georgia and Alabama was collected. Next, in Tables A.2 and A.3, we provide information on employment and investments by the component suppliers located in Georgia and Alabama, respectively. The Georgia data are from the news and articles on the Atlanta Journal Constitution and Georgia Chamber of Commerce website, and the 2011 and the 2013 Troup County Directory of Manufacturers. Address information, if not provided by the previous sources, are from company 10 websites, Google Maps, or www.manta.com. For Georgia, with the exception of Yasufuku USA Inc., all the suppliers are newly established companies in Georgia. These companies came directly to supply the KMMG plant. Our meeting with the Kia management in late July of 2013 confirmed that the vast majority of the Kia suppliers are dedicated to the KMMG plant production. From 2011 to 2013, we can see most of the Kia suppliers have expanded, which is in accord with the expansion of KMMG production. The KMMG impact is still growing as we see many suppliers have exceeded the announced future employment, and there were still new suppliers coming in 2012. The Alabama supplier names and components are from the 2013 Kia supplier list composed and provided by Alabama Department of Commerce and based on the Alabama Industrial Database. Address information, if not provided by the previous list, is again from company websites, Google Maps, or www.manta.com. As we observe from the data, most of the Alabama investments and employments are from expansion of current facilities, which is not the case of Georgia. This is because before KMMG came to Georgia, Hyundai (which took a controlling interest in Kia in 1998) had already built their assembly plant in Montgomery, AL, which has brought a number of suppliers to the state. Then, since Kia and Hyundai share many components, many Hyundai suppliers naturally also become suppliers of Kia, and have expanded because of the newly generated demand from the KMMG plant. So, the KMMG plant has also benefited these companies in Alabama. 2.2. Core versus Non-Core County Classification The objective of the project is to be able to examine the changes in economic and business development across Counties that are the central beneficiaries of the KMMG plant's location, versus those that are not. We use the location of component suppliers to designate Counties as Core versus Non-Core. We assume that the Core counties are those with three or more component suppliers of KMMG. These counties are the ones that are assumed to have received the most economic impact from the KMMG plant. Table 1.1 provides information on these Counties and the number of components suppliers who locate in each. In our study, we compare the changes in the Core Counties with those in the Non-Core, as well as two broader averages related to State-wide and major metropolitan area. For example, we compare changes in the Core Troup County with Non-Core, as well as the Georgia average 11 and Atlanta Metropolitan Statistical Area (MSA) average changes. These comparisons provide a clear picture of the effects in the Core Counties. Table 1.1 Local Components (Auto Parts) Suppliers by County State County AL AL AL AL AL AL AL AL AL AL AL AL AL AL GA GA GA GA GA GA AL GA AL & GA Autauga Bullock Butler Chambers Crenshaw Elmore Lee Lowndes Macon Montgomery Pike Randolph Russell Tallapoosa Harris Heard Meriwether Talbot Troup Upson AL Total GA Total Total Number of Kia Suppliers 0 0 4 7 2 3 14 2 1 11 1 2 0 3 2 0 2 0 20 0 50 24 74 Core N N Y Y N Y Y N N Y N N N Y N N N N Y N Notes: (1) Based on the component supplier information of Table A.1, numbers of Kia suppliers in each of the 20 counties of interest are counted. Among the total 115 identified component suppliers, 74 of them locate in the 20 counties surrounding the Hyundai Alabama and Kia Georgia assembly plants.( 2) Core counties are the counties with three or more component suppliers of KMMG. These counties are the ones that are assumed to have received the most impact from the KMMG plant. See Figure 4.7 below for a mapping of these suppliers. 2.3. Economic and Business Development Data Next, we used the Core v. Non-Core classification to examine various economic and business development indicators. Our objective here is to examine economic and business development in an encompassing manner. 12 To measure the local business and economic impact of the KMMG plant, we use the American Community Survey (ACS) data. The ACS is an ongoing survey that provides data every year giving communities the current information they need to plan investments and services. The data are estimates based on the collected survey answers. For example, the ACS 3year data, for 2005-2007, are based on the answers from the 2005, 2006 and 2007 surveys. In our study, we use the ACS 2005-2007 data to represent the economic situation before Kia started operating in 2008. Similarly, the ACS 2009-2011 data were used to represent the economic situation after Kia started operations. The ACS provides more than 100 variables. In our study, we use specific variables that fit into broad categories such as: employment by occupations; schooling; educational attainment; population migration; and household income. For each variable, we examine the change going from ACS 2005-2007 which is centered in 2006, to ACS 2009-2011 which is centered in 2010. The actual data for our broad set of variables are presented in the tables in Appendix B. In the main body of the report, we discuss the percent changes in the variables, as well as a rough measure of the multipliers, which we discuss later. We collected ACS state-level data for Georgia and Alabama and county-level data for 20 counties surrounding the Kia Georgia and Hyundai Alabama plant. The ACS 3-year data from five counties (Bullock, Crenshaw, Lowndes, Heard, Talbot) are not available because their populations are too small; the ACS 3-year data only report counties of populations larger than 20,000. A few variables for Meriwether and Butler Counties were also not available. To compare the changes in the affected Counties with the Atlanta MSA, we collected the data for all counties located within the Atlanta MSA and aggregated the data to report the ACS variables for the entire MSA. Here we use a somewhat narrower definition of the Atlanta MSA, which is different from the official definition of Atlanta MSA consisting of 25 counties. For our study, the 25-county area is far too expansive and extends right to the border of the area neighboring the Kia plant. Instead of this overly expansive 25-county definition, we picked a slightly narrower 15-county the Atlanta MSA definition. In our view, these are the counties that very likely constitute the core of the Atlanta MSA and serve as the relevant benchmark to compare changes in the Kia activity related counties. The 15 counties we use as the Atlanta MSA definition are: Carroll, Cherokee, Clayton, Cobb, Coweta, DeKalb, Douglas, Fayette, Forsyth, Fulton, Gwinnett, Hall, Henry, Paulding, and Rockdale. This definition of Atlanta MSA is consistently used in our study and reported in the tables. 13 Overall, using the ACS data, we examine a wide range of indicators which fall under broadly defined categories such as employment by major occupational category, migration, education, schooling and household income. As noted earlier, we use the ACS 2005-2007 as the pre-KMMG data, and 2009-2011 as the post-KMMG data. 3. Examining the Economic and Business Development Effects The underlying raw data on the changes in the various ACS variables are presented in Appendix B. In this section, we present the data and discuss changes in two forms. First, in Tables 2.1 to 2.5, we examine the percentage changes in the ACS variables going from preKMMG to post-KMMG. Second, in Tables 3.1 to 3.5, we provide a perspective on some basic multiplier effects. Our attempts to conduct more sophisticated multiplier calculations are constrained by the lack of availability of time-series data both pre-and-post KMMG location. 3.1. Changes in Key Economic Variables Table 2.1 displays the data (as percentage changes) for employment in various major occupational categories made available by the ACS. Some of the key observations are as follows. First, in Georgia, the Core Troup County outperformed all the three counterparts (Georgia average, Atlanta MSA, and non-core counties) in retail trade, transportation and warehousing, and finance and insurance; it also outperformed the Georgia average, Atlanta MSA, but not the non-core counties in sales and office and wholesale trade; and it outperformed non-core counties in manufacturing. However, Troup County didn't show faster increases in the management, service, construction, and education and health care sectors. This shows that although the direct job creation effect of the KMMG plant is in the manufacturing sector, the increase of manufacturing jobs in this region is still weaker compared to other parts of Georgia. However, the induced job creation effect greatly benefited Troup County, mainly in sectors like retail trade, transportation and warehousing, and finance and insurance, which support the manufacturing factories and their workers. It is worth noticing that Harris County experienced significant growths in employment in management, wholesale trade, and education and health care. But these effects, from what we can infer, are due to other changes such as those in military establishments and are not related to KMMG. Second, the Alabama core counties had better 14 Table 2.1 Percentage Change in Employment by Industry State County C o Management r Service Sales and office Construction Manufacturing Whole sale trade Retail Transportation and Finance and trade warehousing insurance Education and health care e 06-10 06-10 06-10 06-10 06-10 06-10 06-10 06-10 06-10 06-10 AL AL 5.18 10.32 -1.58 -13.86 -11.68 -22.22 -2.16 -3.12 -5.15 8.49 AL Core Avg. 14.59 10.36 -5.26 -25.67 -18.17 -15.81 8.55 -2.00 -17.71 22.14 AL Non-core Avg. 12.78 10.11 -1.96 -18.24 15.59 -37.17 -15.45 -1.33 -19.09 18.32 AL Autauga N 14.76 10.74 0.48 -31.19 -14.74 -14.77 5.71 48.87 9.09 30.51 AL Bullock Y N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A AL Butler Y N/A N/A N/A -10.76 -46.48 26.83 14.65 12.06 -10.74 2.09 AL Chambers Y 9.57 -20.14 -11.92 -38.10 -36.25 -8.09 2.49 -25.89 3.59 8.64 AL Crenshaw N N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A AL Elmore Y 37.94 39.53 2.14 -32.36 4.21 -7.90 13.42 -27.72 -4.49 60.83 AL Lee Y 7.25 16.99 -2.84 -21.37 -6.27 -34.97 2.99 17.82 -18.89 11.81 AL Lowndes N N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A AL Macon N -24.52 12.08 -22.55 0.00 66.36 -63.48 -28.47 8.80 -77.44 -2.75 AL Montgomery Y -1.74 15.23 0.51 -23.39 -10.29 -10.81 2.02 0.40 -14.40 6.21 AL Pike N 11.90 3.62 3.84 -14.81 20.41 -23.22 -16.44 -24.43 -24.39 15.30 AL Randolph N 28.07 -13.99 4.22 -32.07 0.29 -51.32 -39.17 -20.85 -17.75 14.03 AL Russell N 33.70 38.09 4.22 -13.13 5.63 -33.04 1.12 -19.04 15.03 34.53 AL Tallapoosa Y 19.93 0.17 -14.19 -28.05 -13.92 -59.89 15.72 11.36 -61.30 43.28 GA GA 3.43 6.87 -5.45 -25.80 -10.55 -15.06 -1.20 -5.59 -12.85 10.64 GA Core Avg. -4.47 -3.56 0.51 -31.39 -10.36 -12.70 19.05 17.66 6.79 -1.17 GA Non-core Avg. 8.65 4.13 3.46 -24.14 -20.08 20.57 7.17 -22.44 -40.05 31.29 GA Atlanta MSA 1.36 7.45 -7.96 -27.54 -5.44 -15.39 -3.16 -5.97 -15.98 10.11 GA Harris N 28.81 2.91 1.16 -3.09 1.05 128.89 2.71 -7.86 -31.47 58.56 GA Heard N N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A GA Meriwether N N/A N/A N/A -34.52 -23.94 2.35 22.71 -35.41 -33.68 7.69 GA Talbot N N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A GA Troup Y -4.47 -3.56 0.51 -31.39 -10.36 -12.70 19.05 17.66 6.79 -1.17 GA Upson N -11.51 5.35 5.75 -34.80 -37.36 -69.54 -3.92 -24.05 -54.99 27.62 Notes: (1) Information on all the following tables are based on American Community Survey (ACS) data. The ACS is an ongoing survey that provides data every yea-r- giving communities the current information they need to plan investments and services. Data are estimates based on the collected survey answers. For example, ACS 3-year Estimates 2005-2007 are based on the answers of 2005, 2006 and 2007. These data are therefore treated as centered on the year 2006 (denoted by 06 above). Similarly, the ACS 3-year Estimates 2009-2011 are based on the answers of 2009, 2010 and 2011. These data are therefore treated as centered on the year 2010 (denoted by 10 above). The percentage difference data are calculated based on ACS 2005-2007, which is used to represent the socioeconomic situations before Kia came in 2008; and ACS 2009-2011, which is used to represent the socioeconomic situations after Kia came. 06-10 indicates that the centers of ACS 2005-2007 and ACS 2009-2011 are 2006 and 2010 respectively.. (2)ACS 3-year estimates of Heard, Talbot, Bullock, Lowndes, and Crenshaw and part of data of Butler and Meriwether are not available as the populations of those counties are too small and the ACS 3-year estimates only cover counties with population larger than 20,000. (3) The concept of Atlanta MSA is different from the official definition of Atlanta MSA which consists of 25 counties, because the 25 counties include many counties in the neighboring area of Kia plant, which we want to be separated from Atlanta. Instead, 15 counties are selected to create a smaller area of Atlanta MSA. Those counties are: Carroll, Cherokee, Clayton, Cobb, Coweta, DeKalb, Douglas, Fayette, Forsyth, Fulton, Gwinnett, Hall, Henry, Paulding, and Rockdale. The same definition of Atlanta MSA is used in every following table. 15 performance than the Alabama average and the Alabama non-core counties in management, wholesale trade, retail trade, and education and health care. Other sectors didn't seem to do better than other parts in Alabama. It is a little surprising that Chambers County, which has many newly established suppliers, didn't show significantly faster growth in any of the selected sectors. The reasons for this need further study. Table 2.2 presents data on the percentage change in migration to the various counties. The migration data provided by the ACS allows broad identification of whether those that came to a particular county migrated from another State or from within the same State, or whether they came from overseas. It also provides information on the citizenship and whether US versus foreign born. The key findings from Table 2.2 are that Troup County experienced a huge inflow of foreign residents. Numbers in residents from abroad, foreign-born citizen, naturalized citizen, and non-citizens almost doubled. For example, non-citizen residents in Troup increased from 1,151 to 2,114. Also, there is a 335% increase in residents from abroad, which is the second highest among all the selected counties, after Pike County. However, from the data on actual number of immigrants, we know that the high growth rate of Pike County is because of its low starting level. From the list of new establishments of suppliers we have, we found that most are Korean companies, thus we infer that a large proportion of the immigrants to the region are Koreans, working in those Korean companies and accompanied by their families. In terms of residents from other states, Troup County decreased less than non-core counties and the Atlanta MSA, which indicates that a part of the decline is offset by the inflow of workers from other states. Under the categories of foreign-born citizen, naturalized citizen, and non-citizens, Alabama core counties are also much higher than non-core counties. Notice the huge increases of Chambers County, where more than 7 suppliers locate. But we can't see a clear trend of domestic migration to the core counties. The major cause of domestic migration in this area may involve some factors other than the arrival of the KMMG plant and its suppliers. Changes in population dynamics, as noted in Table 2.2, will also result in changes in educational profile. Inflows of higher skilled production and services workers, along with management, are likely to alter the composition of educational attainment. Table 2.3 presents these data. Some of the main observations from Table 2.3 are as follows. In Georgia, Troup County had a larger increase in population 25 years or older, population with high school 16 Table 2.2 Percentage Change in Migration State County AL AL Core Residents from other counties, but same state 06-10 -0.16 Residents from other states 06-10 -14.70 Residents from abroad 06-10 -14.97 US born citizen 06-10 3.48 Foreign-born citizen 06-10 28.77 Naturalized citizen 06-10 25.05 Noncitizens 06-10 30.43 AL Core Avg. AL Non-core Avg. AL Autauga N AL Bullock Y AL Butler Y 3.16 12.93 37.09 N/A 29.37 4.41 6.55 4.33 N/A -21.98 40.07 237.44 -49.62 N/A N/A 2.43 4.02 11.93 N/A 2.41 177.82 58.25 31.07 N/A 565 180.34 91.15 -30.79 N/A N/A 98.73 68.45 120.76 N/A N/A AL Chambers Y AL Crenshaw N AL Elmore Y AL Lee Y AL Lowndes N AL Macon N 37.11 N/A -25.46 2.66 N/A 22.47 -38.80 N/A -22.47 7.24 N/A -31.52 136.67 N/A -19.07 -18.63 N/A 231.82 -3.23 N/A 4.50 8.98 N/A -6.40 262.50 N/A 39.79 41.00 N/A 21.91 746.67 N/A -9.80 3.41 N/A 425.00 198.23 N/A 83.41 57.86 N/A -87.50 AL Montgomery Y AL Pike N AL Randolph N AL Russell N AL Tallapoosa Y GA GA GA Core Avg. GA Non-core Avg. -2.50 3.83 -40.83 1.04 -4.42 77.33 688.46 7.52 -32.69 -41.34 -12.96 0.14 42.20 23.96 329.51 6.93 -22.25 98.62 142.22 0.87 -5.95 -26.10 -25.61 3.15 4.20 -28.74 335.09 4.56 22.19 -64.99 -59.29 3.30 32.76 94.45 161.54 -17.73 125.89 11.89 92.27 -11.47 55.58 16.36 78.57 -33.38 105.83 29.17 116.92 49.84 22.29 101.68 192.11 15.21 131.88 3.92 83.67 -26.59 GA Atlanta MSA -4.95 -33.54 -30.04 1.66 8.69 29.34 -0.71 GA Harris N GA Heard N 0.51 -71.79 N/A N/A 12.12 N/A 14.24 N/A -8.61 N/A -1.35 N/A -26.92 N/A GA Meriwether N GA Talbot N GA Troup Y 86.8 -69.41 -90.00 -2.96 N/A N/A N/A N/A 4.20 -28.74 335.09 4.56 -25.60 N/A 92.27 N/A N/A 116.92 N/A N/A 83.67 GA Upson N -20.75 -53.77 -100.00 -1.37 -0.21 101.03 -26.26 Notes: (1) Calculations are based on ACS 2005-2007 and ACS 2009-2011 data. (2) Unit of all the numbers are percentages. Data of Bullock, Crenshaw, Lowndes, Heard, and Talbot, and part of the data of Bulter and Meriwether are not available. (3) The 15 counties are selected to create a smaller area of Atlanta MSA. Those counties are: Carroll, Cherokee, Clayton, Cobb, Coweta, DeKalb, Douglas, Fayette, Forsyth, Fulton, Gwinnett, Hall, Henry, Paulding, and Rockdale. 17 education, and graduate degrees. In population with 9th to 12th grade education, Troup County increased by 13.91%, compared with -9.72 for Georgia average, -15.09 for non-core counties and -12.81 for Atlanta MSA. The number of people with some college education increased from 7,221 to 9,025. However, we observe a decline instead of increase in the number of people with associate or college degrees in Troup County. The combined results show that Troup County has been attracting people, but it is hard to conclude that the arrival of Kia and its suppliers has to date contributed to an accumulation of human capital within the local area. Also, we again observed a large increase in population in Harris County, which is the result of new military establishments. Inward movement of population and workers is also likely to alter the profile of schooling. Table 2.4 presents these data. In Troup County, compared with the Georgia average, Atlanta MSA, and non-core counties, we can see significant larger increases in schooling groups of preschool (27.38%), kindergarten (29.20%), and elementary school (10.03%). In contrast, the number of students in high schools and college or graduate schools didn't show a similarly large increase. This may be the results of workers of new companies bringing their children with them to Troup County. Given the fact that these children are mostly younger than high school age, we can conclude that the people arriving to work in the area's new businesses are mostly young to middle-age adults. Interestingly, Harris County's large increases are in nursery school students, high school students, and college students, which implies a different composition of employees at these military establishments. In Alabama, the two counties of Elmore County and Lee County show similar pattern to those in Troup County. But the reason is not clear why Pike County experienced the largest increase in student number among all the selected counties. Finally, in Table 2.5, we examine changes in household income. While we do not conduct an exercise to assess the tax implications for the County or State, the changes in incomes are a direct signal of economic benefits. For Georgia, the household income of Troup County experienced a larger increase (6.52% in median household income and 10.30% in mean household income) compared to the Georgia average, to Meriwether County and to Upson County. In contrast, Harris County achieved a huge increase of more than 22.3%. As noted before, the increase in Harris County may be the result of its military establishments. Thus, we can still conclude that Kia and its suppliers had significant positive impact on household income in the local area. For Alabama, although the core counties of Elmore and Lee show similar pattern to 18 Troup County, the impact is unclear for other core counties in Alabama. This makes it unclear whether on average the newly generated jobs are higher paid than the previously existing jobs. 3.2. Calculating Multipliers Many studies calculate the impact of State incentives by computing multipliers. Typically, these tend to be related to additional income generated, jobs created and tax revenues collected, among others. As noted above, we take an encompassing view of the effects of the KMMG plant location and operations, and examine a wide array of variables that inform us of the impact of the Kia plant. In Tables 2.1 to 2.5, we displayed the percentage changes in the affected counties, and discussed the changes between core and non-core counties, and how the core county effects compared with the State-wide or Atlanta MSA averages. In this section, we compute some basic multipliers to take a different look at the data. The starting point is the State of Georgia offering approximately $500 million in incentives. Next, the Kia plant locates and begins operations. The Kia plant has a direct employment of approximately 2,500 workers, and a capital investment of $1,200 million. This implies that the $500 million offered in State incentives results in an initial investment of $1,200 million and annual employment of 2,500 workers. These form the narrowest and most direct multipliers. To examine the totality of the effects, we need to consider that the location of the Kia plant resulted in: (1) numerous component suppliers moving to the area; and (2) various counties experiencing broader economic and business development, as manifested by changes in (a) employment in multiple occupations to support the activities of Kia and their suppliers, (b) inward movement of workers and their families with resulting effects on educational attainment and schooling, (c) intra-US and foreign migration patterns, and (d) income changes, among others. In other words, the full effect relates to the overall development of the economic and business ecosystem. As there is no single index we can meaningfully construct to measure this effect, we display the multipliers for each of the variables we consider, compare the multipliers for the core versus non-core counties, and also compare the multipliers for the core counties to State-wide and Atlanta MSA averages. 19 Table 2.3 Percentage Change in Education State County C Population o 25 years and r over e 06-10 Less than 9th grade 06-10 9th to 12th grade, no diploma 06-10 High school graduate (includes equivalency) 06-10 Some college, Associate no degree degree 06-10 06-10 Bachelor degree 06-10 Graduate or professional degree 06-10 AL AL 5.07 -4.23 -8.22 0.97 15.59 14.55 10.74 8.11 AL Core Avg. 5.15 -13.89 -5.33 3.91 14.68 6.25 14.93 21.56 AL Non-core Avg. 4.75 -10.59 -1.91 2.93 21.3 9.97 23.37 -6.93 AL Autauga N 11.75 7.68 5.95 7.12 21.54 18.21 15.59 3.24 AL Bullock Y N/A N/A N/A N/A N/A N/A N/A N/A AL Butler Y 5.08 -0.74 6.47 1.71 14.37 -13.97 4.03 55.9 AL Chambers Y -1.61 -15.22 -0.45 -4.53 10.16 -21.62 3.04 26.93 AL Crenshaw N N/A N/A N/A N/A N/A N/A N/A N/A AL Elmore Y 8.26 -30.3 -16.49 -6.31 32.52 37.61 24.47 46.77 AL Lee Y 11.99 -23.41 12.7 16.97 13.81 9.89 17.4 7.41 AL Lowndes N N/A N/A N/A N/A N/A N/A N/A N/A AL Macon N -3.14 -21.62 -1.68 10.17 13.65 -24.55 -28.36 -3.71 AL Montgomery Y 4.69 9.19 -15.25 4.23 14.12 1.07 14.77 -3.6 AL Pike N 5.75 -43.12 -18.84 11.32 33.02 4.96 32.7 -2.24 AL Randolph N 4.12 17.68 11.06 -8.97 32.28 -5.22 27.42 -35.07 AL Russell N 5.25 -13.56 -6.02 -4.97 5.99 56.46 69.48 3.14 AL Tallapoosa Y 2.49 -22.88 -18.94 11.37 3.11 24.49 25.84 -4.06 GA GA 5.01 -2.48 -9.72 1.23 15.85 9.88 6.98 11.33 GA Core Avg. 7.09 -11.08 13.91 4.8 24.98 -5.56 -3.41 10.25 GA Non-core Avg. 4.54 -8.78 -15.09 -1.01 30.2 -3.7 8.98 7.81 GA Atlanta MSA 3.02 1.9 -12.81 -3.85 11.88 8.34 4.3 9.45 GA Harris N 17.25 -24.67 -17.82 6.3 28.52 22.09 38.41 45.68 GA Heard N N/A N/A N/A N/A N/A N/A N/A N/A GA Meriwether N -2.82 5.85 1.26 -4.85 13.7 -38.99 6.13 -40.05 GA Talbot N N/A N/A N/A N/A N/A N/A N/A N/A GA Troup Y 7.09 -11.08 13.91 4.8 24.98 -5.56 -3.41 10.25 GA Upson N -0.81 -7.52 -28.72 -4.48 48.37 5.8 -17.59 17.81 Notes: (1) The calculations are based on ACS 2005-2007 and ACS 2009-2011 data. (2) Unit of all the numbers are percentages. Data of Bullock, Crenshaw, Lowndes, Heard, and Talbot are not available. ( 3). 15 counties are selected to create a smaller area of Atlanta MSA. Those counties are: Carroll, Cherokee, Clayton, Cobb, Coweta, DeKalb, Douglas, Fayette, Forsyth, Fulton, Gwinnett, Hall, Henry, Paulding, and Rockdale. 20 Table 2.4 Percentage Change in Schooling State County Core Population 3 years and over enrolled in school Nursery school, preschool Kindergarten Elementary school (grades 1-8) High school College or (grades 9-12) graduate school 06-10 06-10 06-10 06-10 06-10 06-10 AL AL 5.75 -2.37 4.78 2.78 2.3 16.19 AL Core Avg. 3.65 7.41 6.50 -0.18 -3.38 23.18 AL Non-core Avg. 11.25 3.94 -2.06 -2.01 11.09 39.37 AL Autauga N 21.77 -17.71 43.2 12.1 25.59 51.62 AL Bullock Y N/A N/A N/A N/A N/A N/A AL Butler Y 8.44 -9.96 -38.53 7.97 0.25 50.74 AL Chambers Y -3.11 19.47 -4.25 -7.1 -14.7 16.26 AL Crenshaw N N/A N/A N/A N/A N/A N/A AL Elmore Y 5.02 65.23 57.3 -6.46 2.45 9.01 AL Lee Y 9.02 29.84 48.57 4.4 -3.19 10.91 AL Lowndes N N/A N/A N/A N/A N/A N/A AL Macon N -17.09 20.3 -36.68 -32.27 11.03 -17.97 AL Montgomery Y 2.88 -14.77 -0.29 3.1 -5.13 13.63 AL Pike N 35.72 78.41 18.08 2.46 3.06 71.68 AL Randolph N -12.86 -55.8 -54.45 4.25 -11.8 -23.68 AL Russell N 28.7 -5.51 19.56 3.41 27.55 115.21 AL Tallapoosa Y -0.37 -45.35 -23.78 -2.96 0.04 38.52 GA GA 7.92 -3.68 7.23 4.07 2.74 23.4 GA Core Avg. 7.62 27.38 29.20 10.03 -8.15 7.48 GA Non-core Avg. 5.39 -2.33 -35.92 -3.36 24.05 30.92 GA Atlanta MSA 9.86 -2.45 7.38 4.58 6.94 27.39 GA Harris N 20.28 26.42 4.38 4.42 39.65 37.93 GA Heard N N/A N/A N/A N/A N/A N/A GA Meriwether N -11.92 -3.24 -55.56 -20.04 12.97 -0.81 GA Talbot N N/A N/A N/A N/A N/A N/A GA Troup Y 7.62 27.38 29.2 10.03 -8.15 7.48 GA Upson N 7.81 -30.18 -56.59 5.53 19.52 55.64 Notes: (1) Calculations are based on the ACS 2005-2007 and ACS 2009-2011 data. (2) Unit of all the numbers are percentages. Data of Bullock, Crenshaw, Lowndes, Heard, and Talbot are not available.(3) 15 counties are selected to create a smaller area of Atlanta MSA. Those counties are: Carroll, Cherokee, Clayton, Cobb, Coweta, DeKalb, Douglas, Fayette, Forsyth, Fulton, Gwinnett, Hall, Henry, Paulding, and Rockdale. 21 Table 2.5. Percentage Change in Household Income State County AL AL AL Core Avg. AL Non-core Avg. AL Autauga AL Bullock AL Butler AL Chambers AL Crenshaw AL Elmore AL Lee AL Lowndes AL Macon AL Montgomery AL Pike AL Randolph AL Russell AL Tallapoosa GA GA GA Core Avg. GA Non-core Avg. GA Atlanta MSA GA Harris GA Heard GA Meriwether GA Talbot GA Troup GA Upson Core N Y Y Y N Y Y N N Y N N N Y N N N N Y N M e dian Hous e hold Income M e an Hous e hold Income 10-Jun 4.8 2.32 11.29 11.28 N/A -7.9 -7.25 N/A 6.71 6.13 N/A 6.58 4.76 28.09 -1.16 11.64 11.46 -1.75 6.52 8.83 N/A 22.3 N/A 5.65 N/A 6.52 -1.46 10-Jun 5.93 2.73 9.64 10.81 N/A -3.63 2.73 N/A 10.11 8.97 N/A -5.55 -0.94 11.77 15.12 16.04 -0.85 0.08 10.3 5.32 N/A 22.36 N/A -0.18 N/A 10.3 -6.23 Notes (1) Calculations are based on the ACS 2005-2007 and ACS 2009-2011 data. (2) Unit of all the numbers are percentages. Data of Bullock, Crenshaw, Lowndes, Heard, Talbot, and Atlanta MSA are not available. (3) 15 counties are selected to create a smaller area of Atlanta MSA. Those counties are: Carroll, Cherokee, Clayton, Cobb, Coweta, DeKalb, Douglas, Fayette, Forsyth, Fulton, Gwinnett, Hall, Henry, Paulding, and Rockdale. 22 Our presentation of the multipliers related to the Kia facility is based on computing the following two ratios: (1) The actual change (pre-Kia to post-Kia) in the specific variable relative to Kia direct employment. We label this MuE. For example, if the actual change in a core county employment was 5,000 workers, then MuE is equal to 2.0 (that is, 5,000 divided by the Kia direct employment of 2,500); and (2) The actual change in the specific variable relative to Kia capital investment. We label this MuCI. For example, if the actual change in a core county employment was 5,000 workers, then MuCI is equal to 4.17 (that is, 5,000 divided by the Kia investment of 1,200). These multipliers are reported in Tables 3.1 to 3.5. In Table 3.1, we report the employment by occupation related multipliers. In Georgia, the core county, Troup County, outperformed its counterparts in the non-core counties in retail trade (0.265 additional jobs per Kia employment, 0.553 jobs per million Kia investments), in transportation and warehousing (0.092 additional jobs per Kia employment, 0.191 jobs per million Kia investments) and in finance and insurance (0.041 additional jobs per Kia employment, 0.085 jobs per million Kia investments). Troup County out-performed the non-core counties in sales and office and services, and was about the same in wholesale trade. In some categories, Troup County did worse than non-core counties; these include, for example, the industry categories of management, construction, and manufacturing. This shows that although the direct job creation effect of the KMMG plant is in the manufacturing sector, the increase of manufacturing jobs in this region is still weaker than in some other parts of Georgia. However, the induced job creation effect greatly benefited Troup County, mainly in retail trade, transportation and warehousing, and finance and insurance, all of which support the manufacturing factories and their workers. It is worth noting that during the analysis period, Harris County experienced significant growth in employment in the management sector, in wholesale trade, and in education and health care, but these are due to the recent arrival of military establishments. For Alabama, the core counties had better performance than non-core counties in the industry categories related to management, service, retail trade, and education and health care. In the other categories, the core counties did not do better than non-core counties. It is a little surprising that that Chambers County, which has many newly established suppliers, didn't show significantly faster growth in any of the selected sectors. 23 The dispersion of the employment results across the core and non-core are perhaps not surprising as many businesses tend to locate in neighboring counties. The precise dynamics of the location of the complementary and supporting businesses can only be addressed in a more detailed study. The effects related to migration are presented in Table 3.2. Troup County experienced a huge inflow of foreign residents. The multipliers for residents from abroad (ranging 0.15 to 0.32), foreign-born citizen (ranging from 0.57 to 1.19), naturalized citizen (ranging from 0.19 to 0.39, and non-citizens (ranging from 0.39 to 0.80) are all relatively large multipliers, and considerably greater than the corresponding numbers for the non-core counties. From the list of new supplier establishments, we found that most of them are Korean companies, and we infer from this that a large proportion of the immigrants are Korean workers and their families. In terms of residents from other states, Troup County decreased less than non-core counties, which indicates that part of the decline is offset by the inflow of workers from other states. For US-born residents, the multipliers for the core and non-core countries are large, and approximately the same. Perhaps the most interesting overall observation relates to the large movement of populations into Troup County that fall under the categories of residents from abroad, foreign-born citizens, naturalized citizens, and non-citizens. For the Alabama core counties, the multipliers for foreign-born citizen (ranging from 2.34 to 4.87), naturalized citizen (ranging from 0.58 to 1.21), and non-citizens (ranging from 1.71 to 3.56) are significantly higher than that state's non-core counties. However, in contrast to the Georgia core v. non-core counties, the Alabama core counties also have much larger multipliers for the US born citizen category (AL range 6.46 to 13.46 versus GA range of 1.11 to 2.32) This leads us to infer that the major cause of domestic migration into the Alabama areas may involve factors other than the arrival of the KMMG plant and its suppliers. This is not surprising as Alabama had been successful in attracting numerous prominent manufacturing facilities, in automobiles (e.g., Honda, Nissan, Hyundai, Toyota, and Daimler) and in other manufacturing (e.g., ThyssenKrupp, Airbus). Table 3.3 presents a transformation in educational attainment. In Georgia, Troup County had larger multipliers in categories related to 9th-12th grade (ranging from 0.29 to 0.60) and high school graduate (ranging from 0.27 to 0.57). For population 25+ and schooling less than 9th grade, Troup County multipliers were about the same as non-core counties. For 24 categories related to some college, associate degrees and bachelor degrees, the multipliers for Troup County were either marginally or considerably lower than non-core. The combined results show that Troup County has been attracting people, but it is hard to conclude that the coming of Kia and its suppliers contribute to an overall accumulation of human capital in that county. This is perhaps not surprising as many workers often tend to stay in adjacent counties, and not necessarily in the county they work in. In Alabama, the core county multipliers are greater than non-core counties in all the educational categories apart from less than 9th grade, and 9th to 12th grade. Overall, there is a significant measured effect on population with higher educational degrees. For the effects related to schooling, the multipliers are presented in Table 3.4. In Troup County, the multipliers are higher in all categories apart from high school, and college or graduate school. This implies that most of the effects are at the lower end of the schooling distribution. This may be the result of workers of new companies bringing their children with them to Troup County. Given the fact that these children are mostly younger than high school age, we can conclude that the people coming to work in the new businesses are mostly young to middle-age adults. Interestingly, Harris County has a somewhat different pattern, which may indicate a different composition of employees at the military establishments in that county. In Alabama, the core counties have systematically higher multipliers than non-core counties, except for the high school category. The population, educational and schooling patterns in the core versus non-core are complicated and difficult to provide a clean interpretation. This is largely due to the fact that workers need not stay in the same county as their work. Finally, in Table 3.5, we present the multipliers related to household income. For Georgia, we can see how the Kia plant has affected the household income of residents in Troup County versus the non-core counties (Meriwether and Upson) in Georgia. Aside from Harris County, the multipliers for Troup County are larger than the other non-core counties. The Harris County numbers are influenced by the military establishments. The Troup County households experienced a larger increase compared to Meriwether County and Upson County (every job brought into the Kia plant brings a $1 increase in the county's median household income and $2 in mean household income; one million dollars of investment in the Kia plant brings a $2 increase in median household income and a $4 in mean household income). Thus, we can still conclude that the Kia location boosted household income in the local area. For Alabama, 25 although the core counties of Elmore and Lee show a somewhat similar pattern of income gains (multipliers) as those in Troup County, the impact of the Kia plant on the incomes in the other core counties in Alabama is unclear. This also makes it unclear whether, on average, the newly generated jobs are higher paid than the existing jobs. Having said this, the Alabama results are more complicated and difficult to interpret due to the location of numerous other automobile and manufacturing plants. In summary, we see tangible evidence that the location of the Kia plant has affected population, schooling, educational and income dynamics in the affected (core) and non-core counties. Since many workers live and work in different counties, it is often difficult to pin down the precise effects in a particular county. But the overall picture is clear, that the location of the Kia plant has had a wide range of effects across the core and no-core counties in Georgia. The Alabama effects are more difficult to interpret due to the location of numerous other automobile and other manufacturing plants in that state. 26 State County AL Core Counties AL Non-core Counties AL Autauga AL Bullock AL Butler AL Chambers AL Crenshaw AL Elmore AL Lee AL Lowndes AL Macon AL Montgomery AL Pike AL Randolph AL Russell AL Tallapoosa GA Core County GA Non-core Counties GA Harris GA Heard GA Meriwether GA Talbot GA Troup GA Upson Total Core N Y Y Y N Y Y N N Y N N N Y N N N N Y N Table 3.1 Multipliers of Employment by Industry Management MuE MuCI 2.174 4.529 1.127 2.348 0.374 0.779 N/A N/A N/A N/A 0.113 0.235 N/A N/A 1.331 2.773 0.615 1.282 N/A N/A -0.220 -0.458 -0.247 -0.514 0.171 0.356 0.215 0.448 0.587 1.223 0.362 0.754 -0.140 -0.292 0.414 0.863 0.613 1.278 N/A N/A N/A N/A N/A N/A -0.140 -0.292 -0.199 -0.414 3.576 7.449 Service MuE MuCI 2.084 4.342 0.686 1.430 0.150 0.313 N/A N/A N/A N/A -0.172 -0.358 N/A N/A 0.638 1.330 0.603 1.256 N/A N/A 0.084 0.176 1.012 2.109 0.036 0.076 -0.071 -0.148 0.486 1.013 0.002 0.004 -0.069 -0.143 -1.085 -2.261 0.018 0.038 N/A N/A N/A N/A N/A N/A -0.069 -0.143 -1.104 -2.299 1.616 3.368 Sales and office MuE MuCI -0.426 -0.888 -0.042 -0.088 0.012 0.026 N/A N/A N/A N/A -0.171 -0.356 N/A N/A 0.071 0.148 -0.172 -0.359 N/A N/A -0.228 -0.474 0.052 0.108 0.054 0.113 0.028 0.058 0.091 0.190 -0.206 -0.428 0.014 0.029 -0.098 -0.205 0.017 0.036 N/A N/A N/A N/A N/A N/A 0.014 0.029 -0.116 -0.241 -0.553 -1.152 Construction MuE MuCI -1.845 -3.844 -0.496 -1.033 -0.260 -0.542 N/A N/A -0.023 -0.048 -0.224 -0.466 N/A N/A -0.436 -0.908 -0.406 -0.845 N/A N/A 0.000 0.000 -0.596 -1.242 -0.050 -0.105 -0.094 -0.197 -0.091 -0.189 -0.184 -0.383 -0.236 -0.493 -0.074 -0.154 -0.012 -0.025 N/A N/A -0.112 -0.234 N/A N/A -0.236 -0.493 0.050 0.105 -2.651 -5.523 Manufacturing MuE MuCI -1.418 -2.955 0.161 0.336 -0.213 -0.444 N/A N/A -0.494 -1.028 -0.632 -1.316 N/A N/A 0.068 0.142 -0.185 -0.386 N/A N/A 0.146 0.304 -0.431 -0.898 0.169 0.353 0.003 0.006 0.056 0.118 -0.238 -0.497 -0.257 -0.536 -0.143 -0.298 0.005 0.011 N/A N/A -0.199 -0.414 N/A N/A -0.257 -0.536 0.050 0.105 -1.658 -3.453 Notes: (1) The table is constructed based on the data in Table B.1. (2) MuE stands for Multiplier for Kia Employment and MuCI stands for Multiplier of Kia Capital Investment. MuE= Change in variable/Kia direct employment; MuCI= Change in variable/Kia capital investment. Kia direct employment is about 2,500 workers and Kia capital investment is $1,200 million. (3) For example, the multipliers under Management for Alabama core counties mean 1 employment in Kia plant on average brings 2.174 management jobs in Alabama core counties and 1 million capital investments in Kia plant on average brings 4.529 jobs in Alabama core counties. (4) By definition, the multipliers of AL core are the sums of the multipliers of the AL core counties; AL non-core, GA core, and GA non-core are calculated in similar ways. The "Total" row is calculated as the sum of all the county multipliers. (5) There exist negative multipliers as the employment in certain categories decrease from ACS 2005-2007 to ACS 2009-2011. (6) Data of Bullock, Crenshaw, Lowndes, Heard, Talbot, and part of the data for Butler and Meriwether are not available. 27 State County Core AL Core Counties AL Non-core Counties AL Autauga N AL Bullock Y AL Butler Y AL Chambers Y AL Crenshaw N AL Elmore Y AL Lee Y AL Lowndes N AL Macon N AL Montgomery Y AL Pike N AL Randolph N AL Russell N AL Tallapoosa Y GA Core County GA Non-core Counties GA Harris N GA Heard N GA Meriwether N GA Talbot N GA Troup Y GA Upson N Total Table 3.1. Multipliers of Employment by Industry ... Cont'd Wholesale trade MuE -0.449 -0.252 -0.052 N/A 0.013 -0.008 N/A -0.027 -0.180 N/A -0.045 -0.112 -0.040 -0.054 -0.060 -0.136 -0.035 -0.036 0.139 N/A 0.002 N/A -0.035 -0.177 -0.772 MuCI -0.936 -0.526 -0.109 N/A 0.028 -0.016 N/A -0.057 -0.374 N/A -0.094 -0.234 -0.084 -0.113 -0.125 -0.283 -0.073 -0.075 0.290 N/A 0.004 N/A -0.073 -0.369 -1.609 Retail trade MuE 0.520 -0.350 0.066 N/A 0.058 0.015 N/A 0.178 0.090 N/A -0.117 0.094 -0.126 -0.184 0.011 0.087 0.265 -0.295 0.017 N/A 0.081 N/A 0.265 -0.393 0.140 MuCI 1.084 -0.730 0.137 N/A 0.120 0.031 N/A 0.370 0.187 N/A -0.244 0.195 -0.262 -0.384 0.023 0.182 0.553 -0.614 0.036 N/A 0.169 N/A 0.553 -0.819 0.293 Transportation and warehousing MuE MuCI -0.114 -0.237 -0.027 -0.057 0.190 0.395 N/A N/A 0.016 0.034 -0.088 -0.183 N/A N/A -0.208 -0.433 0.131 0.273 N/A N/A 0.009 0.018 0.006 0.013 -0.112 -0.233 -0.049 -0.102 -0.065 -0.136 0.028 0.059 0.092 0.191 -0.177 -0.369 -0.016 -0.033 N/A N/A -0.107 -0.223 N/A N/A 0.092 0.191 -0.055 -0.114 -0.226 -0.472 Finance and insurance MuE MuCI -1.045 -2.178 -0.091 -0.190 0.047 0.098 N/A N/A -0.013 -0.027 0.009 0.018 N/A N/A -0.037 -0.077 -0.316 -0.658 N/A N/A -0.165 -0.343 -0.424 -0.883 -0.064 -0.133 -0.016 -0.034 0.107 0.223 -0.265 -0.552 0.041 0.085 -0.325 -0.677 -0.228 -0.475 N/A N/A -0.077 -0.160 N/A N/A 0.041 0.085 -0.020 -0.042 -1.420 -2.959 Education and health care MuE MuCI 3.068 6.391 1.188 2.475 0.429 0.893 N/A N/A 0.012 0.025 0.082 0.171 N/A N/A 1.118 2.328 0.741 1.543 N/A N/A -0.032 -0.067 0.495 1.032 0.199 0.414 0.093 0.193 0.500 1.041 0.620 1.292 -0.027 -0.057 0.630 1.313 0.654 1.363 N/A N/A 0.047 0.098 N/A N/A -0.027 -0.057 -0.071 -0.148 4.859 10.123 28 Table 3.2. Multipliers of Migration State County Core AL Core Counties AL Non-core Counties AL Autauga N AL Bullock Y AL Butler Y AL Chambers Y AL Crenshaw N AL Elmore Y AL Lee Y AL Lowndes N AL Macon N AL Montgomery Y AL Pike N AL Randolph N AL Russell N AL Tallapoosa Y GA Core County GA Non-core Counties GA Harris N GA Heard N GA Meriwether N GA Talbot N GA Troup Y GA Upson N Total Residents from other counties MuE MuCI Residents from other states MuE MuCI Residents from abroad MuE MuCI US born citizen MuE MuCI Foreign-born citizen MuE MuCI -0.860 0.307 0.237 N/A 0.030 0.080 N/A -0.808 0.087 N/A 0.049 -0.075 -0.038 -0.088 0.147 -0.174 0.043 0.210 0.004 N/A 0.316 N/A 0.043 -0.110 -0.301 -1.793 0.639 0.493 N/A 0.062 0.167 N/A -1.683 0.181 N/A 0.102 -0.156 -0.080 -0.183 0.307 -0.363 0.090 0.138 0.138 0.036 N/A -0.016 -0.215 N/A -0.159 0.207 N/A -0.218 0.122 0.236 -0.175 0.259 0.200 -0.276 0.437 0.008 N/A 0.658 N/A 0.090 -0.229 -0.627 -0.474 -0.258 N/A -0.108 N/A -0.276 -0.108 -0.474 0.288 0.288 0.074 N/A -0.033 -0.448 N/A -0.332 0.431 N/A -0.454 0.253 0.492 -0.364 0.540 0.417 -0.575 -0.325 0.367 -0.026 N/A 0.021 0.016 N/A -0.015 -0.090 N/A 0.020 -0.283 0.215 -0.003 0.161 0.026 0.153 -0.988 -0.537 N/A -0.225 N/A -0.575 -0.226 -0.987 -0.057 0.002 N/A -0.022 N/A 0.153 -0.037 0.138 -0.677 0.765 -0.054 N/A 0.044 0.034 N/A -0.031 -0.188 N/A 0.043 -0.590 0.448 -0.006 0.335 0.053 0.318 6.459 3.920 2.265 N/A 0.194 -0.449 N/A 1.328 4.354 N/A -0.566 0.892 0.876 0.013 1.332 0.139 1.112 -0.118 1.120 0.003 N/A -0.045 N/A 0.318 -0.077 0.288 1.534 N/A -0.266 N/A 1.112 -0.148 12.612 13.457 8.168 4.718 N/A 0.405 -0.935 N/A 2.767 9.072 N/A -1.178 1.859 1.826 0.027 2.775 0.289 2.317 2.333 3.195 N/A -0.553 N/A 2.317 -0.308 26.274 2.335 0.397 0.088 N/A 0.045 0.134 N/A 0.184 0.824 N/A 0.031 0.922 0.245 0.101 -0.068 0.226 0.573 -0.054 -0.028 N/A -0.026 N/A 0.573 0.000 3.251 4.865 0.828 0.183 N/A 0.094 0.280 N/A 0.383 1.717 N/A 0.065 1.921 0.511 0.210 -0.142 0.470 1.194 -0.113 -0.059 N/A -0.053 N/A 1.194 -0.001 6.773 Naturalized citizen Non-citizens MuE MuCI MuE MuCI 0.580 0.008 -0.052 N/A 0.000 0.045 N/A -0.021 0.021 N/A 0.129 0.492 0.004 0.013 -0.087 0.044 0.188 0.036 -0.003 N/A 0.000 N/A 0.188 0.039 0.812 1.209 0.016 -0.108 N/A 0.000 0.093 N/A -0.044 0.044 N/A 0.269 1.025 0.008 0.028 -0.181 0.091 0.392 0.075 -0.007 N/A 0.000 N/A 0.392 0.082 1.692 1.710 0.390 0.140 N/A 0.000 0.090 N/A 0.205 0.803 N/A -0.098 0.430 0.242 0.088 0.019 0.182 0.385 -0.165 -0.025 N/A -0.100 N/A 0.385 -0.040 2.320 3.562 0.812 0.291 N/A 0.000 0.187 N/A 0.428 1.673 N/A -0.204 0.896 0.503 0.183 0.039 0.379 0.803 -0.343 -0.053 N/A -0.208 N/A 0.803 -0.083 4.833 Notes: (1) This table is constructed based on the data in Table B.2. (2) MuE stands for Multiplier for Kia Employment and MuCI stands for Multiplier of Kia Capital Investment. MuE= Change in variable/Kia direct employment; MuCI= Change in variable/Kia capital investment. Kia direct employment is about 2,500 workers and Kia capital investment is $1,200 million. (3) Data of Bullock, AL, Crenshaw, AL, Lowndes, AL, Heard, GA, Talbot, GA, and part of the data for Butler, AL and Meriwether, GA are not available. (4) By definition, the multipliers of AL core are the sums of the multipliers of the AL-core counties; AL non-core, GA core, and GA non-core are calculated in similar ways. The total row is calculated as the sum of all the county multipliers. 29 Table 3.3. Multipliers of Education State County Population 25 years 9th to 12th grade, High school graduate Core and over Less than 9th grade no diploma (includes equivalency) MuE MuCI MuE MuCI MuE MuCI MuE MuCI AL Core Counties 8.056 16.784 -0.831 -1.731 -1.437 -2.993 1.726 3.596 AL Non-core Counties 2.67 5.562 -0.366 -0.763 -0.212 -0.441 0.27 0.563 AL Autauga N 1.482 3.088 0.048 0.099 0.078 0.163 0.322 0.67 AL Bullock Y N/A N/A N/A N/A N/A N/A N/A N/A AL Butler Y 0.271 0.564 -0.003 -0.007 0.058 0.122 0.034 0.071 AL Chambers Y -0.154 -0.321 -0.128 -0.268 -0.008 -0.018 -0.154 -0.321 AL Crenshaw N N/A N/A N/A N/A N/A N/A N/A N/A AL Elmore Y 1.624 3.384 -0.344 -0.718 -0.425 -0.885 -0.457 -0.952 AL Lee Y 3.409 7.102 -0.354 -0.737 0.372 0.774 1.254 2.612 AL Lowndes N N/A N/A N/A N/A N/A N/A N/A N/A AL Macon N -0.167 -0.348 -0.111 -0.231 -0.013 -0.027 0.134 0.279 AL Montgomery Y 2.625 5.468 0.249 0.518 -1.048 -2.184 0.649 1.353 AL Pike N 0.415 0.864 -0.267 -0.556 -0.253 -0.527 0.27 0.563 AL Randolph N 0.247 0.515 0.114 0.238 0.107 0.223 -0.214 -0.446 AL Russell N 0.692 1.443 -0.15 -0.313 -0.131 -0.273 -0.241 -0.503 AL Tallapoosa Y 0.282 0.587 -0.25 -0.521 -0.385 -0.803 0.4 0.833 GA Core County 1.13 2.354 -0.143 -0.298 0.288 0.599 0.273 0.568 GA Non-core Counties 1.073 2.236 -0.125 -0.26 -0.538 -1.122 -0.11 -0.228 GA Harris N 1.302 2.713 -0.104 -0.217 -0.125 -0.261 0.143 0.298 GA Heard N N/A N/A N/A N/A N/A N/A N/A N/A GA Meriwether N -0.17 -0.353 0.03 0.062 0.013 0.028 -0.124 -0.258 GA Talbot N N/A N/A N/A N/A N/A N/A N/A N/A GA Troup Y 1.13 2.354 -0.143 -0.298 0.288 0.599 0.273 0.568 GA Upson N -0.06 -0.124 -0.05 -0.105 -0.426 -0.888 -0.129 -0.268 Total 12.929 26.936 -1.464 -3.051 -1.899 -3.957 2.16 4.499 Notes: (1) This table is constructed based on the data in Table B.3. (2) MuE stands for Multiplier for Kia Employment and MuCI stands for Multiplier of Kia Capital Investment. MuE= Change in variable/Kia direct employment; MuCI= Change in variable/Kia capital investment. Kia direct employment is about 2,500 workers and Kia capital investment is $1,200 million.( 3) Data of Bullock, AL, Crenshaw, AL, Lowndes, AL, Heard, GA, and Talbot, GA are not available. (4) By definition, the multipliers of AL core are the sums of the multipliers of the AL core counties; AL non-core, GA core, and GA non-core are calculated in similar ways. The total row is calculated as the sum of all the county multipliers. 30 Table 3.3. Multipliers of Education ... Cont'd State County Core Some college, no degree Associate's degree Bachelor's degree Graduate or professional degree MuE MuCI MuE MuCI MuE MuCI MuE MuCI AL Core Counties 4.015 8.365 0.674 1.404 3.218 6.704 0.691 1.439 AL Non-core Counties 1.571 3.273 0.472 0.983 1.046 2.180 -0.112 -0.234 AL Autauga N 0.579 1.206 0.160 0.333 0.268 0.558 0.029 0.060 AL Bullock Y N/A N/A N/A N/A N/A N/A N/A N/A AL Butler Y 0.130 0.271 -0.063 -0.131 0.020 0.041 0.095 0.198 AL Chambers Y 0.185 0.385 -0.151 -0.314 0.020 0.043 0.082 0.172 AL Crenshaw N N/A N/A N/A N/A N/A N/A N/A N/A AL Elmore Y 1.277 2.660 0.454 0.945 0.604 1.258 0.516 1.075 AL Lee Y 0.791 1.648 0.211 0.440 0.847 1.765 0.288 0.599 AL Lowndes N N/A N/A N/A N/A N/A N/A N/A N/A AL Macon N 0.144 0.301 -0.114 -0.238 -0.187 -0.390 -0.020 -0.043 AL Montgomery Y 1.568 3.267 0.032 0.067 1.437 2.994 -0.262 -0.546 AL Pike N 0.384 0.799 0.012 0.024 0.283 0.589 -0.014 -0.028 AL Randolph N 0.291 0.606 -0.021 -0.043 0.094 0.196 -0.124 -0.258 AL Russell N 0.174 0.362 0.435 0.907 0.589 1.228 0.017 0.035 AL Tallapoosa Y 0.064 0.134 0.191 0.398 0.290 0.603 -0.028 -0.058 GA Core County 0.722 1.503 -0.054 -0.113 -0.066 -0.138 0.112 0.233 GA Non-core Counties 1.142 2.379 0.049 0.103 0.392 0.818 0.262 0.547 GA Harris N 0.447 0.931 0.145 0.303 0.462 0.963 0.334 0.697 GA Heard N N/A N/A N/A N/A N/A N/A N/A N/A GA Meriwether N 0.131 0.273 -0.118 -0.245 0.022 0.045 -0.124 -0.258 GA Talbot N N/A N/A N/A N/A N/A N/A N/A N/A GA Troup Y 0.722 1.503 -0.054 -0.113 -0.066 -0.138 0.112 0.233 GA Upson N 0.564 1.175 0.022 0.045 -0.092 -0.191 0.052 0.108 Total 7.450 15.521 1.140 2.376 4.590 9.563 0.952 1.984 Notes: (1) This table is constructed based on the data in Table B.3 Cont'd. (2) MuE stands for Multiplier for Kia Employment and MuCI stands for Multiplier of Kia Capital Investment. MuE= Change in variable/Kia direct employment; MuCI= Change in variable/Kia capital investment. Kia direct employment is about 2,500 workers and Kia capital investment is $1,200 million. (3) Data of Bullock, AL, Crenshaw, AL, Lowndes, AL, Heard, GA, and Talbot, GA are not available. (4) By definition, the multipliers of AL core are the sums of the multipliers of the AL core counties; AL non-core, GA core, and GA non-core are calculated in similar ways. The total row is calculated as the sum of all the county multipliers. 31 Table 3.4. Multipliers of Schooling State County Core Population 3 years and over enrolled in school Nursery school, preschool Elementary Kindergarten school (grades 1-8) High school (grades 9-12) College or graduate school MuE MuCI MuE MuCI MuE MuCI MuE MuCI MuE MuCI MuE MuCI AL Core Counties 2.906 6.054 0.102 0.213 0.342 0.713 0.232 0.484 -0.415 -0.865 2.645 5.510 AL Non-core Counties 2.845 5.927 -0.062 -0.128 0.071 0.148 0.122 0.253 0.609 1.269 2.104 4.384 AL Autauga N 1.113 2.319 -0.056 -0.118 0.100 0.209 0.296 0.617 0.313 0.653 0.460 0.958 AL Bullock Y N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A AL Butler Y 0.162 0.338 -0.010 -0.020 -0.052 -0.109 0.072 0.149 0.001 0.003 0.151 0.315 AL Chambers Y -0.100 -0.208 0.038 0.080 -0.008 -0.017 -0.107 -0.223 -0.113 -0.235 0.090 0.187 AL Crenshaw N N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A AL Elmore Y 0.377 0.785 0.224 0.466 0.209 0.435 -0.224 -0.466 0.050 0.103 0.118 0.247 AL Lee Y 1.748 3.642 0.233 0.486 0.258 0.537 0.232 0.483 -0.084 -0.174 1.109 2.310 AL Lowndes N N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A AL Macon N -0.574 -1.197 0.027 0.057 -0.034 -0.070 -0.321 -0.668 0.052 0.109 -0.300 -0.624 AL Montgomery Y 0.733 1.527 -0.270 -0.563 -0.004 -0.008 0.309 0.643 -0.270 -0.563 0.968 2.017 AL Pike N 1.297 2.702 0.083 0.173 0.020 0.041 0.030 0.062 0.019 0.039 1.146 2.388 AL Randolph N -0.287 -0.598 -0.100 -0.208 -0.061 -0.128 0.042 0.088 -0.057 -0.118 -0.111 -0.231 AL Russell N 1.296 2.700 -0.015 -0.032 0.046 0.096 0.075 0.156 0.282 0.587 0.909 1.893 AL Tallapoosa Y -0.014 -0.028 -0.113 -0.236 -0.060 -0.126 -0.049 -0.103 0.000 0.001 0.209 0.435 GA Core County 0.515 1.073 0.134 0.278 0.116 0.242 0.293 0.610 -0.130 -0.270 0.102 0.213 GA Non-core Counties 0.506 1.054 -0.033 -0.069 -0.206 -0.430 -0.094 -0.195 0.438 0.912 0.402 0.837 GA Harris N 0.577 1.202 0.048 0.101 0.008 0.018 0.056 0.118 0.254 0.530 0.209 0.436 GA Heard N N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A GA Meriwether N -0.272 -0.568 -0.006 -0.013 -0.110 -0.229 -0.215 -0.448 0.062 0.128 -0.003 -0.006 GA Talbot N N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A GA Troup Y 0.515 1.073 0.134 0.278 0.116 0.242 0.293 0.610 -0.130 -0.270 0.102 0.213 GA Upson N 0.202 0.420 -0.075 -0.157 -0.105 -0.218 0.065 0.135 0.122 0.253 0.195 0.407 Total 6.772 14.108 0.141 0.293 0.323 0.673 0.553 1.153 0.502 1.046 5.253 10.943 Notes (1) This table is constructed based on the data in Table B.4. (2) MuE stands for Multiplier for Kia Employment and MuCI stands for Multiplier of Kia Capital Investment. MuE= Change in variable/Kia direct employment; MuCI= Change in variable/Kia capital investment. Kia direct employment is about 2,500 workers and Kia capital investment is $1,200 million. (3) Data of Bullock, AL, Crenshaw, AL, Lowndes, AL, Heard, GA, and Talbot, GA are not available. (4) By definition, the multipliers of AL core are the sums of the multipliers of the AL core counties; AL non-core, GA core, and GA non-core are calculated in similar ways. The total row is calculated as the sum of all the county multipliers. 32 Table 3.5. Multipliers of Household Income State County Core Median Household Income MuE MuCI Mean Household Income MuE MuCI AL Core Counties 2.749 5.728 3.714 7.737 AL Non-core Counties 6.954 14.488 8.914 18.571 AL Autauga N 2.168 4.516 2.529 5.268 AL Bullock Y N/A N/A N/A N/A AL Butler Y -1.006 -2.097 -0.600 -1.249 AL Chambers Y -0.973 -2.028 0.450 0.937 AL Crenshaw N N/A N/A N/A N/A AL Elmore Y 1.360 2.833 2.434 5.072 AL Lee Y 0.953 1.985 1.840 3.834 AL Lowndes N N/A N/A N/A N/A AL Macon N 0.702 1.462 -0.946 -1.970 AL Montgomery Y 0.800 1.666 -0.229 -0.478 AL Pike N 2.792 5.817 1.936 4.034 AL Randolph N -0.162 -0.338 2.824 5.883 AL Russell N 1.455 3.032 2.570 5.355 AL Tallapoosa Y 1.616 3.368 -0.182 -0.379 GA Core County 1.025 2.135 2.005 4.177 GA Non-core Counties 5.779 12.039 6.627 13.806 GA Harris N 5.088 10.599 6.444 13.426 GA Heard N N/A N/A N/A N/A GA Meriwether N 0.804 1.674 -0.033 -0.068 GA Talbot N N/A N/A N/A N/A GA Troup Y 1.025 2.135 2.005 4.177 GA Upson N -0.112 -0.234 0.215 0.448 Total 16.507 34.390 21.259 44.290 Notes (1) This table is constructed based on the data in Table B.5. (2) MuE stands for Multiplier for Kia Employment and MuCI stands for Multiplier of Kia Capital Investment. MuE= Change in variable/Kia direct employment; MuCI= Change in variable/Kia capital investment. Kia direct employment is 2,500 and Kia capital investment is $1,200 million. (3) Data of Bullock, AL, Crenshaw, AL, Lowndes, AL, Heard, GA, Talbot, GA, and Atlanta MSA are not available. (4) By definition, the multipliers of AL core are the sums of the multipliers of the core counties; AL non-core, GA core, and GA non-core are calculated in similar ways. The total row is calculated as the sum of all the county multipliers. 33 4. Supply Chains: Components and Final Product Flows This section of the report focuses on the various freight flows associated with the automobile manufacturing supply chain, and its uses of local, regional and national highway, rail and waterway (including seaport) networks and cargo transfer facilities. 4.1 Introduction When companies such as Kia Motors move their operations into a new region, they have put a good deal of thought and research into the benefits of doing so. One important consideration is the ability to operate a highly efficient and consistently reliable just-in-time (JIT) materials and parts delivery process, as well as a similarly time sensitive and cost efficient finished products (i.e. finished automobiles) delivery process. This means operating an effective product supply chain that involves a number of sequential, including inter-modal freight movements that are now integral to, and a significant cost component of, the overall production process. Today, a key supply chain requirement is therefore an accessible, reliable, and high capacity global, as well as regional, transportation network. The public sector role in this process includes maintaining and, where necessary, facilitating the expansion of such networks. For States, this public sector role has become a key component in both attracting and retaining large manufacturing facilities, such as Kia's West Point automobile manufacturing plant, and one that deals increasingly with the issue of ensuring that disruptions to the transport of goods into and out of such sites are kept to a minimum. This includes traffic bottlenecks that result from either specific, non-recurring events (crashes, bad weather, roadway damage, necessary network rehabilitation) or that emerge over time from the continued growth in both freight and passenger traffic volumes. 2 As Weisbrod and Fitzroy (2011)3 put it: "From the public perspective, there is a need to make investment, financing and policy decisions based on an understanding of public infrastructure needs, costs and broader economic stakes 2 Cambridge Systematics Inc. and Texas Transportation Institute (2005) Traffic Congestion and Reliability: Trends and Advanced Strategies for Congestion Mitigation. Report to the Federal Highway Administration, Washington, D.C. 3 Weisbrod, G. and Fitzroy, S. (2011) Traffic congestion effects on supply chains: accounting for behavioral elements in planning and economic impact models. Chapter 16 in Supply Chain Management New Perspectives, Renko, S. (Ed.) www.intechopen.com 34 involved. From the perspective of shippers and carriers, there are the day-to-day cost implications of delay and reliability as it affects supply chain management, as well as a longerrange need to assess opportunities, risks and returns associated with location, production and distribution decisions. Both perspectives need to be recognized when considering the full range of impacts that traffic congestion can have on the economy." Bringing these two perspectives together in a productive manner presents a significant challenge, and one that requires, among other things, that public agency transportation planners expand their current efforts beyond traditional measurement of in-vehicle travel times and monetary costs, to broader considerations of industry sector specific production processes. As Holl (2006, page 11)4 puts it, in one of the very few papers to address the role of regional transportation infrastructure investments from the individual firm perspective: "New patterns of production and distribution are emerging that are increasingly dependent on high-quality transportation. In an increasingly time-based competitive environment, access to the higher order road network and issues of reliability and frequency are becoming more important than just pecuniary transport costs." That is, such benefits can go beyond the usual estimates of in-vehicle transit travel time and cost savings, notably by providing opportunities to benefit from logistical reorganization, from market area expansion, and from wider supplier access. She also concludes that "microlevel knowledge" of firm operations is important for correct public sector evaluation of such transportation (infrastructure investment) projects. The trend towards low-inventory JIT ordering and delivery systems in manufacturing industries such as automobile production places a considerable additional burden on the consistent, on-time supply of product inputs. In a summer 2013 interview at Kia Motors as part of this project's activities, the authors were told that a production stoppage to the automated assembly of automobiles of as little as one hour could prove to be very costly. They also learned that dealership orders based on specific vehicle specifications are often met within the southeastern region by dispatching auto-carriers to fulfill such orders, often covering a considerable travel distance over the highway system. A similar on-time imperative has increasingly become the norm in other industries where the delivery of finished products is concerned. Often termed a 4 Holl, A. (20 06) A Review of the Firm-Level Role of Transport Infrastructure with Implications for Transport Project Evaluation. Journal of Planning Literature 21.1: 3-14. 35 "pull" supply chain in which the receiver/final customer determines the delivery schedule, this puts pressure on the manufacturer to maintain reliable transportation services. Solutions to this problem include locating suppliers near to, and sometimes within, a manufacturing plant. Another option is to integrate the manufacture and delivery of components within different branches of the Original Equipment Manufacturer's (OEM's) own company: a movement that is at odds with the vertical disintegration of the production process that has led many manufacturing companies to outsource the creation of sub-components of the firm's core end product. Whatever the "production and delivery model" adopted, transportation costs need to be treated as one component in a series of increasingly interdependent product delivery costs. The rest of this section of the report is used to explore the nature and monetary value of such JIT-based transportation costs, and to assess the availability of existing data sources for doing so. We begin in Section 4.2 with a review of the recent literature on traffic congestion, its recent and projected growth in the south-eastern United States, and its various direct and indirect effects on a manufacturing firm's production costs, examined from the single firm's perspective. This in effect becomes an examination of how we quantify the costs of delays to freight pickups and deliveries. In doing so, and in the context of expected strong and continued growth in overall freight traffic volumes across the south-eastern region 5 , we identify a number of potentially significant freight movement bottlenecks that may affect future movement efficiencies at the firm as well as broader regional level of operation. Sections 4.3 and 4.4 describe our efforts to construct a database of both the flows and costs respectively associated with moving freight inputs and outputs through the case study-based automotive industry supply chain. Section 4.5 then describes a method for bringing these flows and costs into a modeling framework for both identifying and computing the monetary costs of freight movement delays within JIT manufacturing supply chains. 4.2 Measuring the Effects of Congestion-Induced Delays on Firm Transaction Costs 4.2.1 Current and Forecast Congestion in the South-East and Nationally The growing costs of traffic congestion have attracted a good deal of interest over the past decade, notably highway congestion and the efforts of state and regional planners to deal 5 Georgia Statewide Freight and Logistics Plan, 2010-2050. Task 4. Economic Evaluation and Projections. Office of Planning, Georgia DOT. 36 with mixed passenger and freight mobility and place accessibility issues. Both federally and regionally sponsored studies of traffic growth and its impacts in the south-east contain forecasts of freight movement activity that will put a great deal of stress on current transportation networks, including water (seaport) 6 and rail 7 as well as highway 8 travel supporting infrastructure. A recent study by the American Transportation Research Institute 9 estimates that highway congestion cost trucking firms over $9.2 billion in additional operating costs and 141 million hours of lost productivity in 2013: with the Atlanta region rated the 10th most impacted metro area in the country. Based on the manufacturing supply chain data collected and described in Sections 4.3 and 4.4 below, congestion is expected to impact large manufacturing establishments such as the Kia Motors plant principally at three different geographic scales, each of which is also dominated by a specific mode or modal combination: highway congestion both locally and within the south-eastern US congestion on the region's and the nation's rail network, and congestion at both foreign and US seaports associated with the inter-continental transport of waterborne (containerized) freight Highway Congestion Figure 4.1 shows the regional highway traffic forecasts for the year 2040, based on the mixed passenger and truck traffic volumes forecast by the US DOT's Freight Analysis Framework (Version 3: FHWA, 2010). The "VCR40" shown in the map key refers to the forecast traffic volume/roadway capacity (v/c) ratio in 2040, based on the assumption that no significant additional highway capacity has been added to the network. Many links have v/c ratios much higher than 1.2, with many of these links reaching this stage of heavy delay-inducing congestion well before 2040. Based on these ratios, the yellow colored links on this map indicate an average (space mean) speed reduction of over 25% between 2007 and 2040, the blue colored 6 Volpe (2009) Assessment of the Marine Transportation System (MTS) Challenges. Summary Report to the U.S. Army Corps of Engineers by the Volpe Transportation Research Center, Cambridge, MA. 7 Cambridge Systematics Inc. (2007) National Rail Freight Infrastructure Capacity and Investment Study. Chapter 5: Capacity and Performance Analysis. Report to the American Association of Railroads. September, 2007. 8 FAF3 Freight Traffic Analysis (2011) Report to Oak Ridge National Laboratory by Battelle, Columbus, OH. http://faf.ornl.gov/fafweb/Data/Freight_Traffic_Analysis/index.htm 9 Pierce, D. and Murray, D. (2014) Costs of congestion to the trucking industry American Transportation Research Institute, Arlington, VA. http://atri-online.org/ 37 links indicate a greater than 40% reduction, and the red links a decrease of more than 70% in link speeds over the same period. Figure 4.1 Regional Highway Congestion Forecast for 2040* * Data source: See footnote 8. Recognizing the traffic growth potential of the region, the Georgia DOT's Freight and Logistics Plan for 2010-50 (GDOT, 2013)10 analyzed a number of potential network upgrades. These include upgrading both the north-south U.S 27 corridor and east-west Macon-Lagrange Bypass. Both of these corridors are shown in Figure 4.1, and each would offer additional capacity to the highway network in the vicinity of the Kia Motors plant. Figure 4.2 shows the full national picture of this 2040 traffic forecast, highlighting the fact that the south-eastern states including Alabama, Florida, Georgia, Tennessee, and the Carolinas are expected to see a good deal of traffic congestion, comparable to similar slow-downs in between as well as within metropolitan area traffic flows in the north-east, mid-west, Texas, and southern California. 10 Georgia Statewide Freight and Logistics Action Plan. Georgia DOT. http://www.dot.ga.gov/Projects/programs/georgiafreight/Pages/default.aspx 38 Figure 4.2 National Highway Congestion Forecast for 2040* * Data source: See footnote 8. Rail Congestion Similar to the nation's major highways, significant traffic congestion is also expected on many mainline railroads by 2035, failing significant investment in network carrying capacities.11 With only 5% of its Class 1 rail mileage double tracked, and both weight restrictions and bridge clearance issues associated with some shortlines, bottlenecks are already beginning to impact railroad delivery times within Georgia.12 Both Georgia's Rail Plan13 and its more recent Freight and Logistics Plan for 2010-50 recognize this network investment challenge. One potentially positive step forward here is the December 2011 opening of the Cordele Intermodal Center, which might help to link the Kia Motors plant and places to the West to the Port of Savannah's Garden City terminal via CSXT and the Heart of Georgia (HOG) and Georgia Central (GC) 11 See FHWA's on-line Freight Story 2008 report and associated national rail network forecast maps at http://ops.fhwa.dot.gov/freight/freight_analysis/freight_story/congestion.htm#railroad (based on reference in footnote 6 above) 12 See footnote 10. 13 See Footnote 10, as well as Georgia DOT (2009) State Rail Plan. http://www.dot.ga.gov/travelingingeorgia/rail/Pages/StateRailPlan.aspx 39 shortline railroads. Figure 4.3 shows the State's major rail lines, the location of this Cordele intermodal facility, and the State's two seaports with respect to the Kia plant. Also highlighted (in grey) are the rail corridors classified by GDOT's Statewide Freight and Logistics Plan as already experiencing bottleneck conditions: with "significant growth" in rail traffic expected along all but one of these rail lines. Figure 4.3 Major Rail Lines and Current Congestion Levels in Georgia* * Data source: See footnote 7. Seaport Congestion A key question for Georgia and the rest of the east coast states with major seaports is how cost-competitive the impending opening of the newly expanded Panama Canal channel will be.14 From a geographic perspective, the Port of Savannah seems to be well placed to benefit from the 14 MARAD (2013) The Panama Expansion Study. Phase 1 Report. Developments in Trade and National and Global Economies. Maritime Administration, US Department of Transportation, Washington, D.C. 40 `Canal's expanded capacity and the economies of scale it will offer to much larger capacity "Post-Panamax" container ships15, once the port's scheduled harbor deepening is completed. Competition with current land-bridge traffic, using the southern Californian San Pedro Bay ports of Los Angeles and Long Beach to drop off cargo that is shipped by rail or truck to south-eastern states, is anticipated: trading off improved vessel economies of scale and intermodal transfer cost savings against the additional ocean miles required per voyage. And still further competition for cargos may also come from the introduction of even larger trans-Atlantic container vessels with the ability carry more than 13,000 TEU's on board, taking advantage of the soon to be operational and much wider Suez Canal (see Section 4.3.3 below). 4.2.2 Quantifying the Monetary Costs of Freight Network Congestion As noted by a number of authors (see the review by Gong et al, 2012),16 putting a monetary value on the cost of delay due to disruptions to the movement of goods is a conceptual as well as technical challenge. In particular, it depends on the importance of on-time deliveries to the customer in question, especially where that customer is going to be using the goods delivered to produce its own products (such as finished automobiles), and upon just how much JIT operations are a key feature of the production process. Based on both a literature review and their own interviews with manufacturing and wholesale sector shippers and, most notably, with receivers of goods in Texas and Wisconsin, Gong et al (2010)17 found that traffic congestion and associated late product delivery resulted in the following, firm-level operational impacts: Additional fuel, oil, and truck operating costs. Extra in-transit inventory holding costs. A large volume of on-site safety stock and high inventory holding costs. Interrupted work flows at unloading bays. A disturbed production schedule and lower productivity. Dissatisfied customers and potential lost sales. Potential loss of the opportunity to consolidate multiple outbound shipments. 15 Vessels transporting from 5,001 13,000 container TEUs. 16 Gong et al (2012) Assessing public benefits and costs of freight transportation projects: measuring shippers' value of delay on the freight system. UTCM Project 11-00-65.CFIRE Project 04-14 Texas Transportation Institute, College Station, Texas. 17 Gong, Q. et al (2012) ibid. 41 Lost business markets and reduced agglomeration economies. Of these cost elements, until quite recently, only those costs listed in the first three bullets have usually found their way into public agency transportation planning studies: with an emphasis on the over the road vehicle operating costs in the first bullet. And also as a result of this partial coverage, the most reliable statistical data comes from a survey of freight carriers, such as the trucker surveys by the American Transportation Research Institute (ATRI: see Section 4.4.1) and the railroads' reporting to the Surface Transportation Board (STB: section 4.4.2). That is, shippers, and especially receivers of goods are rarely surveyed in a way that leads to robust statistical estimation. Until very recently, most state-based long range transportation planning and infrastructure investment studies also paid limited attention to the cargo handling and storage costs associated with on-site activities at either end of freight deliveries or during enroute terminal based transfers of freight between two different modes. This situation has begun to change, moving studies of benefits versus costs of such public investments to consider what is often now referred to as the total logistics costs of managing (e.g. scheduling), handling, moving, and storing of goods. Efficiencies on total logistics costs matter a good deal to US manufacturing industry. According to the Council of Supply Chain Management Professionals 23rd Annual State of Logistics Report annual report (CSCMP, 2013), these costs accounted for 8.5% of US Gross Domestic Product (GDP) by value in 2011 (of which 5.4% comes from `transportation" or moving the goods between places). Significantly, this is down from 10.1% in 2000 and 16.2% of GDP in 1981: showing both the importance of and potential for monetary gains from more effective logistics. A major future barrier to such continued cost savings will be increased levels of network (including over-the- road /rail/ waterway as well as within-terminal) traffic congestion. A 2005 study by Macrosys Research and Technology 18 for FHWA's Freight Management and Operations Office, made use of Council of Supply Chain Management Professionals (CSCMP) other datasets to estimate an approximate economy-wide, all modes breakdown of total business logistics costs into 63% transportation, 43% inventory carrying, and 18 Macrosys (2005) Logistics Costs and U.S. Gross Domestic Product. Report to the Federal Highway Administration, Washington, D.C. http://ops.fhwa.dot.gov/FREIGHT/freight_analysis/econ_methods/lcdp_rep/index.htm#Toc112735360 42 4% administrative logistics costs: with these inventory-carrying costs broken down further into taxes, depreciation, insurance and obsolescence (63%), warehousing (26%) and interest (8%). It remains a challenge to bring these various freight logistics costs into benefit-cost calculations that can in turn inform transportation system investment decisions at the public agency level. A few studies have begun to shed light on this process, however. In NCHRP Report 436, Weisbrod, Vary and Treyz (2001)19 have extended traditional analysis of congestion costs (i.e. extra travel time and vehicle operating costs) to include additional productivity costs associated with travel time variability, worker time availability, freight inventory and logistics/scheduling, just-in-time production processes, and economies of market access: while noting the lack of prior freight costing studies at that time, but seeing a growing concern for congestion's effect on, in particular, JIT business practices. A more recent survey of businesses in Portland, OR (EDR, 2005 20 ) identified the following direct business costs associated with congestion-induced delays: costs for additional drivers and trucks due to longer travel times costly "rescue drivers" to avoid missed deliveries due to unexpected delays loss of productivity due to missed deliveries shift changes to allow earlier production cut off reduced market areas increased inventories and costs for additional crews and decentralized operations to serve the same market area. reduced access to specialized labor and materials. Drawing on information from their own business interviews as well as prior studies, Weisbrod and Fitzroy (2011) describe 26 different elements of business impact and response to traffic congestion growth, grouped into the following seven broad classes, in their effort to understand the economic consequences that "can only be addressed through more detailed microlevel analysis of business processes and business decision-making": market and fleet size 19 Weisbrod, G., Vary, D. and Treyz, G. (2001) Economic Implications of Congestion. NCHRP 436. Transportation Research Board, Washington, D.C. 20 EDR (2005) The Cost of Congestion to the Economy of the Portland Region. Economic Development Research Group . Portland, Oregon. 43 impacts, business and delivery schedules, inventory management, use of intermodal connections, worker travel, business relocation, and localized interactions with other activities. Based on these and other studies, Figure 4.4 summarizes the principal reported impacts of traffic congestion-induced delays on business costs. For the purposes of further analysis, and eventual model development, these costs are grouped into three classes according to the speed with which they typically manifest themselves. Both congestion-induced en route travel time delays and arrival time variability are listed as causal factors in increasing freight movement costs. Arrival time variability here implies less reliable on-time service, raising an important measurement issue discussed further in Section 4.3 below. Note that the freight cost modeling efforts of interest to this current research effort fall under "Immediate Impacts" and "Short Term Impacts", as they apply to an existing manufacturing plant and its current parts supplier and finished goods distribution center (e.g. autorack rail-to-truck transfer terminal) locations. All of the impacts shown in Figure 4.4 can lead to higher per unit (e.g. per finished vehicle) production costs, which if sufficiently damaging may in turn lead to reduced sales with an eventual effect on reduced economies of production, and hence potentially higher transportation and logistics costs. Finally, changes in production costs, may at some point lead ton changes in the volumes and types of vehicle makes and models offered. Conversely, speedier and more reliable transportation service supports the opposite effects, leading to potentially less costly finished vehicles, lower (and hence more competitive) finished product costs, and increases in vehicle-demand induced production. 4.3 Product Flows in the Automotive Industry Supply Chain 4.3.1 Supply Chain Overview This section focuses on identifying the commodity specific components of the case study industry's supply chain, including the movement of commodities both into and out of the manufacturing plant, from parts sourcing to finished product (i.e. automobile) delivery. Figure 4.5 lists the principal products required to construct today's automobiles, organized into six broad component categories. Automobiles are complex machines requiring a wide variety of premanufactured parts, from nuts and bolts to sophisticated electronic devices that have their own assembly issues. In this research, the focus is placed on the movement of parts from what are commonly termed "Tier 1" suppliers: those companies that deliver their finished products 44 directly to the original equipment manufacturer, or OEM (in our case study, to the Kia Motors plant in West Point, Georgia). Figure 4.4 Business Costs of Traffic Congestion En-Route Travel Delays & Arrival Time Variability Immediate Impacts Increased Delivery Costs: - Labor (mainly driver) Costs o owner-operator and/or hourly employee costs - Fuel Costs o fuel, refueling infrastructure - Vehicle O&M Costs o insurance, licenses o tires, parts, lubricants o cargo conditioning costs Short -Term Impacts Increased Logistics Costs: - Late arrival and/or pick -up induced extra labor costs - Sub-optimal production runs - Inventory carrying costs (use of > safety stock margins), - In-transit inventory costs - Capital depreciation costs - Reduced cross-docking options - Missed intermodal connections Longer -Term Impacts Increased Business Costs - Distribution center relocation - Production site relocation - Longer employee commutes o lower employee retention and recruitment rates Changes in Size and/or Type of Vehicles in Fleet and/or in Vehicle Trip Frequencies Higher Per Unit Production Costs Lost Business Opportunities - Viable Market Area Contraction Reduced economies of scale in production and/ or logistics practices Figure 4.6 shows the major generic freight delivery steps associated with supplying these components. The present research looked at the connections between the activities in the colored boxes, that is, at Tier 1 supplier to OEM and OEM to retail dealer steps. An effort was made to find out what types, volumes, transport modes and shipment distances are involved in moving these various automobile parts, as well how finished automobiles are moved from the OEM to auto dealerships both within and outside the south-eastern region. The transport of replacement parts and the fate of used vehicles and parts were not pursued in this present research effort. 45 Figure 4.5 Principal Components of Automobile Manufacturing Supply Chains Automotive Components A. Axles & Brakes B. Electrical, Electronic & Cooling C. Engine & Transmission D. Suspension, E. Interior F. Exterior Steering & Other & Hydraulics A1. Axles & Components 1.Axles/differentials /transfer cases 2.Bearings 3. CV & u-joints 4. Drive shafts 5. Torsion traction sys. 6. Viscous couplings A2. Brakes & Components 1. ABS components 2. Master cylinders, calipers 3. Pads, shoes 4. Rotors, drums 5. Wheel cylinders, hoses, tubing B2. Electronic Systems & Components 1. Connectors 2.Engine management sys. 3.Optical cable, multiplexing 4 Printed circuit boards 5.Semiconductors, diodes , transistors B3. Cooling Systems & Compnents 1.Fans,clutches 2.Heat exchanges 3.Hoses,belts 4.Radiators 5.Thermostats B1. Electrical Systems & Components 1. Alternators, generators 2. Anti-Theft sys & comps 3. Auto sys & comps 4. Batteries & parts 5. Collision warning systems 6. Switches, fuses, circuit breakers 7. Fuel sys & comps 8. Heating, ventilation, A/C & comps. 9 . Horms,alarms,emergency equip. 10. Ignition sys & comps 11. Instrument clusters & comps 12. Lighting sys & comps 13. Motors & comps. 14. On board radar systems 15. Relays & regulators 16. Sensors & actuators 17. Solenoids 18. Starters 19. Wiring 20. Cruise control C1. Engine & Components 1. Blocks, heads 2. Camshafts, crankshafts 3. Connecting rods 4.Cylinder liners 5. Diesel Engines 6. Emissions equipment 7. Engine bearings 8. Exhaust components 9. Filters (air,fuel,oil) 10. Fuel additives 11. Fuel Sys. And comps 12. Gaskets,seals,packings 13. Gaslolne engines 14. Intake components 15. Intercoolers 16. Pistons & rings 17. Pumps,tubing,hoses,fittings 18. Timing chains,gears & belts 19. Turbo & superchargers 20. Valve covers, oil pans 21. Valvetrain & comps. C2. Transmission & Components 1. Clutches,valves & comps 2. Gears & linkages 3. Housings 4.Manual and automatic transmissions 5. Torque ocnverters 6.Transaxles 7. Transfer cases 8. Transmission bearings D1. Suspension & Components 1. Brushings & bearings 2.Castings/forgings/stampings 3. Dampers 4. Springs 5. Tires 6. Wheels D2. Steering & Components 1. Linkage, hoses, boots 2. Pumps 3. Steering columns 4. Steering grears 5. Steering stacks D3. Hydraulic & Pneumatic Systems 1. Aircompressors 2.Hydraulic cylinders 3. Pumps (nonsteering) 4. Tubing,hoses,fittings 5.Valves & controls E1. Interior 1. Airbags & components 2. Cables 3. Carpeting/floor mats 4. Door systems & trim 5. Headliners 6. Intrument panels, consoles 7. Interior trim 8. Linkages 9. Mirrors 10. Seat belts 11. Seats & components 12. Window systems F1. Exterior 1. Body parts 2. Bumpers & parts 3. Exterior trim 4. Lighting 5. Locks,latches,hinges 6. Mirrors 7. Stampings 8. Sunroofs/convertible tops 9. Wiper blades & arms F2. Fasteners & Adhesives 1.Adhesives 2. Clamps 3. Mechanical fateners 4. Tape F3.Others 46 Figure 4.6 Transportation Links in an Automotive Industry Supply Chain (Generic) Foreign Parts Suppliers Parts Domestic Parts Suppliers Manufacturing Plant (OEM) Parts Foreign (Export) Markets Vehicles Mixing Centers (Rail) Vehicle Distribution Centers Parts Distribution Centers Parts Local Dealerships Re-Sale/ Re-Use Used Vehicles Scrap 4.3.2 Local and Regional Components Suppliers and Their Shipments Many components are delivered to the KMMG plant from local and US based suppliers. These local area suppliers have in this instance come into being for the purposes of supplying parts to the KMMG plant (cf. Table 1.1 above). Unfortunately, much of the information on exactly what and how many parts are supplied to the KMMG plant is private and we do not have any database that tracks these flows. Table A.1 in Appendix A contains a list of these 117 component suppliers of KMMG West Point assembly plant (25 in Georgia, 92 in Alabama) with company names, supplying components and location information, based on combining data from a number of different sources (see notes below table). Figure 4.7 shows the result of translating these addresses into longitudes/latitudes for mapping and highway trip routing purposes, showing the locations of auto parts suppliers in Georgia and Alabama along with the locations of the Kia and Hyundai auto plants. (A version of this file containing the locations of suppliers outside Georgia and Alabama was also created, based on the locations of suppliers in the foreign imports dataset). 47 Figure 4.7 Map of Georgia and Alabama Automotive Parts (Component) Supplier Locations 150 mile radius Table 4.1 lists the data elements in the geo-coded supplier locations database. Table 4.1 Automobile Parts (Component) Supplier Locations Data File ID Long Lat Hnode SIC Emp Tier Zip County Product Class Firm Address OEM Fnode State City Record ID # Longitude Latitude Highway Node # Standard Industrial Classification Code (4-digit) Number of Employees Placeholder (only used to identify OEM locations currently) 5-digit-zip code US County Name SIC Description Company Name Company Address OEM served, if known Other Network Node (NOT CURENTLY USED) 2-Digit State Abbreviation in US City Name 48 4.3.3 Foreign Supplier Shipments US Imports Data Kia being part of Hyundai, and a major multinational company, has deep ties to component suppliers from the Asia-Pacific region. Using US customs related databases, we are able to track these shipments. A number of commercial data sources provide data on the foreign imports of vehicle parts. Upon review, the Panjiva (2013)21 dataset was selected for project use. Shipment data from Panjiva.com are used to analyze the origin and destination port of the shipments coming to the KMMG plant. By searching final destination of Kia Georgia plant, we found 4,253 shipment records in Panjiva for the period 7/7/2008 to 4/25/2013. Each shipment record also contains the type, weight and number of items of commodity shipped, the name and location of the shipper, as well as the foreign port of lading (loading), the US domestic port of unlading, and the final US port of destination. To allow mapping, each of these locations was geo-coded and the results combined to produce the data elements shown in Table 4.2. In the aggregate, Figure 4.8 shows the growth in imports at each unlading port as a percentage of each year's contribution to that port's total mid-2008 through 2012 import totals, based on both annual number of shipments and annual tonnage shipped. The results are shown for the dominant US ports of unlading, i.e. for Savannah, GA and a combined total for the San Pedro Bay ports of Los Angeles and Long Beach, CA; and also (the green bars) for all US ports of lading in the dataset. The final destination of all shipments is the Kia auto plant. Between them, Savannah and Los Angeles/Long Beach accounted for some 98% of all imports during this start-up period, with Savannah accounting for over 60% by tonnage and 50% by number of individually reported shipments (of various sizes and counts). What Figure 4.8 shows is the increased activity levels through Savannah post-2010, when the port is estimated to have accounted for 66% of all imported parts by tonnage (estimated at some 186,000 metric tons from July 2008 through April 2013) and over 55% (over 2,000 bills of lading) by shipment count. Much smaller volumes also entered the US over this four and a half year period including shipments via the ports of Mobile AL, Charleston SC, Jacksonville and Port Everglades FL, New York NY and New Jersey NJ, and Seattle and Tacoma, WA. Most of these imports come from South Korea, via the port of Busan (aka Pusan), with over 82% of shipments by count and over 92% of shipments by weight associated with this foreign port of lading. Other ports of lading include 8 ports in China (including Hong Kong), 3 each in Japan and Panama, 2 each in Germany and Vietnam, and one in France, Guatemala, 21 Panjiva (2013) Investigate Companies, Shipments, and Trade Trends http://panjiva.com/ 49 Jamaica, Malaysia, the Netherlands, Singapore, South Korea and Taiwan during the start-up period (see Figure 4.9). Table 4.2 Geocoded Foreign Import Shipments Data File 50 Percent of 2008-2012 Growth Figure 4.8 Rate of Growth in Imported Auto Parts from 2008-2012 by Selected US Ports of Unlading 60 50 Growth Based on Kilograms of Freight 40 30 Savannah Los Angeles & Long Beach 20 All US Ports of Unlading 10 0 2008 2009 2010 2011 2012 60 Growth Based on 50 Number of Shipments Percent of 2008-2012 Growth 40 Savannah 30 Los Angeles & Long Beach 20 All US Ports of Unlading 10 0 2008 2009 2010 2011 2012 A major gap in the data on freight movements of all kinds within the United States is our limited knowledge of how US imports move inland, once they have been delivered at a US seaport. Perhaps surprisingly, given the considerable and growing importance of foreign imports into the United States, no government data source or combination of available sources exists from which to extract this modal information. 22 While all auto parts currently arrive at the plant by truck, an unknown percentage of the foreign and longer distance within-US cargos may travel overland by rail. 22 Detailed railcar waybills data could tell us a good deal about how much non-container freight moved from ports such as Savannah and Los Angeles/Long Beach to intermodal terminals in close proximity to the Kia Motors plant: but this data was not available to the present study. Containerized data is more difficult to track, while offloading at intermodal truckrail terminals designed for the purpose. Even the federal government's Freight Analysis Framework (FAF) database must estimate these modal percentages. 51 Railroad-operated on-line search sites such as CSXT's ShipCSX 23 and Norfolk Southern (NS) Railroad's Intermodal 24 services sites provide details of intermodal container train schedules originating or terminating in close proximity to the Kia Motors plant: notably the Fairburn CSX intermodal truck-rail terminal a few miles north-east of the Kia Motors plant along I-85 in Fulton County, Atlanta (CSXT also operates the Hulsey intermodal container terminal within Atlanta); and the NS-operated Austell intermodal terminal in Cobb County in north-west Atlanta. Figure 4.9 Example Trans-Pacific Parts Shipment Routes. 4.3.4 Transportation of Finished Automobiles (to Customers/Dealerships) Production and sales data can be obtained from the KIA Motors website, www.kmcir.com. On the website, Kia gives detailed monthly production by model and sales by country statistics of all of its factories around the world. The KMMG data were available starting January, 2010. 23 http://shipcsx.com/public/ec.shipcsxpublic/Main?module=public.ischedule 24 http://www.nscorp.com/content/nscorp/en/ship-with-norfolk-southern/shipping-options/intermodal/terminals- and-schedules.html 52 Kia's annual production and sales, and by model type data are presented in Table 4.3 and Table 4.4, respectively. The KMMG plant began operations with two shifts and about 250,000 vehicles per year. With increase in market demand, especially for Kia Optima, they increased to three shifts and reached the designed maximum capacity of 360,000 vehicles per year. From the sales data, we can see that most of the sales are for the US market with some Kia Sorento being sold in Canada, and a few Kia Optima and Sorento being sold in Latin America. Table 4.3 KMMG Production Statistics 2010 Optima/K5 Sorento 138,071 Kia Total 138,071 Santa Fe 29,051 Total 167,122 Notes: Data are from KIA Motors. 2011 35,132 146,017 181,149 91,155 272,304 2012 128,536 131,572 260,108 98,091 358,199 2013 133,946 129,590 263,536 105,969 369,505 2014(Jan.-Apr.) 49,703 44,687 94,390 35,885 130,275 Table 4.4 KMMG Sales Statistics Optima/K5 US Latin America Sorento US Canada Latin America Total Notes: Data from KIA Motors. 2010 108,202 10,207 710 119,119 2011 21,505 225 130,235 15,105 1,194 168,264 2012 126,797 344 119,597 14,031 1,512 262,281 2013 124,598 340 105,649 14,542 1,295 246,424 2014(Jan.-Apr.) 42,945 88 31,542 3,845 241 78,661 In addition to a discussion with Kia staff, a number of sources were searched in order to both understand and quantify to some degree the level of transportation activity involved in getting finished vehicles from the KIA manufacturing plant to its dealers. While local, including some quite long intra-regional transport of finished cars takes place over the highway (an essential mode for short term order fulfillment), between 40% and 60% of the vehicles Kia produces are now transported from the plant to other parts of the US and Canada by rail. Kia opened its railcar loading facility early 2010 with an initial 36 railcar capacity, which has since been expanded to handle some 90 multilevel autorack railcars25, with plans for further expansion from an initial 80-railcar to 400- railcar 25 Bi-level autorack railcars may carry up to 15 vehicles, while some of the largest autoracks in use may carry as many as 20 or 22 vehicles. For example (only), at 60 carloads x 15 vehicles per car x 250 weekdays per year would = 225,000 vehicles shipped per year. 53 holding capacity.26 By 2014, it is estimated that up to 60 railcars per day will leave the West Point plant for delivery to Kia dealerships. CSX Transportation provided the project with some general information on its railcarloadings with agreement not to report specific volumes. The data show deliveries to Canada, the MidAtlantic, Midwest, North-Eastern, South-Western and West Coast states, and smaller but still significant shipment volumes to the Pacific North West and Mountain West regions of the country. In contrast, South-Eastern states are served largely by truck. Trains heading West need to interline with western- serving railroads at locations such as Birmingham, Memphis, Chicago and Toledo. Kia's westbound vehicles are shipped using a combination of CSX and BNSF Railway, and CSX and Union Pacific Railroad. An on-line 2010 news article in Automotive Supply Chain reported finished vehicles from Kia's West Point manufacturing plant moving on bi-level railcars, each holding ten vehicles, for a total loading capacity at a single time of 1,500 vehicles. 27 Some 1,100 spaces were also reported to be available (in 2010) for plant-side truck pickups, using 10-car auto-transporters to distribute Kia vehicles around the south-east. 28 Both truck and rail options are considered for delivering vehicles to locations in the states of Maryland, West Virginia, Ohio, Indiana, Illinois, Missouri, Kansas, and New Mexico. This same article also reported these railcars being loaded either the same day or the day after the vehicles' release from the plant, with loaded railcars prepared several times a day for CSX nighttime pickups. Information technology is then applied to let truckers at the other end of a rail haul know when the railcars are due to arrive at their rail destination (i.e. offloading) ramps. The location of CSX Corporation (via its Total Distribution Services Inc. subsidiary) automobile- distribution facilities and storage locations, overlaid on a mapping of the CSX rail network can be found on-line at: http://www.csx.com/index.cfm/customers/other-services-partners/tdsi/mapslocations/ Kia Car Dealership locations (addresses) can be found on-line at: http://find.mapmuse.com/brand/hyundai-dealers (with on-line mapping) and http://www.edmunds.com/dealerships/Kia/ 26 http://www.transdevelopment.com/?p=243# 27 http://www.automotivesupplychain.org/features/3/71/news/ 28 Up to 10 vehicles may be carried at one time, and it was reported that some carriers may also be moving Hyundai produced vehicles from the Montgomery, Alabama plant if this is a cost-effective use of the trucker's resource, with Kia (via its Glovis America logistics provider) paying for transport on a per-vehicle basis. 54 4.4 Estimation of Modal and Intermodal Transportation Costs Freight rates are shipment distance and travel time sensitive, and require a sufficiently detailed and accurate source of such data for each of the transportation modes used within both automobile component and finished product links in the supply chain. While published freight rates exist and quotes are often accessible on-line, statistically reliable rate data is in general either hard to come by or expensive to purchase. And for forecasting purposes, it is also important to understand the elements of cost that go into creating such rates, both now and in the future. With this in mind, and given the project's ultimate interest in generating data for use in public sector planning, it was decided to assess the ability of current data sources to produce estimates of such shipment cost elements directly. The focus in the present case study is therefore on estimating truck, rail, and oceanvessel transportation costs (air freight was not considered due to its comparatively limited current use, based on information obtained during OEM interview). The following paragraphs summarize the latest literature on freight cost estimation methods, taking one mode at a time. Section 4.5 then describes how these various mode specific cost formulas can be used, along with a method for estimating shipment distances and associated origin-todestination travel times, in order to compute dollar based fuel, labor, and other per hour or per mile vehicle or vessel operating, maintenance and cargo handling costs: including the cargo handling costs associated with inter-modal transfers. 4.4.1 Trucking (Highway) Costs A recent TRB sponsored review of the needs for and availability of specific freight cost data elements for public agency decision-making concluded (Holguin-Veras et al, 2013, page 56):29 "In the trucking sector, a large assortment of data sources provide bits and pieces of data of various degrees of usefulness and quality, but fail to provide a comprehensive and coherent picture." Elements of Trucking Costs A number of truck operating cost models and supporting software tools have been developed in recent years, usually with a bias towards a specific trucking sector or industrial sector operations. Among the more adaptable models are those by Berwick and Farooq's (2003) Truck Load Analysis 29 Holgun-Veras, J. et al (2013) Freight Data Cost Elements. NCFRP Report 22. Transportation Research Board, Washington, D.C. 55 Model software30, the Owner-Operator Independent Drivers Association's (OOIDA's) per mile truck cost of operation calculator31, and the cost formulas developed by Hussein and Petering (2009)32 who also provide a useful review of some past studies. The American Transportation Research Institute (ATRI)33 also puts out an annual update of truck operating costs, based on survey responses from operators of over 40,000 truckload, less-than-truckload and specialized service trucks. Marginal expenses for motor carriers were divided into vehicle- and driver-based costs. These costs include average costs estimates for each of the following items: Vehicle-based costs: o Fuel and engine oil o Truck/trailer lease or purchase payments o Repair and maintenance o Truck insurance premiums o Tires o Permits o Tolls Driver-based costs: o Wages o Benefits Hussein and Petering (2009)34 use a similar set of cost elements, which they term respectively fuel, labor, depreciation (straight-line depreciation less capital recovery), maintenance, loading and unloading, insurance (of both truck and cargo), overhead (including management and administration staff, property taxes, utilities, advertising, communication equipment, rental of facilities, insurance of facilities, etc)., and extra (highway user and licensing fees and additional costs for transporting hazardous cargo). They provide detailed formulas for each of these cost elements. Also of note, one of their numerical examples is the computation of an auto parts shipment. What this literature makes clear is the considerable number of variables that can affect trucking costs for any given trip or shipment distance. And some of these variables have a 30 Berwick, M., & Farooq, M. (2003). Truck costing model for transportation managers. Upper Great Plains Transportation Institute, North Dakota State University, Fargo, ND. 31 http://www.ooida.com/EducationTools/Tools/costpermile.asp 32 Hussein, M.I. and Petering, M. E.H. (2009) A policy-oriented cost model for shipping commodities by truck. CFIRE Paper No. 94-4. University of Wisconsin, Madison. 33 http://atri-online.org/ 34 Hussein, M., & Petering, M. (2009). Ibid. 56 considerable influence on the final dollar cost of per cargo unit or per vehicle movement. The latest, 2012-based estimates from ATRI (Fender and Pierce, 2013)35 report average carrier costs of just over $1.63 per mile, and $65.3 per hour: with different rates also reported for truckload (TL), less-thantruckload (LTL) and specialized trucking services as well as by region of the country. Of note, fuel costs accounted for some 39% of total operating costs in 2012, with driver costs at 33% of the total. Operating costs are, however, known to vary a good deal in practice, depending on vehicle configuration and age, type of delivery service, and trip distance, among other factors. Hussein and Petering (2009) and Wheeler (2010)36 provide recent reviews, reporting a rather large range of values: from as low as $10 per vehicle operating hour, to over $190 per hour associated with time delays due to traffic congestion. What these studies also show, however, is a generally consistent treatment of the major cost elements involved, which we can summarize here as fuel, labor, operation and maintenance (with or without a vehicle depreciation and other indirect costs, depending on application) and cargo handling (= loading/unloading) costs. To these, we can also add a fifth cost element, storage costs. These can occur whenever a pickup or delivery is late, and in some cases can be quite expensive if special cargo storage conditions (e.g. refrigeration) are required. These delays can occur at both ends of a trip, and however the cargo delivery payments occur between producers, consumers and carriers, any time that is lost due to significant variability in on-time arrivals can prove costly. Pulling these five different cost elements together (i.e. fuel, labor, O&M, loading and storage costs), Figure 4.10 shows how each is a function of not only the type of commodity and type of truck used, but also of an O-D trip's over the highway travel distance and time. Of note, travel speeds can be seen to play a key role in total OD trip costs. In addition to direct impacts of such speeds on trip travel times, and hence labor (driver) costs, speed also impacts fuel consumption rate and therefore fuel costs. Particularly costly are in-transit delays that result in late cargo pickup or delivery where these result in extra cargo storage costs as well as possible additional labor costs associated with rescheduled cargo loading/unloading activity at a trip's origin or destination. These linkages are labeled as Delays in Cargo Pickup/Delivery in Figure 4.10. These include the impacts on in transit speeds 35 Fender and Pierce (2013) An Analysis of The Operational Costs of Trucking: A 2013 Update. American Transportation Research Institute, Alexandria, VA. September, 2013. 36 Wheeler, N.M. (2010) Multi-Criteria Trucking Freeway Performance Measure for Congested Corridors. Master of Science Thesis, Civil and Environmental Engineering, Portland State University, Portland, OR. 57 caused by freight movement disruption events such as major highway accidents, severe weather, road damage and closures, and severe traffic congestion due to mixed passenger and freight traffic volumes exceeding roadway design capacities (see below). The greater the frequency of such events and the lower the percentage of on-time cargo deliveries, the more expensive are the delays incurred. And depending on the type of contract between shipper, receiver and carrier, one or more of these may need to absorb these additional costs, which can be expected, one way or another, to find their way back to the final consumer if they occur with sufficient frequency. Figure 4.10 Elements of Truck-Trip Based Freight Transportation and Logistics Costs Vehicle Type Average Vehicle O M Cost ( /Mile) Average Labor Cost (Driving) ( /Hour or /Mile) Average Fuel Consumption (Gallons/Mile) Vehicle O M Per Trip Cost ( ) Per Vehicle Trip Labor Cost ( ) Trip Travel Time (Minutes) Trip Travel Speed (mph) Average Cost Per Vehicle Trip Commodity Class Average Load (Tons or Units /Vehicle) Fuel Cost Per Trip ( ) Delays in Cargo Pickup/Delivery Average Loading /Unloading Cost ( /Vehicle Load) Storage Handling Cost Per Trip ( ) Average Storage Cost ( /Load/Hour) Delays in Cargo Pickup/Delivery From Assignment Algorithm Vehicle O M cost includes costs associated with tires, oil, parts maintenance and replacement, insurance and licenses. 58 Measuring the Costs of Poor On-Time Reliability Recognition of the importance of this on-time reliability issue to freight shippers and receivers has led to a significant effort by the US DOT and the Transportation Research Board (TRB), among others, to quantify both the magnitude and costs associated with such travel time uncertainty. In particular, TRB's Strategic Highway Research Program 2 (SHRP2) has devoted a good deal of effort to measuring what it terms travel time reliability measures for the purposes of planning and programming studies.37 A literature review by De Maeyer and Pauwels (2003)38 on the role of quality of service attributes and their monetary valuation, as derived by a number of different freight demand models, confirms the importance of service reliability to mode selection, often placed ahead of the value of shipment cost itself. Fowkes et al's (2004) stated preference interviews of 40 shippers, carriers and third party logistics operators (3PLs) identified similar concerns over on-time delivery reliability, especially where JIT deliveries were concerned; placing a higher value on reliability and predictability in delivery times was emphasized as the most important transportation service attribute by industry respondents, more so than minimizing average lead times. 39 However, while a good deal has been written over the past five years about how to measure such reliability for planning purposes, including the use of such measures in the Highway Capacity manual, 40 comparatively little research has been published on how to assign monetary values, or the implied costs, to such reliability measures. Where freight movements are concerned, reliability is recognized as one of the principal variables that affects the choice of both mode and shipment size. Reliability is valued by FHWA's Intermodal Transportation and Inventory Cost Model State Tool (ITIC-ST)41 by measuring the effects of variability in the shipment ordering lead-time. Lead-time here includes the time required for the shipper to receive the order from the customer, to pick the order from his inventory, to arrange for transportation, to wait for a vehicle to arrive at the shipping dock, load the shipment, and finally to move the freight from the shipping point to the customer's destination. The more reliable the 37 http://www.trb.org/StrategicHighwayResearchProgram2SHRP2/SHRP2ResearchReports.aspx 38 De Maeyer, J. and Pauwels, T (2003) Modal choice modelling: a literature review on the role of Quality of Service attributes and their monetary valuation in freight demand models. Department of Transport and Regional Economics, University of Antwerp, Belgium. 39 Fowkes, A.S. et al (2004) How highly does the freight transport industry value journey time reliability--and for what reasons? International Journal of Logistics: Research and Applications 7(1):33-43. 40 Kittelson, P.E. and Vandehey. M., et al (2013) Incorporation of travel time reliability into the HCM. SHRP 2 Reliability Project L08. Transportation Research Board, Washington D.C. 41 Federal Highway Administration (2011). Intermodal Transportation and Inventory Cost Model State Tool. https://www.fhwa.dot.gov/policy/otps/061012/iticst_info.htm 59 delivery schedule, the less safety stock the customer must keep on hand in order to ensure that a stock-out (i.e. the customer runs out of stock) does not occur, noting that safety stock "is typically a larger component of total logistics cost than many of the other costs (with the possible exception of transportation charges) because it must be carried continuously". The method ITIC-ST uses to capture this (un)reliability of on-time service effect is a two parameter Gamma function based on the mean and the standard deviation of the shipment transit time. Treating the mean re-ordering lead-time as equal to the mean in-transit time, the model user inputs a coefficient of variation (COV) equal to the standard deviation of in-transit time for a given mode (truck or intermodal rail in this case) divided by this mean in-transit time. For further technical details, see FHWA (2006).42 Default model parameters are provided. In considering how to quantify travel time variability costs incurred on sections of US freeways (by general traffic), notably as a result of non-recurring incidents, Cohen and Southworth (1999)43 present two different approaches to assigning a user benefit (cost) to more (less) reliable travel times. In the first approach, an additional cost of travel is assigned directly to a measure of trip time variability, i.e. C = a1*T + a2*Var(T) + a3*M (4.1) where C equals the expected cost of a trip, and a1, a2, and a3 are parameters that reflect travelers' relative dislike of, respectively, trip time T, a measure of trip time variability Var(T) (in practice, the standard deviation, SD, was again used), and a monetary travel cost, M. The ratio of (a2/a1) provides a useful measure of the relative importance of changes in travel time variability versus changes in total trip time. The ratio of (a2/a3), often termed a reliability ratio, allows a monetary cost to be assigned to the importance of such variability, i.e. Reliability Ratio (Travel Time) = Value of SD of Travel Time / Value of Travel Time (4.2) With a variation on this idea devoted to late arrivals: Reliability Ratio (Lateness) = Value of SD of Lateness / Value of Lateness (4.3) 42 FHWA (2006) ITIC-ST Version 1.0. Intermodal Transportation and Inventory Cost Model. A Tool for States. Technical Documentation. Federal Highway Administration, Washington, D.C. 43 Cohen, H. and Southworth, F. (1999) On the measurement and valuation of travel time variability due to incidents on freeways. Journal of Transportation and Statistics 2.2. 123-131. 60 where "SD' in both equations (4.2) and (4.3) is short for the standard deviation. A partial review of past studies, in both passenger and freight movement, is provided in a draft 2012 SHRP2 workshop report by Cambridge Systematics and ICF.44 Of the seven freight value of travel time reliability studies cited, none are from North America. Most are from Europe. An earlier review of work on the topic of measuring travel time reliability in the United States, the European Union, and elsewhere, by Grant-Muller and Laird (2006)45 notes that: "At this point in time there is still uncertainty as to what the value of reliability is for both personal and freight related travel. However, there can be no doubt, given the qualitative and increasing quantitative evidence, that these values can be significant and large." This situation still applies today: although a recent US study by Gong et al (2012)46 does begin to shed some light on this topic. Recognizing the difficulty of the task, these authors tried three different approaches to determining the value of fright shipment delay to shippers and receivers. Using a small number of in-depth interviews, along with a larger survey of manufacturers and wholesalers in Texas and Wisconsin, they also develop an analytic approach based on inventory management theory. Using an Analytic Hierarch Process (AHP) and Willingness To Pay approach to their survey instrument, they produce a number of example estimates of the cost of congestioninduced delay. Taken over their entire survey, they suggest a value of 56 per hour for shippers' travel time. They also compute a travel time reliability cost of $0.40 per each percentage of additional delay, where such a percentage represents the hypothetical delay time divided by the normal (average, expected) travel time for a given trip. Based on a series of experiments based on different inventory stock-out policies, different order lead times, and different demand frequency profiles, they found that freight receivers are more likely to find increases in the variability of O-D trip times to be more costly than increases in the means of such O-D travel times: sometimes costing more than twice as much in trucking plus in-transit and warehouse inventory carrying costs. (And of 44 Cambridge Systematics and ICF (2012) Value of travel time reliability. SHRP-2 Workshop Working Paper and Synthesis Report. April 2012. DRAFT. 45 Grant-Muller, S. and Laird, J. (2006) Costs of congestion: literature based review of methodologies and analytical approaches. Report Prepared by the Institute for Transport Studies, University of Leeds, England for the Scottish Executive Social Research agency. http://www.scotland.gov.uk/Publications/2006/11/01103351/0 46 Gong et al (2012 ) Assessing public benefits and costs of freight transportation projects: measuring shippers' value of delay on the freight system. UTCM Project 11-00-65.CFIRE Project 04-14 Texas Transportation Institute, College Station, Texas. 61 note, the values they obtained from their test respondents recorded higher than average delay valuations for those involved in moving automobile parts). An additional wrinkle to Figure 4.10 is the way in which labor costs are determined. Specifically, truck drivers may be paid by the hour or by the mile. If by the latter, then speed of travel will be an especially important factor in a trucking company's revenues (subject to hours of service and other regulations and practices meant to ensure safe driving: see ATRI, 201347). It also again leads to a trade-off between added fuel consumption and time saved. Based on the above literature, the following per vehicle-trip cost formula suggests itself for use in planning studies, allowing the various right-hand-side costs to vary both by commodity carried and vehicle class: Truck Trip Cost = Fuel Cost + Labor (Vehicle) Cost + O&M Cost+ Cargo Handling Costs (4.4) and where cargo handling includes both cargo loading/unloading costs as well as any factory, warehouse or terminal based in-transit storage costs involved, including traffic congestion-induced delay costs.48 An additional term, possibly in the form of a multiplication factor, can also be added as an approximation to the effects of on-time service (un)reliability on overall per trip dollar costs. Finally, truck trip rates should also include a profit margin, for which limited direct information exists in the public domain. 4.4.2 Rail Costs While there are fewer studies of rail cost models and their data sources in the open literature than there are for trucking (Holguin-Veras et al, 2013), two publicly available cost estimation models exist: the Uniform Rail Costing System (URCS)49 and the Intermodal Transportation and Inventory Cost Model State Tool (ITIC-ST) 50 . Of these, Surface Transportation Board's (STB) URCS software offers the most generally useful, annually updated and at the same time, statistically robust railcar costing option. URCS uses average wage rates and other data from its annually collected, 47 Ibid. 48 Loss of cargo value due to within-truck transit delays is considered here to be comparatively small compared to time costs due to extended on-site storage at shipment origin, destination or during intermodal terminal transfer. See Gong et al (2012 ) ibid. 49 http://www.stb.dot.gov/stb/industry/urcs.html 50 Ibid. 62 nationwide sample of railcar waybills to calculate the cost of shipping commodities by a specific railroad and origin-to-destination distance. These freight rates take into account the type of freight and being shipped (e.g. automobiles, or Standard Transportation Commodity Code 37111), the type of train (single or multiple railcar or unit train) and number and type of railcars in the train, whether or not these railcars are railroad or privately owned, and what type of backhauls (returning of empty or loaded railcars) will take place. Figure 4.11 Example Initial Input Screen for Rail Costing Program Figure 4.11 shows the initial data input screen for the URCS when applied to a specific shipment. Default settings can also be used to fix the circuity of the route in which the railcars will travel, the unloaded (tare) weight of the specific type of railcar (e.g. a multi-level flatcar) and a general overhead ratio which allocates administrative and other indirect expenses to variable car-mile and car-day costs for the specific railroad service in question. Using the URCS software, it is 63 possible to estimate the cost of shipping not only parts but also finished automobiles, the latter by multi-level autorack trains. Table 4.5 and Figures 4.12 and 4.13 provide example calculations for four different O-to-D rail distances from the OEM's railcar loading site to destinations whose network mileages are computable using the railroad specific sub-network contained in the ORNL multimodal freight transportation network database (see section 4.5.2 below). In this project, this means using the CSXT railroad sub-network out of the KIA West Point plant. While we DO NOT know the rates charged per vehicle transported from this plant, the URCS software does allow us to compute an approximate transportation rate and cost per destination city, and to do so for a range of cost-impacting factors. The number of railcars per train is set here at 36. Other input parameters are 20 autos per railcar, a weight of 1.75 tons per automobile (for example, the Kia Optima has a curb weight of around 3,200 lbs and the Kia Sorento SUV a curb weight around 3,800 lbs).51 Table 4.5 Privately-Owned Railcars Cost Breakdown by Selected O-D Rail Distances Privately-Owned Railcars 265 miles 514 miles 750 miles 900 miles Cost per shipment 45600 71669 88835 99726 Cost per auto moved 63.3 99.5 123.4 138.5 Cost per auto moved-mile 0.239 0.194 0.165 0.154 Figure 4.12 Results from Rail Costing Scenario Using Railroad-Owned Railcars 51 http://www.edmunds.com/kia/optima/2013/features-specs.html and http://www.edmunds.com/kia/sorento/2013/features-specs.html?style=101424381 64 Figure 4.13 Results from Rail Costing Scenario Using Privately-Owned Railcars Figure 4.14 shows an example of the effect on cost per mile per automobile shipped by moving from a 36 railcar train to a unit train made up of 70 railcars, again for both privately owned ("priv") versus railroad owned ("rr") railcars. Figure 4.14 Effect of Number of Railcars Per Train on Shipment Rate per Automobile (for a shipment distance of 514 miles) $/AUTO MOVE MILE 0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0 rr/multiple/36 rr/unit/70 priv/multiple/36 rr/unit/70 65 4.4.3 Trans-Ocean Shipping Costs With a good deal of auto-parts movement into the US auto manufacturing plants from both Eastern Asia and also Northern Europe, global shipping rates are of considerable interest to cost effective supply chains: for both parts supply and also for finished vehicles. The present project's focus is on the former, since most of the finished vehicles of interest are currently bound for US markets, with some deliveries also into Eastern Canada. Most of these parts are now transported by container ships. With nearly all parts now shipped in steel containers, and with more than 50% of all US seaborne trade by value now moved in containers (UNCTAD , 2013 52 ), data on container transport costs are key inputs to supply chain assessments. Gkonis, et al (2009)53 provide a review of liner shipping costs. Psaraftis (2009)54 describes the following container shipping cost model: Container Shipping Cost = Fuel Cost +Time Charter Costs + Cargo Inventory Costs (4.5) These cost elements are very similar to those reported as equation (4.4) above, with fuel costs for the trans-oceanic portion of a given O-D cargo shipment estimated by the following formula (for example, after Psaraftis, 2009): Fuel Cost = Fuel Price/Gallon* (O-D Distance/V) *(a + b*Vn)*Delta (4.6) where V = average ship speed in nautical miles/day Delta = ship's displacement or loaded ("laden") weight and a, b, n and are empirically derived model parameters. Charter costs are proportional to O-D travel time, while cargo inventory costs are similar to the O&M costs used in the trucking formula. Per unit volume or unit time cargo delay costs here include two elements, or cargo delay rates: one for cargo waiting to be picked up at a port, and a second cost associated with the time a cargo remains within the ship before off-loading. 52 UNCTAD (2013). Review of Maritime Transport 2013. United Nations Conference of Trade and Development, New York and Geneva. 53 Gkonis, K.G. et al (2009) Liner shipping costs and logistics: a literature survey and taxonomy of Problems. Laboratory for Marine Transport, National Technical University of Athens. Greece. 54 Psaraftis, H. (2009) A ship pickup and delivery model with multiple commodities, variable speeds, cargo inventory costs and freight rates. Laboratory for Marine Transport, National Technical University of Athens. Greece. 66 Two useful sources of shipping rates data, including time charter costs, are the following: i) In addition to selling shipping rate information, Searates.com55 provides measures of shipment distances and transit times between major seaports around the world, reported in days and hours. For example, a trans-Pacific vessel transit between the Port of Busan, South Korea and the Port of Los Angeles in southern California represents a roughly 6,100 mile trip, taking some 14.7 days. Continuing overland from Los Angeles to Atlanta, GA increases the trip to around 8,300 miles and 19 days in duration. In comparison, a shipment from Busan to Atlanta via the Panama Canal and the US Port of Savannah covers over 11,300 combined sea plus land miles, with an expected 29 days to complete. ii) VHSS (The Hamburg Shipbrokers' Association)56 reports representative charter shipping rates based on its connections to a Germany shipping industry that owns over half of the container ships on the sea: through its New ConTex and its Hamburg Indices (Containership Time-Charter-Rates) for a range of vessel sizes (based on the twenty-foot equivalent unit, or TEU carrying capacity). Example container shipping rates to the USA are also published by the World Bank for the years 1980 through 2013.57 As with trucking costs, vehicle/vessel speed plays an important role in all three major cost elements: positively in terms of charter time cost savings and negatively in terms of fuel and cargo inventory costs. Fuel costs have fluctuated a great deal over the past decade and a half, in part due to rising and falling per gallon bunker fuel prices, and also to the rapid increases in vessel capacity. While the latter can reduce per cargo unit delivered fuel costs, the increasing costs of handling cargo in the larger vessels once they arrive in port may offset these gains to some extent, notably for long- distance shipping of high value commodities. Complicating the matter, the (spot) price of the goods being shipped may change a good deal during the multi-day course of such trans-oceanic shipments, while containership time charter rates have also been on something of a roller-coaster ride over the past decade and a half58: resulting in either the shipper, ship owner or charterer bearing any extra costs due to in-transit delays. Whoever pays, there is potentially a costly trade-off between fuel, charter time, and also, for long-distance high valued shipments, the cargo inventory carrying costs. One response to the considerable fluctuation in 55 http://www.searates.com/reference/portdistance/ 56 http://www.vhss.de/company 57 http://data.worldbank.org/indicator/IC.IMP.COST.CD 58 see Container Intelligence Monthly. Also Review of Maritime Transport (UNCTAD, 2013 ibid). 67 fuel as well as commodity spot market and charter prices in recent years has been fuel cost savings based on "slow streaming".59 Of particular interest to a study concerned with in-transit congestion, the late arrival of such large ships at seaports (of either embarkation or debarkation) may lead to costly delays in port/terminal transfers, impacting the schedules, and hence costs associated with the use of land- as well as sea-side modal assets. This is a topic that has been visited now by a number of studies, with liner shipping agents seeking to deploy their assets so as to maximize economies of scale associated with time at sea while minimizing any diseconomies of scale associated with time in port.60,61,62 Hsu and Hsieh (2005) estimate late cargo pickup costs as "inventory costs" that are positively correlated with cargo volume, cargo value, and the length of in-transit shipping and at-source or inport storage time. The ability to handle the much larger Post-Panamax vessels at the Port of Savannah also implies a greater potential for increased cargo inventory costs should delays at ports occur. Such delay costs can be significant should a significant port disruption, such as may occur due to a severe weather event.63 4.4.4 Intermodal and Within-Terminal Transfer Costs A significant component of multi-modal "door-to-door" supply chain costs involves freight inter-modal terminals. These costs are often the reason for using trucks to transport some parts, and also to transport finished automobiles significant distances, in order to fulfill limited size and timesensitive orders for specific vehicles. Unfortunately for public agency analysis, the recently published TRB-funded review of freight data sources found that (Holguin-Veras et al, 2013, page 50)64: " there are no regularly published data sources that provide sufficient data on freight terminal costs." 59 See UNCTAD, 2013 ibid. Also Wang, S. and Meng, Q. (2012) Sailing speed optimization for container ships in a liner shipping network. Transportation Research E 48: 701714 60 Cullinane, K. and Khanna, M. (1999) Economies of scale in large container ships. Journal of Transportation Economics and Policy 33(2): 198-208. 61 Payne, B. (2013) Economies of scale, economies of scope and potential diseconomies of scale. AAPA Presentation , September 2010. NYK Line (North America). 62Hsu, C-I, and Hsieh, Y.P. (2005) Shipping economic analysis for ultra-large container ships. Journal of the Eastern Asia Society for Transportation Studies 6: 936 951. 63 F. Southworth, J. Hayes, S. McLeod and A Strauss-Wieder (2014) Enhancing U.S. Port Resiliency as Part of Extended Intermodal Supply Chains. NCFRP Report 37. Transportation Research Board, Washington, D.C 64 Holgun-Veras, J. et al (2013) Freight Data Cost Elements. NCFRP Report 22. Transportation Research Board, Washington, D.C. 68 Based on their review, Holguin-Veras et al (2013) suggest that a useful general model for terminal activity costing includes the following five activities: administrative processing of cargo/vehicle/vessel entry/exit; internal (within-terminal) movements, cargo loading/unloading activities65; storage area organization and sorting; and, ancillary functions (including insurance, security, electricity, administration and other costs of terminal operation). According to Hussein and Petering (2009, page 30), truck drivers are "usually not responsible for loading their vehicles. They may, however, participate in unloading at the destination. Unloading palletized cargo using a forklift costs about $40 per truck and it consumes about 20 minutes. Unloading non-palletized cargo by hand consumes 2-3 hrs and is far more costly." However, such values appear to be illustrative only, rather than statistically grounded. Port or terminal specific tariff publications are suggested as a possible source of information, while noting that commodity and mode specific economies of scale among other aspects of cargo movement (some 119 unique cost elements are identified as potentially influencing such rates) may render such estimates at best approximate. However, according to The Tioga Group, Inc.'s review of marine container terminal operations (2010, page 86)66: "Both ports and marine terminal operators compete on cost, and do not want their costs accessible to either competitors or customers. Negotiated charges to ocean carriers are confidential and sensitive. Labor man-hours and costs are doubly sensitive." Since the present project could not afford to generate such a database from scratch, it was decided to incorporate (and enhance) an existing network data model to allow for such costs to be captured within the shipment source-to-destination routing process (see Section 4.5 below), should suitable sources of data emerge. Consistent with the recently published NCFRP 22 Report on such freight cost elements, this is seen as an important topic for attention, by compiling either a large body of port/terminal on-line cargo handling rates, and/or by modeling each of the above five terminal processing cost elements in some detail. A useful start is offered by the work done on the publicly 65 According to Hussein and Petering (2009) (ibid.), loading and unloading refers to the services of transferring cargo between the inside of a truck's trailer and "any place or point of rest on a wharf or terminal". 66 The Tioga Group, Inc. (2010) Improving marine container terminal productivity: development of productivity measures, proposed sources of data, and initial collection of data from proposed sources. Report to the Cargo Handling Cooperative Program. Maritime Administration, Washington, D.C. 69 available Intermodal Transportation and Inventory Cost (ITIC-ST) modeling software (FHWA, 2006).67 Developed in MS Excel for the purpose of truck and rail freight diversion modeling, ITICST contains spreadsheets for estimating the following components of a shipper's total logistics cost function: ordering cost capital carrying cost in transit capital carrying cost in inventory warehousing cost loading and unloading cost safety stock carrying cost cost of loss and damage claims with cargo handling and storage costs broken down by the different types of expenses associated with bulk, dry, open, or temperature controlled commodities. 4.5 Global Supply Chain Modeling: Putting Freight Costs on Intermodal Networks 4.5.1 Supply Chain Routing In order to compute the transportation costs associated with specific supply chains, or specific components of them, the above modal cost formulas have to be applied to specific origin-destinationcommodity shipments. This means bringing together the sort of geo-location based data on both product inputs and outputs described in Section 4.3 above with the shipping cost formulas described in Section 4.4. While the development of a software tool was not a part of the current research project, an early prototype was constructed in order to test the usefulness of the data collected. Figure 4.15 shows this idea, for what is currently an in-progress software development activity, drawing together the various flow and cost data elements and sub-models, and accessing data via a user interface that allows the analyst to select an appropriate mode of transportation (or tell the program to find the best mode, or mode-combination) in order to ship the goods to be moved. The idea behind this software is for a user to either select default travel speeds as well as per hour cargo holding and intermodal transfer costs, or derive them based on detailed, mode specific cost modeling formulas (based on the four sub-models shown in the blue box) at the bottom left corner of Figure 4.15. 67 ibid 70 Figure 4.16 shows an example of the current user interface (currently coded in MS Excel). These data are then sent to a least cost path-finding algorithm that operates on the global, multimodal link-node freight database described in Section 4.5 below, and computes over-the-network least cost paths based on either O-to-D travel times or distances. The resulting multi-link freight routings are Figure 4.15 Freight Data and Modeling Components of Automobile Manufacturing Supply Chains: Flows and Costs Modeling 1a) OEM Plant Location 1b) Local Area Parts Supplier Locations 3) Multimodal Freight Transportation Network Trucking Cost Model Rail Cost Model Trans-Oceanic Shipping Cost Model Within-Terminal Cost Model 2) Foreign Shipment Locations Shipment Routing Model User Interface Routing model run parameters & model outputs 4) Auto Dealer & Distribution Center Locations GIS Mapping Shipment O-to-D and Route Specific Travel Distances, Travel Times and Monetary Cost Estimates then output for use in dollar valued modal costs formulas such as those described in Section 4.4 above, and in a form that is also suitable for geographic information system (GIS) mapping (Caliper's Maptitude/TransCAD68 software is currently used for this). 68 http://www.caliper.com/ 71 4.5.2 Intermodal Freight Network Data Model To date, only a handful of multimodal freight network models have been developed that span entire countries or international shipments. The most widely reported of these freight network modeling efforts are the following: Figure 4.16 Prototype Supply Chain Routing Model Interface Input Parameters for Running FRSCMOD: Set Mode Specific Routing Impedance Factors: (Ctrl m = All Modes; Ctrl h = Highway; Ctrl r = Rail ; Ctrl n = Non-Rail; Ctrl p = Air) Highway Rail Inland Water Great Lakes Deep Sea Air READ IN MODEL INPUTS 1 0.2875 0.1429 0.1515 0.1724 10.0000 Set Intermodal Terminal Transfer (DEFAULT =2) and Throughput Impedances (DEFAULT =1): 2 1 Run FRSCMOD Set Origin Facility and Destination Facility TIERS for this model run: 15 0 Set ICP = 1 for travel time based routing (DEFAULT); = 2 for distance based routing 1 UPDATE Model OUTPUTS Set ISEA = 1 to include deep water links, = 0 to leave these links out of routing (model runs MUCH faster) 1 Highway Rail Inland Water Great Lakes Deep Sea Air MODE SPECIFIC DEFAULT AVERAGE TRAVEL SPEEDS (in MPH)* 50 22 20 24 25 400 PECIFIC DEFAULT AVERAGE VEHICLE TRAVEL COSTS/HOUR (in DOLLARS) 57 30 20 15 10 100 VERAGE INTERMODAL TERMINAL TRANSFER TIMES (in MINUTES)* 60 120 120 120 120 120 RAGE INTERMODAL TERMINAL TRANSFER COSTS/HOUR (in DOLLARS) 15 15 15 15 15 15 Average Within Rail Terminal Holding Times (in minutes) and Costs (in $/hour): 120 5 Average Within Seaport Terminal Holding Times (in minutes) and Costs (in $/hour): 300 5 Average Within Airport Terminal Holding Times (in minutes) and Costs (in $/hour): 300 5 * Note: default average speeds and intermodal transfer times may be over written by link specific network data - this is usually the case for highways (trucking). STAN (Strategic Planning of National and Regional Freight Transportation) network model, developed by Canadian researchers (Guelat, Florian and Crainic, 1990;69 see also Lubis, et al, 200370); 69 Guelat, J., Florian, M., Crainic, T.G. (1990) A Multimode Multiproduct Network Assignment Model for Strategic Planning of Freight Flows, Transportation Science, Vol.24.1:25-39. 72 NODUS (Geerts and Jourquin, 2000; Beuthe et al, 200171; Jourquin and Beuthe, 200372. 200673), developed in Europe; SMILE (Tavasszy et al, 1998;74 Bovernkerk, 2005 75) also developed in Europe; and ORNL: Oak Ridge National Laboratory's North American multimodal/inter-modal freight transportation network database (Southworth and Peterson, 200076; ORNL, 2013 77). The ORNL database has been used extensively in recent years to compute hundreds of thousands of inter-modal shipment distances for the US Commodity Flow Surveys and to estimate ton-mileage statistics for US DOT's Freight Analysis Framework and provides a very useful starting point for the development of a set of global, including truck, rail, inland water and trans-oceanic shipment routes. Fortunately, a recent version of this carefully documented network database is available in the public domain, and so was selected for use in this present study. Figure 4.17 shows the major modes included in this transportation network database, as well as the structure of the link-node "data model" adopted. An important feature of this network data model is the use of multiple links and nodes to represent detailed inter-modal connections. These intermodal links can carry a good deal of network information, and can be assigned both shipment loading and unloading costs, as well as within terminal/within seaport storage costs (including any time-sensitive delay costs reported). The top diagram (a) shows the modes are modeled. The bottom two diagrams in this figure show how intermodal transfers are incorporated in the network, 70 Lubis, H a-R.S.et al (2003) Multimodal freight transport network planning. Journal of the Eastern Asia Society for Transportation Studies5: 666-680. 71 Beuthe M., et al (2001) Freight transportation demand elasticities: a geographic multimodal transportation network analysis. Transportation Research 37E: 253-266. 72 Jourquin, B., and Beuthe, M. (2003) Multimodal freight networks, analysis with NODUS: methodology and applications. In Across the Border. Building upon a Quarter Century of Transport Research in Benelux. Dullaert, W., Jourquin, B. and Polak, J. (Eds.): 163-184. 73 Jourquin, B., and Beuthe, M. (2006) A decade of freight transport modeling with virtual networks: acquired experiences and new challenges. In Spatial Dynamics, Networks and Modelling. Reggiani, A. and Nijkamp, P. (Eds.):181-200. 74 Tavasszy, L.A., Smeenk, B. and Ruijgrok, C.K. (1998) A DSS for modelling logistic chains in freight transport policy analysis. International Transactions in Operational Research 5(6): 447-459. 75 Bovernkerk, M. (2005) SMILE+: the new improved Dutch national freight model system. European Transport Conference, Strasbourg, France. 76 F. Southworth, F. and Peterson, B.E. (2000) Intermodal and international freight network modeling. Transportation Research 8C:147-166. 77 ORNL (2013) On-Line Tools: CTA Transportation Networks. Intermodal Transportation Network Center for Transportation Analysis. Oak Ridge National Laboratory, Oak Ridge, TN. http://cta.ornl.gov/transnet/ 73 Figure 4.17 The ORNL Multi-Modal/ Inter-Model Freight Network Data Model a) Network Modes National Highway Network Database National Rail Network Database Operational Rail Network Database National Waterway Network Database Global Seaways Network Database Operational Waterways Network Database Intermodal, Truck, Rail and Water Terminals Databases Combines Inland , Intra- Coastal, Great Lakes & Trans-Oceanic Links ORNL Unified Multimodal/Intermodal Freight Network b) Example truck-rail-truck O-to-D shipment terminal to rail highway to transfer Terminal rail line haul terminal (storage) transfer Railroad #1 Railroad #2 Terminal (storage) c) Example Intermodal terminal transfers origin Railroad interline highway network link(s) notional local loading/access link to highway network destination Route Impedance = modal line-haul travel costs + intra-terminal storage/holding costs + inter-carrier (interlining) costs + local network to terminal transfer costs + terminal to local network transfer costs See Southworth, F. and Peterson, B.E. (2000) for details 74 for (b) the most common form of truck-rail-truck intermodal door-to-door shipments, and (c) in terms of moving freight through an inter-modal terminal, such as a seaport (see Southworth and Peterson, 2000 for technical details.).78 The original network database contains three separate files: a link attributes file, a node attributes file, and a link shape-point file. Of most importance to the current project are the link attributes (see Table 4.6), since link lengths determine shipment distances and to which can be added, or from which can be computed specific path (or route) based shipment costs. For this present project, the latest version of this periodically updated network database was expanded to include additional nodes and links for each of the foreign shipper locations as well as foreign seaports of lading reported to have handled West Point, GA bound shipments of automotive parts between mid-2008 and mid-2013. In doing so, each foreign seaport is represented by a four node/three link set of connections between existing land and water inks, as shown in Figure 4.18. Table 4.6 Network Link Attributes File Variable 1 LkSeqNum 2 Block 3 LinkID 4 A-node 5 B-node 6 Length 7 Imped 8 Access 9 Domestic 10 Oneway 11 Heading 12 Mode 13 Cargo 14 Name 15 NumPts Notes Link Sequence Number Mode or Mode Combination Identifier Unique Link Identifier A Node sequence number B Node sequence number Link length (in miles) Link Impedance (default or user defined) Type of link access restrictions CONUS, AK or HI, Canada, Mexico, et al. Flow direction retriction code Compass headings (N,S,E,W,blank) Truck, Rail, Water, Terminal, .... Code for cargo class specific routes 8-character link name (for Highways = Sign Route) Number of vertices in polyline (from link shape-file, for GIS mapping purposes) 78 Ibid. 75 Figure 4.18 Simplified Foreign Seaport Link-Node Representation Approach (land) link Transfer link Storage Link Transfer link Approach (water) link FOREIGN SEAPORT 5. Study Summary and Conclusions The research effort documented in this report set out to explore the little understood linkages between the micro-foundations of industry dynamics and economic activity, and the macro-congestion aspects of freight transport. A major barrier to such understanding has been the difficulty of obtaining the necessary data with which to carry out in-depth empirical analysis. The economic impacts and transportation requirements associated with the rapid development of Kia Motor's large automobile manufacturing plant in West Point, Georgia was chosen for study. Detailed, company and location specific data were collected from multiple sources, in order to piece together the economic impacts of the plant's impacts on the local and regional economy. This included a detailed functional and spatial mapping of the plant's supply chain inputs and outputs, and the demands they place on the broader south-east regional, as well as on the global and multimodal freight transportation system. Recognizing the limitations of past efforts to link specific instances of economic growth to both local business activity and regional transportation needs, the principal study effort went into collecting and merging the necessary data elements in sufficient detail to allow for in-depth empirical analysis. First, to get a clear understanding of the manufacturing and related processes, we developed a taxonomy of the automobile supply chain, identifying the major component 76 categories (Section 2). We also provided some details on which types of components are likely to be produced and delivered from areas relatively close to the KMMG plant, and which components may come from locations outside the region. The location of the numerous component suppliers in nearby areas was found to provide a substantial boost to the overall economic activity in the region. We identified all of the component suppliers that have located in Georgia and neighboring Alabama Counties following the decision of Kia Motors to locate in West Point. For these component suppliers, we provided some information on the types of components they manufacture and supply to KMMG, as well as their size, employment and investment amounts. Next, using the American Community Survey (ACS) database, we studied the economic impacts on counties, both core and non-core. Core counties are defined as those where a meaningful number of component suppliers are located, whereas non-core are neighboring counties with lack of meaningful component suppliers. We examined a comprehensive set of variables, including those related to employment in a wide range of occupations, schooling, educational attainment, and population and migration patterns, among others. In our examination of the data and computation of multipliers, we found that in some categories of economic and business development, the core counties showed substantial differences compared to non-core counties, while in other areas the differences were less clear. To understand the inflow of components to the KMMG plant, and outflow of finished automobiles, we collected data on the various freight flows associated with the automobile manufacturing supply chain, and its uses of local, regional and national highway, rail and waterway (including seaport) networks and cargo transfer facilities. Specific freight costing software tools were identified for use in future network-based flow modeling applications. To better understand the nature of these shipments, we enhanced and modified an existing global truck-rail-trans-oceanic freight network database to allow routing and mapping of individual product shipments, in order to better model the door-to-door costs involved. This included a brief exploration of the potential for significant freight movement bottlenecks, based on an interpretation of recent highway and rail traffic forecasts for the next two to three decades. While more detailed analytical and econometric modeling is needed to better understand the exact magnitudes of the effects of the KMMG plant's decision to locate in West Point, Georgia, the database constructed during the project represents an excellent starting point for such an effort. The project also demonstrates the level of effort needed to construct similar datasets for other manufacturing plant-based studies. 77 Appendix A. Component Suppliers Table A.1. List of Kia Motors Manufacturing Georgia (KMMG) Suppliers in Georgia and Alabama Company Autorica LLC DaeHa America Daehan Solutions Georgia LLC Daewon America DongWon Metals GLOVIS America, Inc. GLOVIS Georgia LLC Hamco America, Inc. Hanil E-Hwa Co., Ltd. Hiteco USA Inc. Hysco America Illinois Tool Works, Inc. DaeLim USA (ITW DaeLim) I-Master Corp. Johnson Controls KSI Kyungshin (Kyungshin Lear) Mando Corp. Mobis Alabama, Georgia Plant Nalara Georgia LLC Powertech America Pretty Products Prowill, LLC Sejong Georgia LLC Sewon America Inc. Sumika Polymer Compounds Yasufuku, USA, Inc. A1 Bar Code Systems AJin USA (Joon LLC) Alabama Bolt & Supply Inc. Alabama Graphics & Engineering Supply Inc. American Pipe & Supply Co. Inc. Arcadian Services Atchley Steel Company Inc. Bar Bender Steel Inc. Barloworld Handling LP Bermco Aluminum C & J Tech Alabama Changer & Dresser Corporation Chowel Weldparts CNC Enterprises Inc. CNJ Inc. Component Factory Automation Systems Plastic resin pellets NVH parts and interior components Suspension system Door Frames, roof molding, side absorbers, cross bars Vehicle Processing Center Integrated logistics Wheel & Tire Assembly Plastic Auto Trim Parts Cutting Tool and Machine Tool Accessory Manufacturing Steel supplier Plastic interior trip parts Automated Systems Automotive seating and door panels Wiring harnesses Electric power-steering gears and anti-lock brakes Front-end modules, front-rolling chassis Warehousing, Quality Engineering Automatic transmissions manufacturing Floor mats Industrial supplies and services Muffler and exhaust systems Stamped component and decorative trim Plastic Parts Plastic Injection & Blow Molding Parts Bar Coding Systems; Badge Printers Automotive Metal Stamping & Robotic Welding Hoses, Hydraulics, Fasteners Reprographics Pipes, Valves, Fitting & Plumbing Distribution Car Wash Chemicals/Systems Steel Fabrication Reinforcement Steel Cutting & Bending Services; Structural Steel & Concrete Distribution Distributes, Leases & Repairs Forklift Trucks Aluminum Refining & Smelting Automotive Plastic Injection Molding Resistance Welding Supplies Automotive Welding Electrodes Metal Fabrication & Equipment Maintenance Spec. Precision Mach. of Auto Brake Discs & Knuckles County Troup Troup Harris Troup Meriwether Troup Troup Troup Troup Troup Troup Troup Troup Harris Troup Meriwether Troup Troup Troup Troup Troup Troup Troup Spalding Troup Baldwin Chambers Montgomery Montgomery Jefferson Lauderdale Lee Montgomery Montgomery Jefferson Tallapoosa Calhoun Crenshaw Pike Lee State GA GA GA GA GA GA GA GA GA GA GA GA GA GA GA GA GA GA GA GA GA GA GA GA GA AL AL AL AL AL AL AL AL AL AL AL AL AL AL AL Zip Code 30241 30240 31833 30241 30230 31833 31833 31833 30240 31833 31833 30241 30241 31833 31833 30230 31833 31834 31833 30240 30241 30240 30241 30223 30240 36526 36852 36108 35233 35222 35630 36874 36117 36108 35222 35010 36207 36049 36081 36832 Address 23 Busch Dr, Lagrange, GA 201 Piedmont Circle LaGrange, GA 791 S Progress Pkwy, West Point, GA 20 Piedmont Circle, LaGrange, GA 475 Meriwether Park Drive, Hogansville GA 2000 Webb Rd, West Point, GA 6101 Sorento Rd, West Point, GA 6101 Sorento Rd, West Point, GA 104 Wiley Road, Lagrange, GA 6801 Kia Parkway West Point, GA 6501 Forte Rd, West Point, GA 50 Sl White Blvd, Lagrange, GA 112 Corporate Park East LaGrange, GA 1700 S Progress Pkwy, West Point, GA 1201 O.G. Skinner Dr, West Point, GA 955 Meriwether Park Dr, Hogansville, GA 7001 Kia Pkwy, West Point, GA 7001 Kia Parkway, Suite 201 West Point, GA 6801 Kia Parkway West Point, GA 1513 Redding Drive, Lagrange, GA 106 Corporate Park E Drive, LaGrange, GA 1641 Lukken Indus Drive W, Lagrange, GA 1000 Sewon Blvd., LaGrange, GA 109 E Solomon St, Griffin, GA 1 Yasufuku Place, LaGrange, GA PO Box 3046, Daphne, AL 1500 County Road 177, Cusseta, AL 630 Air Base Blvd, Montgomery, AL 2801 5th Ave S, Birmingham, AL 4100 Eastlake Blvd, Birmingham, AL 3109 Northington Ct, Florence, AL 12505 US Highway 280 E, Salem, AL 1143 Dozier Rd, Montgomery, AL 3001 Hayneville Rd, Montgomery, AL 3230 Messer Airport Hwy # K, Birmingham, AL 145 Plant 10 Drive Alexander City, AL 1527 Itc Way, Anniston, AL 5826 Montgomery Hwy, Luverne, AL 1708 Highway 231 N, Troy, AL 265 Teague Ct, Auburn, AL 78 Cumberland Plastic Systems LLC Cutting Tool Engineers Inc. Cutting Tool Engineers Inc. DaeDong Hi-Lex America Inc. (DDHLA) Daeil USA Corporation (Daeil) Daewon America DAS North America Davison Oil Company Inc. Die-Tech Inc. Dongwon Autopart Technology AL Plastic Automotive Components Lee AL Custom Cutting Tools for Metal Cutting Shelby AL Custom Cutting Tools for Metal Cutting Shelby AL Door Hardware & Module Systems; Power Window Chambers AL Struts & Parts for Automotive Industry Chambers AL Suspension Bars & Coils Lee AL Automotive Seat Components Montgomery AL Lubricant Manufacturing Mobile AL Die Cast Dies Lauderdale AL Door Frames, Side Impact Beams, Roof Molding,Console Brackets Crenshaw AL DSW Converting Knives Inc. Industrial Knives Jefferson AL Dudley C. Jackson Inc. Fastenal Company Glovis of Georgia Halla Climate Systems Alabama Corporation Hanil USA HPM Alabama Corporation (HONAM Petrochemical Corporation) HS Automotive Alabama Inc. Hwashin America Corporation HYSCO America Company Hyundai Motor Manufacturing Alabama LLC Hyundai Polytech America Company Inc. ILJIN Alabama Corporation Industrial Machine & Supply Inspec Tech Inc. Jay Mid-South LLC JIT Industries KC Sol-Tech Company Ltd. Key Safety Restraint Systems Keyport KwangSung America Corp. LeeHan America Mando America Corporation Alabama Merryweather Foam Inc. MGM Machining Inc. Mitchell Plastics Mobis Alabama LLC Motion Industries Inc. MP Tech America LLC Nemak USA Inc. Neocon USA Nitto Denko Automotive Alabama LLC OMI (Opelika MetalFab Inc.) Specialized Industrial Pumping & Spraying Systems Industrial Supplies, Safety, Jan-san, Tools Auto Warehousing & Logistics Sequencing Services HVAC Units, Front End Modules/FEM Plastic & Steel Tube Component Assembly for Fuel Sys. Injection Moldable Fiber Reinforced Thermoplastics Weather Stripping, Tubing & Rubber Hoses Chassis & Drive Train Automotive Body Parts Steel Coil & Sheeting for Chassis & Auto Body Parts Theta Gasoline Direct Injection & Multi Port Injection Engines, 1.8 Liter Nu Engine Anti-vibration Rubber & Thermoplastic Auto Parts Mechanical Power Transmission Equipment Man. Machine Parts Industry Specific Labels Metal Seat Frames Industrial Hydraulic Cylinders Tool & Die Steering Wheels, Air Bags, Seatbelts Warehousing & Distribution Blow Plastic Tubing for Auto Ventilation Systems; Automotive Plastic Injection Molding Automotive Air Filtration Systems Braking, Steering & Suspension Systems Gaskets, Auto Dunnage, Waterjet Cutting,Sound Control, Packaging Medical Foams Copper Electrodes Automotive Interior Components Chassis, Plastic Injection Molding, Distribution Bearings, Power Transmissions, Electric Motors, Pumps, Hoses Plastic Injection Molded Interior Parts Engine Blocks Powdered Metal Components Automotive Seals & Gaskets Automotive Shipping Racks, Steel Tubing, Material Shelby AL Montgomery AL Chambers AL Macon AL Elmore AL Lee AL Coffee AL Bulter AL Bulter AL Montgomery AL Barbour AL Russell AL Talladega AL Dekalb AL Etowah AL Morgan AL Lee AL Butler AL Baldwin AL Tallapoosa AL Chambers AL Lee AL Talladega AL Shelby AL Madison AL Montgomery AL Jefferson AL Chambers AL Talladega AL Madison AL Walker AL Lee AL 36832 35124 35124 36852 36863 36801 36105 36608 35630 36049 35205 35080 36108 36801 36075 36080 36832 36330 36037 36037 36105 36027 36869 35160 35989 35904 35640 36832 36037 36551 36853 36852 36801 35151 35080 35811 36108 36108 36852 35150 35805 35501 36804 229 Teague Ct, Auburn, AL 208 Commerce Parkway, Pelham, AL 208 Commerce Parkway, Pelham, AL 1195 County Road 177, Cusseta, AL 3509 45th St SW, Lanett, AL 4600 N Park Dr, Opelika, AL 201 County Ct, Montgomery, AL 8450 Tanner Williams Road, Mobile AL 4504 Helton Dr, Florence, AL 12865 Montgomery Hwy, Luverne, AL 1506 Reverend Abraham Woods Jr Blvd, Birmingham, AL 177 Mullins Dr, Helena, AL 4560 Newcomb Avenue, Montgomery, AL 404 Fox Run Ave, Opelika, AL 676 Hala Bama Drive, Shorter, AL 50 Hanil Drive, Tallassee AL 765 W Veterans Blvd, Auburn, AL 100 Sonata Dr, Enterprise, AL 693 Sherling Lake Rd, Greenville, AL 200 Team Member Ln, Greenville, AL 700 Hyundai Blvd, Montgomery, AL 112 Lakepoint Indus Park Road, Eufaula, AL 14 Downing Dr, Phenix City, AL 101 Costner Street, Talladega AL 46 Inspec Dr, Valley Head, AL 140 Thomas Dr, Gadsden, AL 2201 Hwy 31 S, Hartsellt, AL 1127 W Veterans Blvd, Auburn, AL 200 Pleasant Hill Ct, Greenville, AL 30427 County Road 49, Loxley, AL 217 Thwthet Industrial Park, Dadeville, AL 1230 County Road 177, Cusseta, Alabama 4201 N Park Dr, Opelika, AL 1212 Wynette Rd, Sylacauga, AL 117 Hicks Drive, Helena AL 1619 Highway 72 E, Huntsville, AL 1395 Mitchell Young Rd, Montgomery, AL 540 Trade Center Street, Montgomery, AL 1450 County Road 177, Cusseta AL 2100 Old Sylacauga Highway, Sylacauga AL 4950 Gilmer Drive NW, Huntsville AL 3611 Industrial Pkwy, Jasper, AL 1200 Steel St, Opelika, AL 79 Handling Equipment Opelika Scrap Material, Inc. Chambers AL 36804 2000 Steel St. Opelika, AL Posco America Corporation (POSCO-AAPC) Processed Steel Jefferson AL 35111 6500 Jefferson Metro Pkwy, mc Calla, AL Prolific Plastics Plastic Products (Injection Molding) Lee AL 36801 1304 Fox Run Avenue, Opelika AL Pyongsan America Inc. Automotive Plastics Components Lee AL 36832 760 W Veterans Boulevard, Auburn AL R O Deaderick Company Inc. Metalworking Machinery Merchant Wholesaler Madison AL 35824 350 Electronics Boulevard SW Huntsville, AL REHAU Automotive LLC Plastic Injection Molding, Painting & Assembly Cullman AL 35055 2424 Industrial Drive SW, Cullman AL Richway Transportation Services Trucking Terminal- Steel Rolls, Finished Products Butler AL 36033 572 Highway 31 S, Georgiana, AL Sabel Steel Service Steel Distribution Montgomery AL 36104 749 N Court Street, Montgomery AL SaeHaeSung Alabama Corp. Vehicle Welding & Stamping Parts Covington AL 36421 202 Progress Dr, Andalusia, AL Saudi Basic Industries Corp.(SABIC) Innovative Plastics Engineered Plastics Lowndes AL 36752 1 Plastics Dr, Lowndesboro, AL SCA Inc. Automotive Trimmed Exterior Plastic Parts Lee AL 36832 2230 Pumphrey Avenue, Auburn AL Sejin Alabama Plastic Injection Molded Automotive Parts Tallapoosa AL 36853 274 Thweatt Indus Boulevard, Dadeville, AL Sejong Alabama LLC Mufflers & Exhaust Systems Lowndes AL 36032 450 Old Fort Rd E, Fort Deposit, AL Seohan Auto USA Corporation Auto Front & Rear Axle Assembly Lee AL 36832 247 Teague Ct, Auburn, AL Seohan-NTN Driveshaft USA Constant Velocity Joints for Drive Shaft Assembly Lee AL 36832 249 Teague Ct, Auburn, AL Seoil America Inc. Automotive Adhesives & Sealants Elmore AL 36078 9 Twin Creeks Drive, Tallassee AL Simcoe Wood Products Inc. Wooden Pallets Cullman AL 35058 3730 al Highway 69 N, Cullman AL SMI Auto USA Inc. Automobile Parts Stamping Lee AL 36832 155 Alabama Street, Auburn, AL Southern Metal Fabricators Inc. Industrial Metal Fabricator Marshall AL 35950 1215 Frazier Road, Albertville, AL SteelFab Inc. of AL Structural Steel Randolph AL 36274 389 Steel View Dr, Roanoke, AL Store Room Fasteners Industrial Distribution Montgomery AL 36109 2361 Cong W L Dickinson Drive, Montgomery AL Sumitomo Electric Carbide Inc. Machine Tools & Supplies Merchant Wholesaler Madison AL 35805 5650 Sanderson Street NW # J, Huntsville AL Sung Woo USA Corporation Imports, Warehouses, Charge & Transport Batteries Montgomery AL 36116 6177 Perimeter Parkway, Montgomery AL TekLinks Computer Software Engineering Design Services Jefferson AL 35209 201 Summit Parkway, Birmingham AL Thompson Tractor Company Inc. Excavating Equipment, Engines Leasing & Distr. Jefferson AL 35217 2401 Pinson Valley Pkwy, Birmingham, AL ThyssenKrupp System Engineering Inc. Engineering Services (Systems) Madison AL 35758 485 Production Avenue 1/2, Madison AL Tomita USA Inc. MRO Supplies; Japanese OEM Machinery Parts Distribution Calhoun AL 36207 1400 Commerce Blvd # 8, Anniston, AL Tool Smith Company Inc. Power Hand Tools Merchant Wholesaler Jefferson AL 35233 1300 4th Ave S, Birmingham, AL Turner Supply Company Industrial Supplies Distribution Mobile AL 36602 250 N Royal St, Mobile, AL Vulcan Painters Inc. Industrial Painting Service Jefferson AL 35126 1549 Red Hollow Rd, Pinson, AL WESCO Distribution Inc. Electrical Supplies & Equipment Distribution Jefferson AL 35233 125 32nd St S, Birmingham, AL YE Tech Alabama Corporation Automation Equipment for Auto Assembly Randolph AL 36274 182 Industrial Ave, Roanoke, AL YESAC Corporation Automation Machinery, Pallets & Racks Elmore AL 36078 40 Yesac Drive, Tallasse, AL Yura Corporation Wiring Harnesses, Spark Plug Stick Coils Houston AL 36301 2431 W Main Street # 301, Dothan, AL Notes: (1) Table A.1 gives a list of 117 component suppliers of KMMG West Point assembly plant (25 in Georgia, 92 in Alabama) with company names, supplying components and location information. (2) Only information for suppliers in Georgia and Alabama is collected. Alabama data are from the 2013 Kia supplier list composed and provided by Alabama Department of Commerce based on their Alabama Industrial Database. Georgia data are from the news and articles on Atlanta Journal Constitution, Georgia Chamber of Commerce website, and 2011 and 2013 Troup County Directory of Manufacturers. Address information, if not provided by the previous sources, is from company websites, Google Maps, or www.manta.com. 80 Table A.2. Employment and Investments of KMMG Suppliers in Georgia Company Component Year Location Jobs 2011 2013 Employment Employment Investment Autorica LLC Factory Automation Systems 2008 Troup, GA 10 DaeHa America Plastic resin pellets 2012 Troup, GA 23 Daehan Solutions Georgia LLC NVH parts and interior components 2009 Harris, GA 300 191 $35 million Daewon America Suspension System 2012 Troup, GA 45 $14 million DongWon Metals Automotive stamplings, Door Frames, bumper side absorbers, 2008 Meriwether, GA 300 224 carriers 275 $30 million Glovis America Inc. Vehicle Processing Center 2009 Troup, GA 250 285 GLOVIS Georgia LLC Integrated logistics 2009 Troup, GA 600-700 224 275 $60 million Hamco America, Inc. Wheel & Tire Assembly 2009 Troup, GA 4 Hanil E-Hwa Co., Ltd. Interior parts 2010 Troup, GA 173 124 225 $8.45 million Hiteco USA Inc. Cutting Tool and Machine Tool Accessory Manufacturing 2010 Troup, GA Hysco America Steel supplier 2008 Troup, GA 50 7 9 Illinois Tool Works, Inc. Plastic interior trip parts 2008 Troup, GA 75 200 200 DaeLim USA I-Master Corp. Automated Systems 2008 Troup, GA 5 8 Johnson Controls Automotive seating 2009 Harris, GA 310 661 670 KSI Kyungshin Wiring harnesses 2009 Troup, GA 50-70 50 50 $3.5 million Mando Corp. Electric power-steering gears and anti-lock brakes 2011 Meriwether, GA 426 200 $200 million Mobis AL LLC Front-end modules, front-rolling chassis 2008 Troup, GA 600 350 840 $60 million Nalara Georgia LLC Warehousing, Quality Engineering 2012 Troup, GA 9 Power Tech America Transmissions 2010 Troup, GA 355 331 481 $150 million Pretty Products Floor mats 2008 Troup, GA 130-185 151 $6.5 million Prowill, LLC Industrial supplies & services 2011 Troup, GA Sejong Georgia LLC Muffler and exhaust systems 2009 Troup, GA 250 116 176 $27.8 million Sewon America Inc. Stamped component and decorative trim 2009 Troup, GA 700 800 912 $170 million Sumika Polymer Compounds Plastic Parts 2009 Spalding, GA 50 Yasufuku, USA, Inc. Plastic Injection & Blow Molding Parts 2005 Troup, GA 38 50 Notes: (1) Georgia data are from the news and articles on Atlanta Journal Constitution and Georgia Chamber of Commerce website, and 2011 and 2013 Troup County Directory of Manufacturers. Address information, if not provided by the previous sources, is from company websites, Google Maps, or www.manta.com. (2) Jobs and investment are the announced job numbers and investment when the projects were launched. (3) 2011 and 2013 employment data are from 2011 and 2013 Troup County Directory of Manufacturers respectively. (4) Pretty Products filed bankruptcy in 2010. See: http://www.burbageweddell.com/2010/06/17/pretty-products-bankruptcy-backgound/ 81 Table A.3. Employment and Investments of KMMG Suppliers in Alabama Company A-Jin USA (Joon LLC) A-Jin USA (Joon LLC) Alabama Bolt & Supply Inc. Alabama Graphics & Engineering Supply Inc. American Pipe & Supply Co. Inc. Atchley Steel Company Inc. Bar Bender Steel Inc. Barloworld Handling LP Bermco Aluminum C & J Tech Alabama CNJ Inc. CNJ Inc. CNJ Inc. Cumberland Plastic Systems LLC Cumberland Plastic Systems LLC Cumberland Plastic Systems LLC DaeDong Hi-Lex America Inc. (DDHLA) DaeDong Hi-Lex America Inc. (DDHLA) Daeil USA Corporation (Daeil) Daewon America Daewon America DAS North America DSW Converting Knives Inc. DSW Converting Knives Inc. DSW Converting Knives Inc. Fastenal Company Glovis of Georgia Halla Climate Systems Alabama Corporation Hanil USA Hanil USA Hanil USA HPM Alabama Corporation Hyundai Motor Manufacturing Alabama LLC KC Sol-Tech Company Ltd. KwangSung America Corp. Component Automotive Metal Stamping & Robotic Welding Automotive Metal Stamping & Robotic Welding Hoses, Hydraulics, Fasteners Reprographics Pipes, Valves, Fitting & Plumbing Distribution Steel Fabrication Reinforcement Steel Cutting & Bending Services; Structural Steel & Concrete Distribution Distributes, Leases & Repairs Forklift Trucks Aluminum Refining & Smelting Automotive Plastic Injection Molding Spec. Precision Mach. of Auto Brake Discs & Knuckles Spec. Precision Mach. of Auto Brake Discs & Knuckles Spec. Precision Mach. of Auto Brake Discs & Knuckles Plastic Automotive Components Plastic Automotive Components Plastic Automotive Components Door Hardware & Module Systems; Power Window Door Hardware & Module Systems; Power Window Struts & Parts for Automotive Industry Suspension Bars & Coils Suspension Bars & Coils Automotive Seat Components Industrial Knives Industrial Knives Industrial Knives Industrial Supplies, Safety, Jan-san, Tools Auto Warehousing & Logistics Sequencing Services HVAC Units, Front End Modules/FEM Plastic & Steel Tube Component Assembly for Fuel Sys. Plastic & Steel Tube Component Assembly for Fuel Sys. Plastic & Steel Tube Component Assembly for Fuel Sys. Injection Moldable Fiber Reinforced Thermoplastics Theta Gasoline Direct Injection & Multi Port Injection Engines, 1.8 Liter Nu Engine Tool & Die Blow Plastic Tubing for Auto Ventilation Systems; Automotive Plastic Injection Molding Year 2008 2010 Location Lanett, Chambers, AL Lanett, Chambers, AL Montgomery, AL Montgomery, AL 2008 Birmingham, Jefferson, AL Lee, AL Montgomery, AL 2011 2011 2007 2010 2011 2008 Montgomery, AL Bessemer, Jefferson, AL Alexander City, Tallapoosa, AL Auburn, Lee, AL Auburn, Lee, AL Auburn, Lee, AL Auburn, Lee, AL 2009 Auburn, Lee, AL 2010 Auburn, Lee, AL 2008 Cusseta, Chambers, AL 2010 Cusseta, Chambers, AL 2010 2007 2010 2008 2009 2011 2010 2007 Lanett, Chambers, AL Opelika, Lee, AL Opelika, Lee, AL Montgomery, AL Birmingham, Jefferson, AL Birmingham, Jefferson, AL Birmingham, Jefferson, AL Montgomery, AL Valley, Chambers, AL Shorter, Macon, AL 2007 2008 2010 2011 2007 Tallassee, Elmore, AL Tallassee, Elmore, AL Tallassee, Elmore, AL Auburn, Lee, AL Montgomery, Montgomery, AL Auburn, Lee, AL 2008 Dadeville, Tallapoosa, AL Jobs 450 150 Na Na Investment $89 million $50 million Na Na 10 $2 million Na Na Na Na Na 10 150 25 25 18 10 Na $12 million $9.8 million $15.10 million $7.29 million $20.76 million $1 million 51 $1.9 million 5 $0.5 million 103 $10.9 million 30 $6.5 million 70 Na Na Na 5 2 0 Na 200 130 $10.7 million $7.3 million $6.2 million Na $0.55 million $0.21 million $0.25 million Na $20 million Na 3 $15 million 90 Na 60 $3 million 30 $9.25 million 522 $270 million Na Na 200 Na 82 KwangSung America Corp. KwangSung America Corp. KwangSung America Corp. LeeHan America Mando America Corporation Alabama Mando America Corporation Alabama Mando America Corporation Alabama Mando America Corporation Alabama Mobis Alabama LLC Mobis Alabama LLC Mobis Alabama LLC Motion Industries Inc. MP Tech America LLC OMI (Opelika MetalFab Inc.) Posco America Corporation (POSCO-AAPC) Prolific Plastics Pyongsan America Inc. Pyongsan America Inc. Pyongsan America Inc. Sabel Steel Service SCA Inc. SCA Inc. SCA Inc. Sejin Alabama Sejin Alabama Sejin Alabama Sejin Alabama Seohan Auto USA Corporation Seohan Auto USA Corporation Seohan Auto USA Corporation Seohan-NTN Driveshaft USA Seohan-NTN Driveshaft USA Seoil America Inc. SMI Auto USA Inc. SteelFab Inc. of AL SteelFab Inc. of AL Store Room Fasteners Sungwoo USA Corporation TekLinks TekLinks Thompson Tractor Company Inc. Blow Plastic Tubing for Auto Ventilation Systems; Automotive Plastic Injection Molding Blow Plastic Tubing for Auto Ventilation Systems; Automotive Plastic Injection Molding Blow Plastic Tubing for Auto Ventilation Systems; Automotive Plastic Injection Molding Automotive Air Filtration Systems Braking, Steering & Suspension Systems Braking, Steering & Suspension Systems Braking, Steering & Suspension Systems Braking, Steering & Suspension Systems Chassis, Plastic Injection Molding, Distribution Chassis, Plastic Injection Molding, Distribution Chassis, Plastic Injection Molding, Distribution Bearings, Power Transmissions, Electric Motors, Pumps, Hoses Plastic Injection Molded Interior Parts Automotive Shipping Racks, Steel Tubing, Material Handling Equipment Processed Steel Plastic Products (Injection Molding) Automotive Plastics Components Automotive Plastics Components Automotive Plastics Components Steel Distribution Automotive Trimmed Exterior Plastic Parts Automotive Trimmed Exterior Plastic Parts Automotive Trimmed Exterior Plastic Parts Plastic Injection Molded Automotive Parts Plastic Injection Molded Automotive Parts Plastic Injection Molded Automotive Parts Plastic Injection Molded Automotive Parts Auto Front & Rear Axle Assembly Auto Front & Rear Axle Assembly Auto Front & Rear Axle Assembly Constant Velocity Joints for Drive Shaft Assembly Constant Velocity Joints for Drive Shaft Assembly Automotive Adhesives & Sealants Automobile Parts Stamping Structural Steel Structural Steel Industrial Distribution Imports, Warehouses, Charge & Transport Batteries Computer Software Engineering Design Services Computer Software Engineering Design Services Excavating Equipment, Engines Leasing & Distr. 83 2009 Dadeville, Tallapoosa, AL 2011 Dadeville, Tallapoosa, AL 2011 Dadeville, Tallapoosa, AL 2011 Cusseta, Chambers, AL 2007 Lanett, Chambers, AL 2007 Opelika, Lee, AL 2008 Opelika, Lee, AL 2010 Opelika, Lee, AL 2007 2010 2011 2011 2008 Montgomery, Montgomery, AL Montgomery, Montgomery, AL Montgomery, Montgomery, AL Birmingham, Jefferson, AL Cusseta, Chambers, AL Opelika, Lee, AL 2009 McCalla, Jefferson, AL 2007 2008 2010 2010 2007 2010 2011 2007 2009 2010 2011 2007 2009 2011 2007 2009 2008 2007 2009 2011 2009 2011 2007 Opelika, Lee, AL Auburn, Lee, AL Auburn, Lee, AL Auburn, Lee, AL Montgomery, Montgomery, AL Auburn, Lee, AL Auburn, Lee, AL Auburn, Lee, AL Dadeville, Tallapoosa, AL Dadeville, Tallapoosa, AL Dadeville, Tallapoosa, AL Dadeville, Tallapoosa, AL Auburn, Lee, AL Auburn, Lee, AL Auburn, Lee, AL Auburn, Lee, AL Auburn, Lee, AL Elmore, AL Auburn, Lee, AL Roanoke, Randolph, AL Roanoke, Randolph, AL Montgomery, Montgomery, AL Montgomery, Montgomery, AL Birmingham, Jefferson, AL Birmingham, Jefferson, AL Birmingham, Jefferson, AL 170 $5 million 100 $8.30 million 1000 $8 million 51 $3.2 million 16 $3.3 million 77 $21 million 200 $25 million 5 $4.3 million 140 250 133 100 250 Na $55.6 million $59.7 million $38.97 million Na $30 million Na 60 $17 million 20 90 100 100 Na 40 180 21 300 50 70 160 74 97 10 96 32 Na 33 52 Na 5 Na 10 0 25 $0.4 million $5.4 million $5.5 million $5 million Na $8.2 million $15.1 million $1.3 million $30 million $5 million $7 million $15.79 million $22 million $9.69 million $7.9 million $16 million $6.9 million Na $3 million $0.5 million $1 million $1.85 million Na $0.5 million $1 million $2 million Tool Smith Company Inc. Power Hand Tools Merchant Wholesaler Jefferson, AL Na Na Vulcan Painters Inc. Industrial Painting Service Jefferson, AL Na Na WESCO Distribution Inc. Electrical Supplies & Equipment Distribution Jefferson, AL Na Na YE Tech Alabama Corporation Automation Equipment for Auto Assembly 2009 Roanoke, Randolph, AL 5 $0.2 million YESAC Corporation Automation Machinery, Pallets & Racks 2008 Tallassee, Elmore, AL 60 $0.29 million Notes: (1) Alabama supplier names and components are from the 2013 Kia supplier list composed and provided by Alabama Department of Commerce based on their Alabama Industrial Database. Address information, if not provided by the previous list, is from company websites, Google Maps, or www.manta.com. (2) Jobs and investment are provided by 2007, 2008, 2009, 2010, and 2011 Alabama New and Expanding Industry Announcement prepared by Alabama Development Office. Expansions of the same company in different years are entered as separate entries. Some Kia suppliers are not recorded in the Announcements. 84 Appendix B. Economic and Business Effects The tables in appendix B present the actual data based on which the percentage changes and the multipliers are calculated. Table B.1. Actual Change in Employment by Industry State County Core Management 2006 2010 Service 2006 2010 Sales and office 2006 2010 Construction 2006 2010 Manufacturing 2006 2010 AL AL 611,109 642,794 301,058 332,117 507,261 499,240 161,695 139,286 308,148 272,147 AL Core Avg. 14,608 13,421 6,925 6,897 11,272 9,533 3,022 2,243 5,532 4,736 AL Non-core Avg 3,686 4,250 2,443 2,786 3,919 3,898 1,201 953 2,231 2,312 AL Autauga N 6,333 7,268 3,501 3,877 6,515 6,546 2,084 1,434 3,615 3,082 AL Bullock Y N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A AL Butler Y N/A 2,051 N/A 1,550 N/A 1,902 539 481 2,655 1,421 AL Chambers Y 2,948 3,230 2,130 1,701 3,582 3,155 1,467 908 4,356 2,777 AL Crenshaw N N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A AL Elmore Y 8,770 12,097 4,037 5,633 8,264 8,441 3,368 2,278 4,040 4,210 AL Lee Y 21,224 22,762 8,868 10,375 15,184 14,753 4,745 3,731 7,387 6,924 AL Lowndes N N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A AL Macon N 2,239 1,690 1,746 1,957 2,523 1,954 606 606 550 915 AL Montgomery Y 35,556 34,939 16,617 19,148 25,706 25,836 6,370 4,880 10,473 9,395 AL Pike N 3,589 4,016 2,513 2,604 3,515 3,650 851 725 2,073 2,496 AL Randolph N 1,913 2,450 1,265 1,088 1,636 1,705 736 500 2,412 2,419 AL Russell N 4,356 5,824 3,190 4,405 5,408 5,636 1,729 1,502 2,505 2,646 AL Tallapoosa Y 4,542 5,447 2,971 2,976 3,623 3,109 1,640 1,180 4,282 3,686 GA GA 1,442,258 1,491,797 653,198 698,071 1,130,911 1,069,270 384,108 285,015 498,708 446,074 GA Core Avg. 7,822 7,472 4,831 4,659 6,863 6,898 1,883 1,292 6,208 5,565 GA Non-core Avg 19,681 20,199 22,927 21,571 3,105 2,982 1,381 1,319 1,834 1,714 GA Atlanta MSA 892,757 904,916 326,268 350,562 645,929 594,522 206,636 149,737 204,954 193,808 GA Harris N 5,321 6,854 1,582 1,628 3,698 3,741 972 942 1,235 1,248 GA Heard N N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A GA Meriwether N N/A N/A N/A N/A N/A N/A 814 533 2,076 1,579 GA Talbot N N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A GA Troup Y 7,822 7,472 4,831 4,659 6,863 6,898 1,883 1,292 6,208 5,565 GA Upson N 34,041 33,544 44,272 41,513 2,511 2,222 2,356 2,482 2,190 2,316 Notes: (1) This table contains the actual raw data of Table 3.1. (2) All data are from ACS 2005-2007 and ACS 2009-2011. Data of Bullock, Crenshaw, Lowndes, Heard, and Talbot, and some data for Butler and Meriwether are not available. (3) 15 counties are selected to create a smaller area of Atlanta MSA. Those counties are: Carroll, Cherokee, Clayton, Cobb, Coweta, DeKalb, Douglas, Fayette, Forsyth, Fulton, Gwinnett, Hall, Henry, Paulding, and Rockdale. 85 Table B.1. Actual Change in Employment by Industry ... Cont'd State County Core Wholesale trade Retail trade Transportation and warehousing Finance and insurance Education and health care 2006 2010 2006 2010 2006 2010 2006 2010 2006 2010 AL AL 70,087 54,517 245,235 239,950 104,934 101,662 119,194 113,057 401,690 435,798 AL Core Avg. 945 758 4,377 4,594 1,547 1,500 2,596 2,160 7,932 9,210 AL Non-core Avg 444 318 1,897 1,722 761 747 899 853 2,989 3,583 AL Autauga N 887 756 2,873 3,037 970 1,444 1,298 1,416 3,514 4,586 AL Bullock Y N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A AL Butler Y 123 156 983 1,127 340 381 298 266 1,436 1,466 AL Chambers Y 235 216 1,488 1,525 846 627 613 635 2,373 2,578 AL Crenshaw N N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A AL Elmore Y 861 793 3,309 3,753 1,872 1,353 2,049 1,957 4,593 7,387 AL Lee Y 1,284 835 7,504 7,728 1,835 2,162 4,182 3,392 15,675 17,527 AL Lowndes N N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A AL Macon N 178 65 1,029 736 250 272 532 120 2,911 2,831 AL Montgomery Y 2,600 2,319 11,589 11,823 3,764 3,779 7,352 6,293 19,931 21,169 AL Pike N 435 334 1,910 1,596 1,142 863 656 496 3,249 3,746 AL Randolph N 265 129 1,177 716 585 463 231 190 1,654 1,886 AL Russell N 454 304 2,496 2,524 856 693 1,777 2,044 3,617 4,866 AL Tallapoosa Y 566 227 1,387 1,605 625 696 1,080 418 3,581 5,131 GA GA 157,458 133,743 513,804 507,617 264,611 249,818 298,689 260,308 800,785 886,003 GA Core Avg. 685 598 3,481 4,144 1,297 1,526 1,502 1,604 5,789 5,721 GA Non-core Avg 585 555 1,703 1,457 482 335 1,219 948 1,687 2,213 GA Atlanta MSA 98,991 83,758 277,532 268,755 153,041 143,899 198,783 167,021 395,212 435,161 GA Harris N 270 618 1,584 1,627 496 457 1,811 1,241 2,792 4,427 GA Heard N N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A GA Meriwether N 213 218 894 1,097 754 487 570 378 1,534 1,652 GA Talbot N N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A GA Troup Y 685 598 3,481 4,144 1,297 1,526 1,502 1,604 5,789 5,721 GA Upson N 1,273 830 2,631 1,648 197 60 1,276 1,226 736 559 Notes: (1) This table contains the actual raw data of Table 3.1 Cont'd. (2) All data are from ACS 2005-2007 and ACS 2009-2011. Data of Bullock, Crenshaw, Lowndes, Heard, and Talbot are not available. (3) 15 counties are selected to create a smaller area of Atlanta MSA. Those counties are: Carroll, Cherokee, Clayton, Cobb, Coweta, DeKalb, Douglas, Fayette, Forsyth, Fulton, Gwinnett, Hall, Henry, Paulding, and Rockdale. 86 Table B.2. Actual Change in Migration State County Core Residents from other counties Residents from Residents from other states abroad US born citizen Foreign-born citizen Naturalized citizen Non-citizens 2006 2010 2006 2010 2006 2010 2006 2010 2006 2010 2006 2010 2006 2010 AL AL 152,684 152,446 133,710 114,048 17,909 15,228 4,422,755 4,576,478 130,790 168,416 40,389 50,505 90,401 117,911 AL Core Avg. 4,391 4,033 3,152 3,210 535 400 84,202 86,894 2,302 3,275 885 1,176 1,873 2,728 AL Non-core Avg 1,170 1,324 1,662 1,731 81 265 33,797 35,757 566 764 248 252 317 512 AL Autauga N 1,596 2,188 2,056 2,145 131 66 47,460 53,122 708 928 419 290 289 638 AL Bullock Y N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A AL Butler Y 252 326 182 142 0 53 20,167 20,653 20 133 N/A N/A N/A N/A AL Chambers Y 539 739 1,384 847 30 71 34,707 33,585 128 464 15 127 113 337 AL Crenshaw N N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A AL Elmore Y 7,934 5,914 1,771 1,373 194 157 73,744 77,064 1,156 1,616 541 488 615 1,128 AL Lee Y 8,173 8,390 7,139 7,656 1,208 983 121,278 132,164 5,024 7,084 1,555 1,608 3,469 5,476 AL Lowndes N N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A AL Macon N 543 665 1,729 1,184 22 73 22,084 20,670 356 434 76 399 280 35 AL Montgomery Y 7,493 7,306 7,930 8,234 1,734 1,026 215,204 217,435 7,036 9,341 2,213 3,443 4,823 5,898 AL Pike N 2,170 2,074 763 1,353 78 615 29,131 31,322 649 1,262 55 64 594 1,198 AL Randolph N 670 451 1,057 620 54 47 22,280 22,312 156 408 42 75 114 333 AL Russell N 872 1,240 2,705 3,353 122 524 48,031 51,361 959 789 650 433 309 356 AL Tallapoosa Y 1,955 1,520 507 1,007 45 109 40,113 40,460 448 1,012 103 212 345 800 GA GA 484,463 455,652 354,713 262,126 58,687 43,655 8,385,737 8,649,632 841,282 941,301 265,429 342,847 575,853 598,454 GA Core Avg. 2,573 2,681 2,401 1,711 114 496 60,914 63,694 1,553 2,986 402 872 1,151 2,114 GA Non-core Avg 1,394 1,568 597 202 62 14 25,455 26,389 516 471 229 389 287 225 GA Atlanta MSA 252,021 239,550 203,755 135,408 36,137 25,283 4,010,963 4,077,738 658,631 715,880 206,099 266,562 452,532 449,318 GA Harris N 1,947 1,957 897 253 33 37 26,926 30,760 825 754 591 583 234 171 GA Heard N N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A GA Meriwether N 909 1,698 389 119 60 6 22,453 21,789 250 186 0 N/A 250 N/A GA Talbot N N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A GA Troup Y 2,573 2,681 2,401 1,711 114 496 60,914 63,694 1,553 2,986 402 872 1,151 2,114 GA Upson N 1,325 1,050 504 233 92 0 26,987 26,617 474 473 97 195 377 278 Notes: (1) This table contains the actual raw data of Table 3.2. (2) All data are from ACS 2005-2007 and ACS 2009-2011. Data of Bullock, Crenshaw, Lowndes, Heard, and Talbot, and some data of Butler and Meriwether are not available. (3) 15 counties are selected to create a smaller area of Atlanta MSA. Those counties are: Carroll, Cherokee, Clayton, Cobb, Coweta, DeKalb, Douglas, Fayette, Forsyth, Fulton, Gwinnett, Hall, Henry, Paulding, and Rockdale. 87 Table B.3. Actual Change in Education State County Core Population 25 years and over Less than 9th grade 9th to 12th grade, no diploma High school graduate (includes equivalency) 2006 2010 2006 2010 2006 2010 2006 2010 AL AL 3,015,910 3,168,795 199,377 190,935 404,661 371,406 973,126 982,556 AL Core Avg 54,310 57,667 3,218 2,872 7,152 6,554 16,197 16,917 AL Non-core Avg 22,167 23,502 1,754 1,571 3,276 3,170 7,728 7,863 AL Autauga N 31,540 35,246 1,550 1,669 3,276 3,471 11,290 12,094 AL Bullock Y N/A N/A N/A N/A N/A N/A N/A N/A AL Butler Y 13,332 14,009 1,084 1,076 2,258 2,404 4,966 5,051 AL Chambers Y 23,974 23,589 2,109 1,788 4,626 4,605 8,505 8,120 AL Crenshaw N N/A N/A N/A N/A N/A N/A N/A N/A AL Elmore Y 49,147 53,208 2,842 1,981 6,440 5,378 18,104 16,962 AL Lee Y 71,089 79,611 3,776 2,892 7,313 8,242 18,465 21,599 AL Lowndes N N/A N/A N/A N/A N/A N/A N/A N/A AL Macon N 13,312 12,894 1,281 1,004 1,905 1,873 3,293 3,628 AL Montgomery Y 140,048 146,610 6,765 7,387 17,192 14,571 38,352 39,975 AL Pike N 18,034 19,071 1,547 880 3,355 2,723 5,965 6,640 AL Randolph N 14,988 15,606 1,618 1,904 2,415 2,682 5,966 5,431 AL Russell N 32,960 34,691 2,773 2,397 5,430 5,103 12,125 11,522 AL Tallapoosa Y 28,269 28,973 2,732 2,107 5,085 4,122 8,792 9,792 GA GA 5,945,347 6,243,020 378,127 368,744 678,996 612,985 1,799,261 1,821,432 GA Core Avg 39,856 42,681 3,223 2,866 5,169 5,888 14,209 14,891 GA Non-core Avg 17,458 18,353 1,331 1,227 2,692 2,244 6,408 6,316 GA Atlanta MSA 3,032,660 3,124,259 157,105 160,092 247,153 215,493 778,884 748,913 GA Harris N 18,880 22,136 1,054 794 1,756 1,443 5,671 6,028 GA Heard N N/A N/A N/A N/A N/A N/A N/A N/A GA Meriwether N 15,049 14,625 1,265 1,339 2,609 2,642 6,371 6,062 GA Talbot N N/A N/A N/A N/A N/A N/A N/A N/A GA Troup Y 39,856 42,681 3,223 2,866 5,169 5,888 14,209 14,891 GA Upson N 18,446 18,297 1,675 1,549 3,712 2,646 7,181 6,859 Notes: (1) This table contains the actual raw data of Table 3.3. (2) All data are from ACS 2005-2007 and ACS 2009-2011. Data of Bullock, Crenshaw, Lowndes, Heard, and Talbot are not available. (3) 15 counties are selected to create a smaller area of Atlanta MSA. Those counties are: Carroll, Cherokee, Clayton, Cobb, Coweta, DeKalb, Douglas, Fayette, Forsyth, Fulton, Gwinnett, Hall, Henry, Paulding, and Rockdale. 88 Table B.3. Actual Change in Education ... Cont'd State County Core Some college, no degree 2006 2010 Associate's degree 2006 2010 Bachelor's degree 2006 2010 Graduate or professional degree 2006 2010 AL AL 601,540 695,347 199,544 228,587 402,159 445,355 235,503 254,609 AL Core Avg. 10,649 12,322 3,438 3,719 8,061 9,402 5,595 5,883 AL Non-core Avg 4,353 5,139 1,374 1,610 2,216 2,739 1,466 1,410 AL Autauga N 6,717 8,164 2,196 2,596 4,292 4,961 2,219 2,291 AL Bullock Y N/A N/A N/A N/A N/A N/A N/A N/A AL Butler Y 2,261 2,586 1,124 967 1,215 1,264 424 661 AL Chambers Y 4,547 5,009 1,744 1,367 1,678 1,729 765 971 AL Crenshaw N N/A N/A N/A N/A N/A N/A N/A N/A AL Elmore Y 9,817 13,009 3,015 4,149 6,171 7,681 2,758 4,048 AL Lee Y 14,319 16,297 5,338 5,866 12,173 14,291 9,705 10,424 AL Lowndes N N/A N/A N/A N/A N/A N/A N/A N/A AL Macon N 2,645 3,006 1,165 879 1,650 1,182 1,373 1,322 AL Montgomery Y 27,765 31,685 7,457 7,537 24,326 27,919 18,191 17,536 AL Pike N 2,904 3,863 585 614 2,162 2,869 1,516 1,482 AL Randolph N 2,252 2,979 996 944 857 1,092 884 574 AL Russell N 7,247 7,681 1,927 3,015 2,120 3,593 1,338 1,380 AL Tallapoosa Y 5,184 5,345 1,948 2,425 2,802 3,526 1,726 1,656 GA GA 1,130,853 1,310,045 379,421 416,902 1,026,571 1,098,226 552,118 614,686 GA Core Avg. 7,221 9,025 2,445 2,309 4,868 4,702 2,721 3,000 GA Non-core Avg 3,075 4,027 1,109 1,150 1,731 2,058 1,111 1,330 GA Atlanta MSA 584,362 653,766 196,847 213,267 709,138 739,612 359,171 393,116 GA Harris N 3,916 5,033 1,643 2,006 3,010 4,166 1,830 2,666 GA Heard N N/A N/A N/A N/A N/A N/A N/A N/A GA Meriwether N 2,395 2,723 754 460 881 935 774 464 GA Talbot N N/A N/A N/A N/A N/A N/A N/A N/A GA Troup Y 7,221 9,025 2,445 2,309 4,868 4,702 2,721 3,000 GA Upson N 2,915 4,325 931 985 1,302 1,073 730 860 Notes: (1) This table contains the actual raw data of Table 3.3 Cont'd. (2) All data are from ACS 2005-2007 and ACS 2009-2011. Data of Bullock, Crenshaw, Lowndes, Heard, and Talbot are not available. (3) 15 counties are selected to create a smaller area of Atlanta MSA. Those counties are: Carroll, Cherokee, Clayton, Cobb, Coweta, DeKalb, Douglas, Fayette, Forsyth, Fulton, Gwinnett, Hall, Henry, Paulding, and Rockdale. 89 Table B.4. Actual Change in Schooling State County Core Population 3 years and over enrolled in school 2006 2010 Nursery school, preschool 2006 2010 Kindergarten 2006 2010 Elementary school (grades 1-8) 2006 2010 High school (grades 9-12) 2006 2010 College or graduate school 2006 2010 AL AL 1,165,158 1,232,117 70,367 68,700 63,445 66,480 495,915 509,710 251,292 257,080 284,139 330,147 AL Core Avg. AL Non-core Avg AL Autauga N AL Bullock Y AL Butler Y AL Chambers Y AL Crenshaw N AL Elmore Y AL Lee Y AL Lowndes N AL Macon N AL Montgomery Y AL Pike N AL Randolph N AL Russell N AL Tallapoosa Y 25,445 9,426 12,783 N/A 4,796 8,038 N/A 18,781 48,423 N/A 8,405 63,536 9,076 5,574 11,291 9,095 26,656 10,848 15,566 N/A 5,201 7,788 N/A 19,723 52,793 N/A 6,969 65,368 12,318 4,857 14,531 9,061 1,458 507 796 N/A 241 493 N/A 857 1,954 N/A 335 4,577 264 448 690 624 1,500 476 655 N/A 217 589 N/A 1,416 2,537 N/A 403 3,901 471 198 652 341 1,133 390 581 N/A 340 471 N/A 911 1,326 N/A 229 3,116 271 281 588 635 1,276 426 832 N/A 209 451 N/A 1,433 1,970 N/A 145 3,107 320 128 703 484 9,491 3,914 6,118 N/A 2,247 3,777 N/A 8,659 13,182 N/A 2,485 24,934 3,009 2,472 5,486 4,149 9,588 3,975 6,858 N/A 2,426 3,509 N/A 8,100 13,762 N/A 1,683 25,706 3,083 2,577 5,673 4,026 5,039 1,908 3,060 N/A 1,223 1,919 N/A 5,067 6,542 N/A 1,188 13,150 1,535 1,203 2,555 2,332 4,866 2,213 3,843 N/A 1,226 1,637 N/A 5,191 6,333 N/A 1,319 12,475 1,582 1,061 3,259 2,333 8,324 2,707 2,228 N/A 745 1,378 N/A 3,287 25,419 N/A 4,168 17,759 3,997 1,170 1,972 1,355 9,426 3,759 3,378 N/A 1,123 1,602 N/A 3,583 28,191 N/A 3,419 20,179 6,862 893 4,244 1,877 GA GA 2,531,690 2,732,121 186,497 179,634 138,732 148,769 1,070,865 1,114,437 542,397 557,274 593,199 732,007 GA Core Avg GA Non-core Avg 16,897 6,425 18,184 6,846 1,220 525 1,554 497 993 1,283 479 307 7,299 2,933 8,031 2,855 3,974 1,449 3,650 1,814 3,411 1,038 3,666 1,373 GA Atlanta MSA 1,287,880 1,414,816 101,654 99,165 69,932 75,090 551,958 577,255 271,957 290,838 292,379 372,468 GA Harris N 7,109 8,551 458 579 480 501 3,188 3,329 1,604 2,240 1,379 1,902 GA Heard N N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A GA Meriwether N 5,714 5,033 494 478 495 220 2,679 2,142 1,187 1,341 859 852 GA Talbot N N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A GA Troup Y 16,897 18,184 1,220 1,554 993 1,283 7,299 8,031 3,974 3,650 3,411 3,666 GA Upson N 6,451 6,955 623 435 463 201 2,931 3,093 1,557 1,861 877 1,365 Notes: (1) This table contains the actual raw data of Table 3.4. (2) All data are from ACS 2005-2007 and ACS 2009-2011. Data of Bullock, Crenshaw, Lowndes, Heard, and Talbot are not available. (3) 15 counties are selected to create a smaller area of Atlanta MSA. Those counties are: Carroll, Cherokee, Clayton, Cobb, Coweta, DeKalb, Douglas, Fayette, Forsyth, Fulton, Gwinnett, Hall, Henry, Paulding, and Rockdale. 90 State AL AL AL AL AL AL AL AL AL AL AL AL AL AL AL AL AL GA GA GA GA GA GA GA GA GA Table B.5. Actual Change in Household Income County AL Core Avg Non-core Avg Autauga Bullock Butler Chambers Crenshaw Elmore Lee Lowndes Macon Montgomery Pike Randolph Russell Tallapoosa GA Core Avg. Non-core Avg Harris Heard Meriwether Talbot Troup Upson Core N Y Y Y N Y Y N N Y N N N Y N N N N Y N Median Household Income 2006 2010 40,052 41,973 38,692 39,838 33,147 36,624 48,052 53,471 N/A N/A 31,829 29,313 33,570 31,137 N/A N/A 50,675 54,075 38,849 41,231 N/A N/A 26,670 28,424 41,973 43,972 24,849 31,829 34,908 34,503 31,256 34,894 35,256 39,297 48,540 47,690 39,313 41,875 31,039 35,854 57,045 69,764 N/A N/A 35,560 37,569 N/A N/A 39,313 41,875 511 230 Mean Household Income 2006 2010 54,830 58,084 51,399 52,947 45,793 50,250 58,461 64,783 N/A N/A 41,342 39,843 41,119 42,243 N/A N/A 60,188 66,274 51,287 55,888 N/A N/A 42,622 40,258 60,857 60,284 41,147 45,988 46,687 53,747 40,050 46,476 53,602 53,147 65,227 65,279 48,639 53,651 40,014 45,537 72,053 88,164 N/A N/A 46,042 45,960 N/A N/A 48,639 53,651 1,948 2,486 Notes: (1) All the data are calculated based on ACS 2005-2007 and ACS 2009-2011 estimates. (2) Unit of all the numbers are percentages. Data of Bullock, Crenshaw, Lowndes, Heard, Talbot, and Atlanta MSA are not available. (3) 15 counties are selected to create a smaller area of Atlanta MSA. Those counties are: Carroll, Cherokee, Clayton, Cobb, Coweta, DeKalb, Douglas, Fayette, Forsyth, Fulton, Gwinnett, Hall, Henry, Paulding, and Rockdale. 91