Division of Economics
AJ Palumbo School of Business Administration
Duquesne University
Pittsburgh, Pennsylvania
STATE GOVERNMENT SIZE AND ECONOMIC GROWTH
A PANEL DATA ANALYSIS OF THE UNITED STATES
OVER THE PERIOD 1986-2003
Kristyn K. Brady
Submitted to the Economics Faculty
in partial fulfillment of the requirements for the degree of
Bachelor of Science in Business Administration in Economics
December 11, 2007
Faculty Advisor Signature Page
Antony Davies, Ph.D. Date
Associate Professor of Economics
STATE GOVERNMENT SIZE AND ECONOMIC GROWTH A PANEL DATA ANALYSIS OF THE UNITED STATES OVER THE PERIOD 1986-2003
Kristyn K. Brady, BSBA
Duquesne University, 2007
Abstract
There is a significant body of literature examining the effect of government size on economic growth in an international context. In this analysis, I narrow the field to include only the 50 states in the Union for the period 1986-2003 and measure the impact of state government size on individual states’ Real Gross Domestic Product. The size of state government is represented by a variety of factors including expenditures on consumption as a percent of Gross Domestic Product (GDP), transfer payments and subsidies paid out as a percent of GDP, the level of government employment, and the total tax burden using fixed effects to control for initial conditions within the states.
The results of this analysis indicate that increases in various determinants of government size slow economic growth within a state. The variances in the size of state governments raise important policy questions concerning the optimal size of government. The elasticities calculated from this model form a foundation for recommendations in future policy reform.
Key words: economic growth, government size, state government
Table of Contents
II. / Literature Review……………………………………………………………………. / 6
III. / Methodology……………………………………………………………………...... / 12
IV. / Results and Analysis………………………………………………………………… / 21
V. / Economic Implications………………………………………………………………. / 25
VI. / Suggestions for Future Research…………………………………………………….. / 26
VII. / Conclusion……………………………………………………………………...... / 27
VIII. / References……………………………………………………………………...... / 29
Appendices
AI. / Correlation Tests…………………………………………………………………… / 31
AII. / Stationarity Tests……………………………………………………………………. / 32
I. Introduction
The term “size of government” is somewhat vague, connoting different meanings to people of various educational and political backgrounds. Should the size of a government be determined by its geographic domain? Can government size be measured by the number of legislators currently employed? Definitions of government size vary widely across time and across countries. One of the most commonly-cited definitions of government size among economic researchers is the percentage of Gross Domestic Product (GDP) spent by the government on consumption.
Those in favor of government intervention and larger government typically cite the social benefits resulting from publicly provided goods. Proponents assert that the government’s scope allows it to perform certain functions more efficiently than private enterprise, such as infrastructure building financed via taxes, public education, and income redistribution programs. Those opposed to a larger government maintain that the government should exercise only limited power and authority over issues of property rights, protection of personal liberty, and other constitutional provisions. An empirical determination of the impact of the size of government requires a definition of not only government size but also a reasonable measurement of benefit. One such measure is the rate of economic growth in a country or in this case, a state, consistent with Ram (1986), Ghali (1999), and Conte and Darrat (1988).
This paper is an empirical analysis of the relationship between sizes of governments and economic growth. The intent is to determine how the size of a state’s government impacts the state’s economic prosperity. The method is to examine the effect of the size of US state governments on the respective states’ economic growths as measured by the rate of growth in state real Gross Domestic Product (RGDP). This model differs from existing research by focusing on public education expenditures and government transfer payments and subsidies as components of government expenditures, and by examining annual data across states rather than across countries.
II. Literature Review
There has been a significant body of work conducted on this topic on a cross country basis and on a state by state basis to a lesser degree. There is also no generally accepted modeling method. The model specification varies widely by study. The findings on economic growth primarily fall into two categories: Keynesian growth and Classical/Neoclassical growth. The former show evidence of government consumption expenditures as enhancing economic growth, where a greater proportion of expenditures relative to GDP speeds the pace of economic growth. Reasons for this include governmental influence in reconciling the differences between private and social interests, guidance toward a “socially optimal” growth path, and protection of the country against foreign exploitation (Ghali 1999). This view is supported by the findings of Ram (1986), Lin (1994), Ghali (1999) and Loizides and Vamvoukas (2005). Under a Classical growth assumption, government consumption expenditures hamper the path of economic growth. The underlying theory is that the collection of taxes to fund government spending crowds out both consumption and saving by the private sector. In addition, government spending introduces distortionary effects and incentives that serve to further destabilize the economy (Hyman, 2005). These effects are felt particularly with respect to income redistribution in the forms of welfare payments and subsidies. Another cited reason for a detrimental effect of government size is the inefficiency of government operations. Ghali (1999) gives the following examples:
Public investments undertaken by heavily subsidized and inefficient state-owned enterprises in agriculture, manufacturing, energy and banking and financial services have more often reduced the possibilities for private investment and long-run economic growth.
In the United States, these inefficiencies are seen in the subsidization of ethanol production, steel tariffs, and the government sponsored enterprises: Fannie Mae and Freddie Mac (White, 2007). Negative effects of government size on economic growth were found by Barro (1991), Gallaway and Vedder (2002), and Schaltegger and Torgler (2006).
a. Keynesian Growth
Ram (1986) found that the overall impact of government size on growth was positive. He used pooled cross-sectional data from 115 countries over the period 1960-1980. He measured economic growth by growth in total output expressed as GDP in international dollars. Lin (1994) found that differences in growth studies can be traced to model settings. He used OLS, 2SLS and 3SLS models to examine a 62 country data set over the period 1960-1985. Lin’s analysis shows evidence that government size as measured by the share of consumption expenditure relative to GDP has a positive impact on the economic growth as proxied by the annual rate of growth in real per capita GDP. These results hold in the short term under each model specification, but in the intermediate run (25 years) changes in government size had little or no significant impact.
Ghali and Loizides and Vamvoukas took the relationship a step further. As regression uncovers correlation between elements and not causality, these researchers were unsatisfied with only using regression in their analyses. Both used the Granger causality framework to determine the causation between government size and economic growth. Ghali (1999) used a vector error correction model to analyze quarterly data on ten Organisation for Economic Co-operation and Development (OECD) countries over the period 1970:1-1994:3. This study was not concerned with the impact of the allocation of government expenditures but only on the causal relationship between government size and economic growth. He found that government size, as proxied by the ratio of total government spending to GDP causes growth in GDP. Loizides and Vamvoukas’ (2005) analysis covers a forty year period across a three country sample: the United Kingdom, Ireland, and Greece. These three countries were chosen to provide a basis for comparison. At the time, the United Kingdom was considered to be “developed” by the authors while Ireland and Greece were considered to be “developing.” They found that the relative size of government as measured by real government expenditure as a percent of gross national product (GNP) fosters economic growth (real per capita GNP growth) in all three countries using a bivariate error correction model.
b. Classical Growth
Barro (1991) found that over 25 years, from 1960-1985, and across 98 countries that per capita economic growth was negatively related to the ratio of government consumption expenditures to real per capita GDP. Economic growth, the dependent variable, was defined as the growth rate of real per capita GDP over the period studied. This work used an endogenous growth model in which the chosen regressors impact not the long run level of output, but rather the slope of the long run output curve.
Gallaway and Vedder (2002) studied the effect of government size on economic growth on the national level in the United States over the period 1966-1998. The authors were concerned in particular with the “transfer state” or the level of income transfer facilitated by the government. This study regressed economic growth (real per capita GDP) on real per capita income transfers, average productivity of labor, and the unemployment rate. The authors used both level and first differenced models with comparable results. The authors concluded that as the “transfer state” expanded over the period studied, the deadweight loss to society grew and thus retarded economic growth. Specifically, in 1998 every additional $1 of income transfers increased deadweight loss by $1.90. These findings support the Misesian theory that transfers and subsidies are a negative-sum game rather than a zero-sum game.
Schaltegger and Torgler (2006) used data from all state and local governments in Switzerland to examine the claim that a negative relationship between government size and economic growth only applies in wealthy countries with large public sectors. They confirmed that over the period 1981-2001 GDP growth per capita was negatively affected by government expenditures as a percent of GDP. In Switzerland, a reduction in consumption expenditures by one percentage point of GDP related to a 0.06 percentage point increase in the economic growth rate.
c. Other Findings
Not all economic growth studies found significant effects, either positive or negative, of government spending on economic growth. Conte and Darrat (1988) found that the expansion of the public sector could not be held accountable for a slowdown in the economy. In their 22 OECD country data set, 15 countries exhibited no effect of government spending on economic growth[1]. Conte and Darrat defined the public sector as the “real value of government outlays, real tax collections, and changes in the real monetary base” all expressed as percentages of GDP. The authors studied the countries over the period 1960-1984.
Among the models, there is significant evidence, particularly in cross-country studies, supporting a convergence theory of growth. This theory emphasizes the importance of including initial levels of RGDP in an analysis when measuring various economic growth rates. Countries with low initial RGDP are likely to realize greater benefits from government expenditures due to “diminishing returns to capital as predicted in the neoclassical model” (Barro, 1996, 3). These countries are most likely to reap the benefits of direct economic help from the government while countries with higher initial RGDP see a decline in growth from additional government expenditure. This is likely because “as an economy prospers, the return on investment declines and the growth rate tends accordingly to decrease” (Barro, 1996, 3). A convergence theory is cited by: Barro (1991, 1996), Dowrick (1996), Gupta, Madhavan and Blee (1998) and Tomljanovich (2004).
Bania et al. (2007) applied a convergence theory approach to the United States on a state level with similar results. The authors concluded that there was evidence of the existence of “growth hills” in the economic development of states that mirrored the convergence effect in cross-country studies. This finding implies that states with lower levels of initial GDP benefit more from government intervention and grow more rapidly. Over time, diminishing returns to capital impart a negative effect on economic growth.
While most of the studies on economic growth have been conducted across countries, a few researchers have applied the framework to a single country analysis. Schaltegger and Torgler (2006) examine the impact of government spending on the economic growth of states in Switzerland, a wealthy country. This is an attempt to control for the convergence theory by limiting analysis to an already wealthy and economically developed country. The researchers examined all state and local governments within the country. Similarly, Loizides and Vamvoukas (2005) limited their analysis to just three countries, Greece, the United Kingdom, and Ireland, while Bania et al. (2007 ) and Gallaway and Vedder (2002) studied the just the United States.
There has been substantial research conducted on the impact on economic growth due to government spending on particular components. Transfer payments and subsidies are commonly examined in this context as such programs are often financed with tax dollars. Gallaway and Vedder (2002) studied the transfer state extensively, concluding that the effect of transfers and subsidies on economic growth in a region is negative. Likewise, Helms (1985) found that while productive end uses of tax dollars in the form of improved public services may increase the level of GDP output, tax revenues used to fund transfers tend to negatively impact economic growth within a state. They conclude that because of this, income redistribution should not be undertaken at the state level. Romans and Subrahmanyam (1979) also found that higher levels of income transfers within a state correspond to slower economic growth. Weber (2000) sought to explain the economic slowdown in the US during the 1970’s. He determined that the federal government’s fiscal policy change from purchases to transfer payments more than accounted for this economic slowdown.
Prior literature on this subject is divided on the merits of government size. While most of the literature thus far has been conducted on cross-country samples, there have been some studies performed within a country across state levels. These studies have supported a Classical theory of economic growth in which growth in government size is a detrimental force. The model in this paper breaks down government consumption into a variety of components such as education expenditure and transfers and subsidies as well as including the total tax burden as a source of government revenue. Significant effects resulting from any of these will help to define the relationship between government size and economic growth more precisely.