Is public expenditure productive? Evidence from the manufacturing sector in U.S. cities, 1880-1920.

AuthorYeoh, Melissa
PositionReport

This article provides the first examination of the relationship between public expenditures mad labor productivity that focuses on municipalities, rather than states or nations. We use data for 1880-1920, a period of rapid industrialization in which there were both high levels of public infrastructure spending and rapid growth of productivity. We use a simple Cobb-Douglas production function to model labor productivity in the manufacturing sector, letting total factor productivity depend on "productive" public expenditure by city govermnents--that is, on public spending that may raise the productivity of labor and encourage human capital accmmulation.

Using a data set of 45 of the largest cities in the United States, we find no statistically significant relationship between productive public expenditure and labor productivity in the manufacturing sector during this period. These findings are robust to three different econometric approaches. We do, however, find a strongly positive mad statistically significant relationship between private capital and labor productivity. Our results are consistent with those of much of the literature examining this same relationship in states and nations and they have important implications for contemporary public policy issues.

An Overview

The decline in labor productivity in the United States during the 1970s created a challenging puzzle for economists to solve. Aschauer (1989) found that public capital had a strongly positive relationship with productivity in the United States, and argued that the productivity decline had been caused by a decline in public expenditure on infrastructure. Munnell (1990) and others found similar results. These initial Findings were used by politicians and policymakers as evidence of the need for large increases in government spending on infrastructure. Some critics identified flaws in the econometric approach of this early work and, after correcting those flaws, found either a negative relationship or no statistically significant relationship between public capital and productivity. Peterson (1994) found that the marginal rate of return on public capital had declined substantially since 1950 and was substantially lower than that on private capital. He suggested that policies to increase private capital would contribute more to the growth of output than would the increases in public infrastructure recommended by Aschauer and others. The early literature on both sides of the debate is summarized in Munnell (1992) and Gramlich (1994). Few have been able to replicate the large effects found by Aschauer. Work in this area has slowed, but there remains no consensus (see, e.g., Kalyvitis and Vella 2011, and Lithgart and Suarez 2011). Moreover, virtually all of the existing literature has focused on national, regional, or state data and has analyzed contemporary time periods. While there is a substantial empirical literature investigating the relationship between local government spending and economic growth, (1) there appear to be none that examine the relationship with local labor productivity. (2)

One of Asehauer's (1989: 177) key findings was that "'a 'core' infrastructure of streets, highways, airports, mass transit, sewers, water systems, etc. has [the] most explanatory power for productivity." Since the bulk of that infrastnleture spending is done by local govermnents, we take a different approach than previous researchers and focus on local-level data and do so for a period of rapid industrialization, 1880-1920. During that time, there was a great deal of public expenditure on the construction of infrastructure and a rapid increase in the productivity of labor. (3) If public expenditure is positively associated with productivity, as some have claimed, then that relationship should be readily apparent in the data we have chosen.

The period between 1880 and 1920 saw tremendous growth in cities and wide variations in the labor productivity and public expenditure in areas across the United States. (4) The United States became a world leader in manufacturing during this period of rapid industrialization and much of the industrialization was correlated with city growth. Some cities recorded rapid population growth rates (for example, Detroit grew at an average of 73 percent per decade from 1880 to 1920), while other cities had slower population growth rates (such as Albany's average of 6 percent per decade).

Since manufacturing generated more than half of the total value of output in the United States by the late 19th century and was centered in the largest cities in the nation (Gallman 1960: 9.6), we focus our attention on whether public expenditure played any significant role in raising the labor productivity of manufacturing workers specifically. Labor productivity directly affected the profitability of manufacturing establishments and thus the overall economic growth of a city. Cities, in turn, were allocating large quantities of resources toward "public capital" such as roads, water supply systems, sanitation, education, and health. Furthermore, local governments were responsible for the bulk of government activity during this period, so focusing on city goveruments is most appropriate. (5)

Some argue that public expenditure, particularly in education and health, increases human capital and thus raises labor productivity in cities that invested heavily in such areas (Glaeser, Scheinkman, and Shleifer 1995). However, not every type of public expenditure will raise labor productivity. Some types of public expenditure, such as spending on the maintenance of public buildings and the salaries of city employees in the legislative and judicial branches of government, will not raise labor productivity in manufacturing. For that reason, we focus on productive public expenditure.

Economic theory posits that productive public capital--such as roads, water supply systems, sewers, education, and health--lowers the cost of doing business and raises the marginal product of other forms of capital. As a result we should see businesses flourishing in cities that invested heavily in infrastructure. For example, public expenditure on roads, bridges, highways, and waterways lowers the cost of transportation and facilitates the movement of goods and labor throughout the United States. (6) Public expenditure on education, health, sanitation, and water supply systems may increase human capital accumulation by making the labor force (or the future labor force, in the ease of school children) more literate and healthier. (7)

We can model the growth of a city using the augmented Solow growth model, assuming that the city is a small economy. This model suggests that physical and human capital accumulation should go a long way in explaining the differential income levels of cities. According to Barro (1997: 2) in his cross-country study of economic growth and convergence, "The concept of capital in the neoclassical model can be usefully broadened from physical goods to include human capital in the forms of education, experience, and health." The effects of physical capital accumulation (Romer 1986) and human capital accumulation (Lucas 1988) on economic growth are modeled and documented in many studies, such as Barro's (1997) cross-country empirical study and Barro and Sala-i-Martin's (1991) study of income convergence in U.S. states. (8) Stansel (2005 and 2009) found similar results for the relationship between human capital and the growth of population and employment in U.S. metropolitan areas.

Holtz-Ealkin and Schwartz (1995) and Ranch (1994, 1995) provide formal theoretical models of the relationship between public expenditure and productivity at the sub-national level that are closely related to the subject of this article. Those models come to opposite conclusions about that relationship. Holtz-Eakin and Schwartz's (1995) article develops a neoclassical growth model explicitly incorporating infrastructure investments and providing a tractable framework to empirically analyze the significance of public capital accumulation to productivity growth. Examining a panel of state data, they find no statistically significant relationship between public sector capital and the growth of productivity. Their results suggest that higher infrastructure outlays were not associated with a significant increase in productivity growth in U.S. states between 1971 and 1986.

Rauch (1994) develops a formal model to study the effect of municipal reform in the Progressive Era (from 1902 to 1931) on city governments' allocation of public expenditure and on city growth, using the rates of growth in manufacturing employment and value-added output as measures for city growth. He finds that city governments' expenditure on roads, sanitation, and the water supply system are statistically significant in explaining manufacturing employment growth in both panel and cross-sectional analyses. However, expenditures on roads and sanitation are not statistically significant in explaining growth in manufacturing's value-added output in the panel regression. (9)

Our dependent variable is based on the same data as those used by Ranch, but we use file level (rather than the growth) of the log of real dollar value added by manufactures per worker (rather than the total), that is, we use productivity not overall output growth. Our analysis differs from Ranch's in four important ways. First, we explicitly model labor productivity (not output growth) as a function of productive public expenditure. Second, we include education and health spending in our public expenditure measure so that it will capture those additional potential benefits to human capital and thereby productivity. Third, we examine an earlier time period, 1880-1920 (one that avoids the potentially contaminating effects of the Great Depression and that includes the last two decades of the 19th century, which saw large public...

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT