Despite several decades of U.S. anti-poverty policies, up to 13 percent of the population still lives below the poverty line, and many areas, such as Appalachia and the Mississippi Delta consistently, have significantly larger percentages of residents living in poverty. The economic downturn affecting the U.S. from 2008 through 2010 requires identification of the kinds of projects that have the most positive impact on the economy and on areas that have the greatest need for public assistance. This research project will complete the first in-depth analysis of the role of local, state, and federal government spending on the economic development of U.S. counties from 1996 to 2007. Using a unique dataset that captures the counties' economic characteristics and public spending, the investigator will identify what types of investment (such as education, transportation, business support, human services) are the most effective in promoting growth among counties. The novelty of this approach lies in assessing whether public programs that are not designed for economic development purposes may also affect the economy of a county. It also will account for the presence of spillover effects that take place when public spending is allocated to a recipient county by means of spatial statistics and spatial econometrics. An example would be the construction of new transportation infrastructure that will impact the economy of the recipient county as well as that of its neighbors. Previous empirical estimations of the impact of public policies have failed to account for this type of spatial dependence. The investigator also will explore the potential presence of a "crowding-out" of local and state expenditures by federal spending. Identification of circumstances where the volume of federal expenditures becomes overwhelming could help support changes in the design of federal programs to enable funds from all levels of government to be used most effectively. The investigator also will test for the potential endogeneity of public spending and the robustness of the results to different spatial scales.
This project will assess the efficiency of public policies in promoting economic growth across counties. Project results will suggest what types of projects have the greatest impact on economic growth. Particular attention will be devoted to measuring the role and the extent of spillover effects in promoting growth. Information and insights from this project therefore will give policy makers a better understanding of how investments in one place diffuse to neighboring places. This project will lead to new insights regarding the ways allocation criteria should be defined and recipient areas assisted so that programs such as the American Recovery and Reinvestment Act and future regional development policies can become more efficient.
This project provides new insights on the impact of federally-funded investments on the growth patterns of the US states and counties. Their capacity to promote growth has been hardly studied in the literature even though such expenditures are a source of intense discussions among economists and policy-makers as seen in 2009 during the implementation of the Obama administration’s stimulus package. Our results are relevant to the regional economies of the U.S. and to the federal government since the country is running a budget deficit and the aftermath of the recent economic crisis is still observable. The intellectual merits and broader impacts of our project are numerous. First, we have created the most complete and recent dataset of federal programs for regional development purposes. We have relied on data from the Census Bureau’s Consolidated Federal Funds Reports that we have classified according to 15 axes of "place-based" and "people-based" development (e.g. spending for education, for transportation, for the unemployed). The use of actual spending data has led to new insights compared to previous contributions. The latter use proxies or limited datasets that, we believe, explain the level of uncertainty about the role of public intervention. This classification has allowed us to reach our second objective, namely identifying the most efficient development programs. We use an instrumental variable approach to control for the endogeneity of public funding and a beta-convergence framework to model the growth dynamics. While overall federal spending is found to support regional economic growth, we are more interested in uncovering the role of each specific program. We find that programs for health purposes promote growth across all counties while support to farmers has a positive impact in the non-metropolitan counties only. On the other hand, programs that support the poor and education have a negative impact on growth. The former result is due to the counterproductive nature of this type of transfer while the latter result comes from federal support for education being targeted to poor, slow-growing, counties. The federal government ought to follow this approach if it wants to achieve greater growth at the least cost, improve how its activities are carried out and justify its choosing of future development strategies. Yet, we recognize that measuring the impact of federal spending on other objectives, such as social inequalities or labor productivity, is necessary before we can righteously advocate for an elimination of some programs. Our third objective consisted in testing the sensitivity of our results to public spending classified by agency or department. We find that the U.S. Department of Agriculture is the only one whose activities promote growth significantly. We believe it comes from its programs focusing on farmers and rural areas only. Other departments may be spreading their efforts to too many objectives. Our work accomplished another, fourth, objective in measuring the spillovers of economic development. Public investments affect economic growth in the recipient area and beyond. Yet, the literature still treats counties as isolated entities. We have remedied to this problem by estimating a Cobb-Douglas production function where the technological progress encompasses elements from various economic growth theories including spillover effects. Our results indicate that federally-funded capital has a positive impact on income and on growth but only when spillover effects are considered. In addition, when private capital increases in a state, it promotes growth in neighboring states. From a policy perspective, these results suggest that supporting a region encompassing multiple counties could be more relevant than dealing with each county individually. Finally, the fifth accomplishment of this project has been to challenge the traditional idea that federal spending crowds-out state/local spending. Our results indicate that a significant crowding-in takes place overall and in the programs supporting transportation, energy, the unemployed and businesses. On the other hand, programs for the poor and the low-income suffer from a crowding-out effect where any extra federal dollar leads local authorities to reduce their spending by 10%. Our results support changes in the design of federal programs such as a reduction of the matching requirements in the case of crowding-in and an increase in the limit of locally and state-provided funds (called the cap) in case of crowding-out. This project has provided financial support to three graduate students and has constituted the PhD-thesis of one of them. As of today, the project has led to two articles published as chapters in peer-reviewed international books, one submission in a journal and one article in preparation. The research results have been presented in fifteen venues, ranging from professional meetings to presentations at the Economic Development Agency. Finally, a webGIS that displays the data and some regression results is available online at the following address: http://webgis.arizona.edu/reis/. Publications, presentations and website provide the opportunity for further work aiming at maximizing the impact of future public interventions on regional growth.
