CHOICES
A publication of AAEA
The United States is blessed with exemplary federal data provision about local areas, such as counties, that we view as second to none. The data about our towns, cities, counties, metropolitan areas, and states have both augmented private sector productivity and produced more effective and accountable government. These data programs are now threatened by budget cuts to reduce the budget deficit, most immediately by the sequester process. We contend that the cuts are “penny wise, but pound foolish,” given their miniscule share of the total federal data budget, and argue that it is critical for these efforts to continue.
House Bill H.R. 1638, “Census Reform Act of 2013,” is one recent effort to balance the federal budget by curtailing public spending on an essential public service (U.S. Government Printing Office, 2013). It is also the tip of the iceberg of cutbacks leading to statistical data-related collection and reporting suspensions and delays currently underway in various federal agencies as a result of sequestration.
This article summarizes many of the proposed cutbacks and examines their economic consequences for the nation. We argue that federal statistics are essential for understanding key socioeconomic and demographic changes occurring at local and regional levels nationwide, and for analyzing the effects of local, state, and federal policy on people and businesses in U.S. counties and congressional districts. Indeed, state and local governments cannot perform their roles as “laboratories” that “try novel social and economic experiments” if successes and failings cannot be tracked, monitored, or evaluated.
If these data programs are discontinued, federal, state, and local policymaking will increasingly occur in the dark as policies are implemented in a vacuum of data and knowledge. At a more fundamental level, these data are vital to the existence of a well-informed population that is at the core of any democracy: How else will voters know if the policies put in place by their elected leaders work in their districts, regions, cities, towns, and communities?
The Association of Public Data Users (2013) summarizes the data collection efforts undermined by the sequester. For example, at the Bureau of Economic Analysis (BEA), important parts of the Local Area Personal Income program providing data at county and metropolitan levels will be eliminated. This is the only source of such data and, as we show below, vital for many critical economic analyses, including those related to understanding sources of job creation and economic development. In addition, the BEA’s Foreign Direct Investment Analytical Products and its Regional Input-Output Modeling System (RIMS-II) will be phased out due to the sequester.
Due to space limitations, we focus on cutbacks at the BEA, but also note other critical data programs at risk. At the Census Bureau, funding shortfalls threaten the timely delivery of key statistics as well as critical investments in the data collection infrastructure needed to keep costs low while ensuring data accuracy. The Bureau of Labor Statistics is no longer maintaining its International Labor Comparisons and Mass Layoff Statistics programs, among others. And there are perennial efforts to eliminate the vital American Community Survey (ACS) —the nation’s best source of demographic and income data for geographic areas as small as census tracts, which approximate neighborhoods.
And this list goes on. At the Energy Information Administration, the weekly petroleum and natural gas statistical reports are being delayed, and the Annual Energy Review and Energy Perspectives publications have been suspended. At the USDA’s National Agricultural Statistics Service, numerous statistical surveys and reports have been suspended due to sequester, with effects expected to be permanent in many cases.
At a time of global economic upheaval, major emerging challenges and critical changes in regional economic growth conditions related to new forms of energy development and other forces, make the timely and accurate collecting and reporting of data on states and localities more essential. Instead of cutting back, we argue that benefit-cost analysis would support a significant expansion of these investments in data collection, reporting, analysis, and public provision. These data are needed by local governments responsible for our nation’s business climate as well as by the businesses themselves to help our nation remain competitive.
For example, public data allows small firms to identify new business opportunities, thereby keeping the economy innovative and dynamic. There is compelling evidence that small, locally owned firms are key engines both of economic growth (Fleming and Goetz, 2011) and of job creation (Neumark, Wall, and Zhang, 2011); consequently, it is vital to provide the BEA statistics to all potential users, as a public good. Without public data, only large firms will know about new opportunities because they can afford to purchase or generate the data themselves. But large firms may not act on the information or they may suppress it to stifle competition; small firms would clearly be at a disadvantage. A more tenuous argument is that widespread reporting of key state and local variables is essential to safeguard against economic calamities that produce even higher public—and private—costs, such as the Great Recession of 2007-2009.
Furthermore, services that have compiled data are to be discontinued. Examples include the U.S. Statistical Compendium and USA CenStats Database (U.S. Department of Commerce, 2013a, 2013b). While the source data may be available in some cases, these discontinuations mean that each user must exert time and effort to acquire data. This duplication of effort is one reason why cutting the efficient federal data programs is ‘pound foolish.’ Now congressional staffers and federally funded researchers will have to duplicate several times the effort, indefinitely into the future, at a cost to taxpayers that far exceeds any budgetary savings.
Effective and accountable state and local government needs good data to develop plans and strategies to enhance economic well-being and competitiveness. Good data is necessary to provide cost-effective public programs and policies as well as to evaluate them.
Local Area Personal Income (LAPI) provides the only annual comprehensive and detailed measures of county level economic activity, transfer program payments, and personal well-being. The LAPI program has been the only source of consistent annual estimates at the county level covering farm as well as non-farm establishments, government, and private sectors at the detailed sectoral level. The LAPI program is the only source of annual estimates of the earnings of proprietors and commuters by place of residence. LAPI data is also particularly important due to its accuracy because it is developed from unemployment insurance, tax returns, and other administrative record data. And it is the only data reconciled to state and national totals.
The LAPI employment data by industry indicates the importance of sectoral activity to residents in an area. Public servants use the data on employment by sector to project infrastructure needs for water, highways, or energy, for example. Manpower experts use it to understand labor market trends and gaps. Private businesses use it to understand the market, and investors use it to identify new economic opportunities that range from housing and retail establishments to energy. And sectoral detail is essential for understanding how national business cycles both affect and are affected by economic conditions at the sub-national level (Goetz, Loveridge, and Albrecht, 2013).
Without LAPI data, it would be prohibitively expensive if not impossible for civil servants, analysts, or scientists to measure “how well are we doing” at the county level of disaggregation. It would undermine the ability of citizens in America’s cities, towns, and rural areas to understand where they are economically, to plan where they want to be, or to assess if their economic development efforts are having the expected effects.
The LAPI estimates of county per capita personal income are also used by the Census Bureau in its small area income and poverty estimation (SAIPE) to measure median household income. SAIPE provides timely estimates of income and poverty statistics used to administer federal programs and allocate federal funding to local jurisdictions. For thousands of non-metro communities that traditionally are home to the elderly, the LAPI’s annual estimates of transfer income by county are particularly relevant, just as they are for America’s war veterans who depend on veteran pensions—and who disproportionately reside in rural areas (Montgomery, 2012).
Annual estimates of transfer payment receipts are obviously essential for analyses of the effectiveness of anti-poverty programs at the local level. Officials also use the data to identify deserving people or businesses that are underserved, or to identify other problems such as fraud. Human services specialists consult these data to identify areas where training may bring workers into more productive work and where other economic development efforts may be successful.
Some publically provided alternative estimates to the LAPI’s sectorally detailed, county-level employment and income data exist, but each has shortcomings. Like the LAPI, the Quarterly Census of Employment and Wages (QCEW) prepared by the Bureau of Labor Statistics (BLS) starts with state and county estimates of all earned wages and salaries based on administrative records, including quarterly unemployment insurance (UI) contributions filed by employers with state employment security agencies. However, QCEW documents only employees in industries subject to state UI laws, and it captures only about 47% of total personal income.
Another source, the Census Bureau’s County Business Patterns (CBP) program, provides essential counts of establishments, employees (jobs), and employee compensation by sector at the county level of geographic disaggregation. However, CBP data excludes the self-employed and proprietors who own businesses, regardless of how many employees they have. Also, CBP earnings data are based on the place of employment and we do not know where those earnings go, or which market areas they benefit. Further, CBP data are a snapshot of employment during just the week of March 12. This is a problem for industries with seasonal employment, such as tourism and some food processing activities.
The Bureau of Labor Statistics (BLS) provides monthly Current Employment Statistics (CES) for metropolitan areas only that are survey-based and thus not as accurate as the annual LAPI data. Unlike LAPI, the CES also excludes agriculture, forestry, fishing, and hunting; private households; the military; and self-employment. Most significantly, the BLS’ CES data are not available for rural counties.
Unlike these other data sources, each of which is important in their own right, the LAPI estimates of compensation by sector measure the full scale and industrial mix of an area’s economy. In the LAPI all jobs are counted, even multiple jobs held by the same person, which is important as part-time employment has grown (U.S. Bureau of Economic Analysis, 2012). Because self-employment is widespread in low population density areas, the BEA's LAPI data is essential for the analysis of rural economies.
There is no alternative to the comprehensive LAPI data that enables us to know how prosperous or poor people are in different localities, or what the local market potential or tax bases are, for example. Losing the LAPI sectoral employment data would force rural area analysts to prepare their own employment and earnings estimates, using data they must acquire from state employment security agencies on their own. When analysts nationwide have to construct their own estimates, and at questionable quality, costs of the duplicative effort will by far exceed the savings from cutting the single federal LAPI program at the BEA.
Open, accountable government and high-quality private sector and academic research require good data. Unfortunately, we cannot expect private vendors to adequately fill the void if the federal government scales back its efforts to provide state and local data. The main reason is that the data has characteristics of what economists call a “public good.” These characteristics are: (1) once they are produced, everyone has access to the goods; that is, they are non-rival; and (2) it is not possible to properly charge people for the use of the good because once they are produced, people receive the benefits of the good whether they pay for it or not; that is, they are non-excludable. Private markets cannot efficiently operate in such an environment and would under-produce public goods in the absence of government intervention. National Defense is the classic example of everyone benefitting from its production and where a private company would be unable to adequately charge for its provision. In the case of data, even with strict confidentiality agreements, data and information could leak out to the broader public, reducing the return to private vendors. Hence, if left to the private vendors, economic data would be under-produced because firms could not exclude nonpayers—so-called free riders.
Left to the private sector, the loss of free, high quality, government-provided data will negatively affect new and small businesses more than larger companies that may be able to pay for the data needed to make business decisions. Both large and small businesses in rural towns and smaller metropolitan areas, however, are likely to be disadvantaged by private provision of local area data. Private sector data companies may not document small areas at all, instead focusing on larger metropolitan areas where there is sufficient demand to cover the costs of provision. It costs about the same to prepare a large county’s data as a small one, but the larger county contains more customers and potential revenue. If private vendors do not receive enough revenue to cover costs, they will not provide data about rural and small metropolitan areas at all.
The lack of data about local area sectoral employment patterns and local area income would also undermine the competitiveness of rural and small town America in attracting businesses. Firms consult such information to select new establishment sites. A company wants to know about the labor market, demographics, and the income in an area before it opens a new establishment. Opportunities in rural towns may be overlooked simply for a lack of information. Likewise, rural and small metropolitan area governments will not be able to evaluate projects or programs if the data is not available. The risk of embracing sub-optimal policies would thus be greater in the smaller area, further undermining small town place-competitiveness.
Furthermore, public provision of the fundamental data enables private businesses that process and add value to the data. Examples of the value added by private firms to public data include economic forecasting, the construction and application of economic models, and the provision of very specialized data for specific data users. If high-quality state and local data were no longer federally provided, these firms will have to spend significant resources to collect and clean data. The quality of the private data products is likely to decline as well as rise in cost which will be passed on to customers, making the data less accessible to smaller governments and businesses.
Private vendors may be reluctant to acknowledge that the quality of their products will be compromised by the loss of publically available, high quality, primary data from the federal programs. Compounding the problem, the lack of publically available data will make it impossible to verify the content of private data products. The private products will have a greater “black-box” feel, making customers more skeptical and less likely to purchase them. In sum, private companies that attempt to fill the void of free government data not only face the “public good” challenges noted above, but also face reduced demand.
The ACS and BEA LAPI data are foundational in examining state and local socioeconomic outcomes in academic and policy research. Desmet and Fafchamps (2005) use LAPI data to better understand which places are creating more jobs and whether certain industrial structures are particularly important. Hence, such research supports strategic planning to enhance local and regional competitiveness. The LAPI sector detail under threat of discontinuation is crucial because industrial structure plays a key role in affecting whether low-skilled workers can find work (Partridge and Rickman, 2006). Going forward, ACS will become even more crucial, because it will be the only source of local data on demographics and income dynamics (Peters, 2012, is a good example).
Evaluations of specific government programs often rely on LAPI data, and will increasingly rely on the ACS as the replacement for the Decennial Census detailed income and demographic data, given the loss of the long form from the 2010 Census. Stephens and Partridge (2011) heavily use LAPI information in their assessment of the role of self employment and small business development as contributing to economic prosperity of the Appalachian region. Their work also underscores how American work patterns have undergone a dramatic yet rarely noticed transformation over the last four decades: the rate of self-employment has skyrocketed recently from one in eight to nearly one in four workers. We know about this transformation thanks to data provided by BEA’s LAPI program (Goetz, Fleming, and Rupasingha, 2012).
In another example, Kandilov and Renkow (in press) apply modern program evaluation methods using BEA’s LAPI data to determine if broadband loans made a difference in agricultural commodity sales at the county level. With BEA data they were able to determine that even though agricultural producers benefitted from the broadband program, other sectors may have been weakened. These nuanced yet important details are needed for effective policymaking about broadband—and they require data of sufficient geographic and industry detail.
Putting together spatial data from different federal agencies, Loveridge and Reimer (2013) show a high association between county-level Supplemental Nutrition Assistance Program (SNAP) payments and rates of diabetes among the population. They suggest a new policy experiment that could save taxpayers vast amounts of money if successful, and more than enough to pay for the data collection efforts many times over. Knowledge about how changes play out in different counties and congressional districts is critical for Americans and their elected representatives if the nation is to continue to prosper.
LAPI data along with other federal data have been essential for teaching the next generation of business location specialists, economists, planners, and policy experts. Due to its availability, ease of use, and high quality, students use this data to learn about the industry structure of particular communities, how to forecast future economic growth, and how to evaluate policy. For example, students of regional and urban economics use this data to statistically understand if tax policies and incentives affect the location of business activity, or whether improving educational attainment will attract higher-paying jobs or new industries. Students also use the data to learn how to prepare for and mitigate the impacts of natural disasters, or to understand the role of transfer payments. Using real data, students can construct plans for business start-ups or develop plans for energy or water use. Planning students use this data to understand how infrastructure placement affects industry structure and energy demand. Because American students have had real world data, they have been immediately employable and their training is much more relevant than if they had to use contrived data or data from other countries.
We are at a crossroads in terms of potentially losing a relatively low-cost but vital resource for understanding fundamental economic changes and adjustments underway in the over 3,000 U.S. counties that make up the United States of America. Facing sector- and region-specific economic and environmental forces that are without parallel in history, our need to understand how different regions are impacted has never been greater. It is one thing to track national-level macroeconomic indicators about the business cycle, the price level, and interest rates. It is quite another to know how these key variables are playing out in the different sectors and regions—rural and urban. How do the performances of specific economic sectors in these sub-regions contribute to (or detract from) national economic well-being? Without detailed data we will not know which sectors in different regions are leading the rebound, and which are holding it back.
Lacking state and local data, it will be harder for start-ups or new, high-growth companies to identify market opportunities. Governments will become less accountable, and policy successes (or failures) will be less transparent to voters. Without data on employment and income we cannot determine if particular policies are working: contributing to job growth and regional development, or just wasting tax dollars.
Defunding these data programs today may save taxpayers pennies in the short run, but it will cost hundreds of dollars in terms of direct outlays to produce alternatives, in waste, and in terms of undocumented local economic opportunities foregone. Spending will be duplicated at the local and state levels to construct alternative data sets, and business opportunities missed. Both market and policy failures are likely. The nation cannot afford to lose these data programs.
Association of Public Data Users. (2013). Sequestration-Related Budget Cuts; Regional, State, and Local Impacts. Arlington, VA: Association of Public Data Users. Available online: http://apdu.org/uncategorized/sequestration-related-budget-cuts-regional-state-and-local-impacts/)
Desmet, K. and Fafchamps, M. (2005). Changes in the spatial concentration of employment across US Counties: A sectoral analysis 1972-2000. Journal of Economic Geography 5(3), 261-84.
Fleming, D. and Goetz, S.J. (2011). Does local firm ownership matter? Economic Development Quarterly, 25(3), 277-81.
Goetz, S.J. (2008). Self-employment: The new economic reality. Rural Realities, vol. 2, no. 3. Available on-line: http://www.ruralsociology.org/wp-content/uploads/2012/03/Rural-Realities-2-3.pdf.
Goetz, S.J., Fleming, D., and Rupasingha, A. (2012). The economic impacts of self-employment. Journal of Agricultural and Applied Economics, 44(3), 315-321.
Goetz, S.J., Loveridge, S., and Albrecht, D. (2013). Where the jobs went after 2007. The National Agricultural and Rural Development Policy Center, Data Brief No. 1. Available online at http://www.nardep.info/uploads/DataBrief1_Jobs.pdf
Kandilov, I.T. and Renkow, M. (2013, in press). The impact of the USDA broadband loan program on U.S. agriculture. The National Agricultural and Rural Development Policy Center, Policy Brief.
Loveridge, S. and Reimer, A. (2013). Diabetes and SNAP: An opportunity for change? The National Agricultural and Rural Development Policy Center Data Brief No. 3. Available online at http://www.nardep.info/uploads/Brief_DiabetesSNAP.pdf
Montgomery, S. (2012) Characteristics of rural veterans: 2010 – Data from the American Community Survey, Washington, D.C.: National Center for Veterans Analysis and Statistics, July. Available online at http://www.va.gov/vetdata/docs/SpecialReports/Rural_Veterans_ACS2010_FINAL.pdf.
Neumark, D., Wall, B., and Zhang, J. (2011). Do small businesses create more jobs? New evidence from the National Establishment Time Series. Review of Economics and Statistics, 93, 16-29.
Partridge, M.D. and Rickman, D.S. (2006). The geography of American poverty: Is there a role for place-based policies?, Kalamazoo, Mich: W.E. Upjohn Institute for Employment Research.
Peters, D. (2012). Understanding micro and meso scale income inequality in the Midwestern United States, 1979-2009. Rural Sociology, 77: 171-202.
Stephens, H. and Partridge, M.D. (2011). Do small businesses matter for economic growth in Appalachia?” Growth and Change, (42): 431-465.
U.S. Bureau of Economic Analysis. (2012). Local Area Personal Income and Employment Methodology. Washington, DC: Bureau of Economic Analysis. Available online: http://www.bea.gov/regional/pdf/lapi2011.pdf
U.S. Department of Commerce. (2013a). USA CenStats Database. Washington, DC: U.S. Department of Commerce. Available online: http://censtats.census.gov/usa/usa.html
U.S. Department of Commerce. (2013b). What is the Statistical Abstract? Washington, DC: U.S. Department of Commerce. Available online: http://www.census.gov/compendia/statab/
U.S. Government Printing Office. (2013). HR1638. Washington, DC: U.S. Government Printing Office. Available on-line : at http://www.gpo.gov/fdsys/pkg/BILLS-113hr1638ih/pdf/BILLS-113hr1638ih.pdf.
