Printer Friendly

Procyclical state spending in the USA: the impact of political ideology.


Government spending is procyclical when expenditure increases at a faster rate than income during an economic upturn and decreases at a faster rate during a recession. While economists anticipate countercyclical spending (Alesina et al., 2008), there is growing evidence that spending is procyclical (Gavin et al., 1996; Kaminsky et al., 2004; Talvi and Vegh, 2005; Woo, 2009). In this literature, studies focus on central government spending (Sorensen et al., 2001 is an exception).

In this paper, the objective is to focus on sub-central governments' motivation to spend procyclically. Procyclical spending can maximize a community's welfare (Alesina et al., 2008; Abbott and Jones, 2011) but it is not obvious that sub-central governments choose to spend procyclically to maximize welfare. With evidence that sub-central government spending is procyclical in the USA, this paper considers two hypotheses to explain the presence of procyclical spending.

The first hypothesis is that sub-central governments are more likely to spend procyclically if this is what voters prefer. The more local politicians compete for votes the more they are likely to respond to the preferences of the median voter. The implication is that the pattern of cyclical sub-central government spending across US states will be correlated with the pattern of voters' preferences across US states. This prediction can be tested by measuring differences in voters' preferences on a scale set by the ideological spectrum that separates the Republican Party and the Democratic Party. Empirical studies report the choices that parties have made with reference to their different political ideologies. In this paper, the intention is to focus on Republicans' and Democrats' spending responses to changes in state GDP. Is it possible to explain different responses with reference to different preferences?

The second hypothesis is that differences in sub-central governments' willingness to spend procyclically depend on differences in the availability of finance. Studies that explore the determinants of central government procyclical spending highlight the importance of 'voracity effects' (Lane and Tornell, 1996; Tornell and Lane, 1999). 'Voracity effects' occur if competition for increased public spending between different groups in the community intensifies as national income increases. This paper focuses on competition between state governments for intergovernmental transfers (competition that takes the form of a zero-sum game). If competition intensifies as national income increases, it is also the case that 'recessions have a chilling effect on fiscal competition.' (Lane, 2003, p. 2665). As competition is likely to increase when there are a large number of claimants, 'voracity effects' are increasingly likely in more federal systems of government. A recent study of procyclical spending in the OECD indicates that voracity effects and federal governments' willingness to provide procyclical intergovernmental transfers are relevant (Abbott and Jones, 2011).

Sub-central governments are more likely to spend procyclically if intergovernmental transfers are procyclical. The question is whether differences in procyclical spending across different states depend on differences in the influence states are able to exert (via different political networks). The evidence is that political networks are important when explaining the allocation of intergovernmental transfers (see Kehmani, 2004 for a survey of this literature). This paper explores the possibility that the likelihood of sub-central procyclical government spending depends on the existence of a coincidence between the political affiliations of the incumbent US President and the incumbent state governor. The second testable prediction is that sub-central governments are more likely to spend procyclically if they are able to rely on a 'closer' political network between the state and the federal government.

To begin, it is necessary to estimate the extent to which US states' fiscal activities are procyclical. Are all state expenditures procyclical? Are states' own tax revenues procyclical? With this information it will be possible to test the two hypotheses.

In the next section of the paper the objective is to present testable predictions for each hypothesis. Section 3 of the paper describes the data that is to be tested and the tests that are to be employed. Section 4 reports the estimation results. Section 5 presents the conclusions.


It is possible to define the set of circumstances in which procyclical spending maximizes welfare but this set is quite specific and quite narrow. Even if procyclical spending might increase the community's welfare, this might not be the government's motivation to spend procyclically. When governments are motivated to spend procyclically, it is not necessarily the case that procyclicality will increase the community's welfare. So, why do sub-central governments choose to spend procyclically?

When markets are working perfectly, the welfare-maximizing direction of a change of government spending in a neoclassical model depends on the degree of substitutability between government and private consumption. If they are substitutes, government expenditures should be countercyclical. If they are complements, a benign government might increase spending in a procyclical fashion (Lane, 2003). Of course, if there is market failure, procyclical spending might also maximise a community's welfare. Focusing on developing countries, Alesina et al. (2008) suggest that procyclicality occurs because there are imperfections in financial markets. They note that " bad times, many developing countries cannot borrow, or can do so only at very high interest rates, therefore they cannot run deficits and have to cut spending; in booms, they can borrow more easily and choose to do so, increasing public spending...." (p. 1007).

While these arguments are important when justifying procyclical spending, the objective here is to explain the motivation to spend procyclically. This section of the paper compares two hypotheses. The first focuses on different political ideologies. The second focuses on different political networks.

2.1 Procyclical Spending and Political Ideology

An established literature demonstrates that welfare gains derived from fiscal federalism depend on differences in voters' preferences in different jurisdictions (Oates, 1972). The gains from decentralization are higher the more that citizens are able to 'vote with their feet', i.e. by moving to sub-central jurisdictions that offer a closer match to their fiscal preferences (Tiebout, 1956). If the expectation is that levels and compositions of sub-central government spending differ across sub-central government jurisdictions, the expectation is also that procyclical spending will differ across sub-central government jurisdictions.

Differences between political parties' willingness to spend as national income changes are well documented in the 'political business cycle' literature. When researchers analyze politicians' willingness to spend, they distinguish between statements made by the Republican and Democratic parties and between actions taken by the Republican and Democratic parties. Studies indicate that the Democratic Party is more willing to increase government expenditure to pursue social goals, for example to provide social services (see Mueller, 2003 for a survey of this literature). The federal government has the responsibility to manage the macroeconomy (Oates, 1972). When the Democratic Party is in office the government is far more willing to increase government spending (to achieve social goals). When the Republican Party is in office the government is more inclined to argue that private spending is more efficient than public spending, and the government is more sensitive to the risks of inflation if national income is increasing (Mueller, 2003).

With different ideologies, voters are able to predict that Democratic sub-central governments are more likely to increase public expenditure when national income increases. By comparison, both parties are sensitive to the electoral costs they will incur if there is a sub-central budget deficit. Sub-central governments are not expected to borrow to act counter-cyclically (as this role is assigned to central government). They would find it more difficult to borrow than central government (as they would be in competition with each other), and in the USA there are restrictions on sub-central government borrowing (Sorensen et al., 2001). The implication is that differences in political ideology (between the Republican and Democratic parties) are relevant when voters express their preferences. Differences in estimates of cyclical government spending across different sub-central governments may reflect differences in voters' preferences.

If the political process works 'efficiently' (in terms of representation), the testable prediction is that government spending is more likely to be procyclical when the state governor is a Democrat than when the state governor is a Republican.

2.2 Procyclical Spending and Political Networks

The second hypothesis suggests that there may be 'failings' in the political process and the likelihood of procyclical spending depends on sub-central governments' capacity to respond to 'voracity effects'. The literature on pro cyclical spending highlights the importance of 'voracity effects', which occur when competition between groups for increased government spending intensifies as national income increases. Studies that focus on central government spending highlight the importance of voracity effects that are particularly relevant when government institutions are weak and when there are significant differences between the preferences of different groups in the economy, for example between the preferences of producer groups, consumer groups and ethnic groups (Akitoby et al., 2006).

Here the focus is on sub-central government spending and it is impossible to ignore the competition that exists between local politicians for intergovernmental transfers. When local politicians 'win' intergovernmental transfers they are also able to increase electoral support (by increasing local expenditure without increasing local taxation--Mueller, 2003). (1) With this incentive there is a 'common pool' problem and there are 'voracity effects' as national income increases (Abbott and Jones, 2011).

Studies suggest that local politicians are more likely to receive intergovernmental transfers when they are members of the same political party as the incumbent US president (a survey of this literature is presented by Larcinese et al., 2006). Cox and McCubbins (1986) argue that this ideological relationship proves important because the objective is to maximize presidential support by the use that is made of these funds. (2) In this context, it has been argued that intergovernmental transfers prove more 'efficient' if they are targeted on government supporters ("....core supporters can be targeted in a more efficient way because parties know their preferences better"--Larcinese et al., 2006: 448).

If levels of spending across US states are influenced by the availability of funds from federal governments, changes in the levels of spending across different US states (as GDP changes) are also likely to depend on differences in the availability of intergovernmental transfers. The question is whether the likelihood of procyclical spending is sensitive to the existence of this preferential network (when the state governor and the president are members of the same political party). (3)

Arena and Revilla (2009) test the proposition that differences in the cyclicality of spending across different states in Brazil depends on a coincidence between the political affiliations of the president and the state government. They use "... the number of years (that).... the same political party held both the presidential and gubernatorial offices ..." and they report that "... this variable has a significant impact on the coefficient of current expenditure procyclicality ..." (p.20).

In this paper the hypotheses are: (i) that differences in procyclical subcentral government expenditures in the USA depend on differences in voters' preference and (ii) that differences in procyclical sub-central government expenditures in the USA depend on differences in the political networks that are available when competing for intergovernmental transfers.


Three testable predictions are presented to distinguish between the two competing hypotheses. These predictions are: i) there are differences in the cyclicality of government spending across the US states; ii) the likelihood of procyclical spending depends on the political ideology of the state governor; iii) the likelihood of procyclical government spending depends on the political network that exists between the state governor and the US President.

To investigate the cyclicality of state level spending, we analyse data from the US Census Bureau's publication The Annual Survey of State Government Finances. As well as aggregate values, this publication provides detailed statistics on spending disaggregated by object. Total spending can be split into intergovernmental expenditure and direct expenditure, the latter including Capital Expenditure; Current Expenditure; Assistance & Subsidies (4); Insurance Benefits & Repayments; Wages & Salaries; and Interest Payments. By estimating the cyclicality regression equation for aggregate spending, as well as its components, we hope to ascertain whether some components are more procyclical than others, and whether some components are acyclical or counter-cyclical. All figures are denominated in current US dollars and are collected for the 50 US states plus the District of Columbia, over the period 1963 to 2006. We consider the spending of state government and local government together. (5) As well as aggregate spending and its components, we utilize data on government revenue (both Intergovernmental Revenues (6) and General Revenue from Own Sources) to identify whether the funding of government spending exhibits the same pattern of cyclicality as expenditure itself. Finally, we also test for the cyclicality of government borrowing.

To derive the estimated cyclicality coefficients two econometric specifications are used:

[DELTA][] = [[alpha].sub.i] + [[lambda].sub.t] + [beta][DELTA][] + [[epsilon]] (1)

[DELTA][] = [[alpha].sub.i] + [[lambda].sub.t] + [delta][DELTA][] + [beta][DELTA][] + [[pi].sub.1][] + [[pi].sub.2] [] + [[epsilon]] (2)

for i=1, ..., N and t=1, ..., T

where [] denotes the log of the government spending component for state i at period t, y is the log of state Gross Domestic Product, [[alpha].sub.i] are individual effects and [[lambda].sub.t] represents the time dummies. The individual effects control for any cross-sectional variation, for example the relative size of the state, while the time fixed-effects control for aggregate time-series variation, such as shocks to national GDP or inflation shocks. State GDP data comes from the US Bureau of Economic Analysis and is presented in millions of US dollars. The first specification (based on Lane, 2003), provides an index of cyclicality. With log-differences, [beta] is the elasticity of government expenditure with respect to output growth. [beta]<0 indicates government spending is counter-cyclical, [beta]>0 suggests government spending is procyclical, while a value greater than unity implies a more-than-proportionate response to output fluctuations. A value of [beta] greater than unity is consistent with the 'voracity effect'. This version of the cyclicality equation is estimated using the fixed effects estimator.

The second specification extends (1) to include a lagged value of the dependent variable, as well as lagged levels of government spending and output. This Error Correction Model (ECM) specification, while also producing an index of cyclicality, controls for a potential dynamic adjustment process in government spending and can also be used to model the long-run relationship between income and government spending, as predicted (for example) by Wagner's Law (Wagner, 1911). A similar specification has also been estimated by Akitoby et al. (2006) and Abbott and Jones (2011). Equation (2), which is our preferred specification, is estimated by a system GMM dynamic panel data estimator (Blundell and Bond, 1998). This methodology has two advantages. Firstly, equation (1) includes a lagged dependent variable as a regressor, [DELTA][], and estimating it with a standard panel data estimator would produce a biased estimate of the autoregressive component (Nickell, 1981). Secondly, it could be possible that government spending and one or more of the regressors are simultaneously determined, thus leading to a breakdown of the assumption that the estimated residuals and regressors are orthogonal. The SYS-GMM estimator partially overcomes the endogenity problem through the simultaneous estimation of level and first difference equations, whereby all the regressors of the level equation are instrumented using lagged first differences and lagged levels act as instruments for the explanatory variables in the first difference equation.


In this section of the paper the first set of results describe the pattern of procyclical sub-central government fiscal activity in the USA. The second section presents the results that are used to explain why some states are more likely to spend procyclically.

4. 1 Procyclical sub-central government fiscal activity in the USA

Table 1 presents estimates of the cyclicality coefficient, [beta], for our full sample of observations. The estimation results point to the procyclicality of sub-central spending: [DELTA][] is individually significant in all four cases and all P coefficients are positively signed. The mean coefficient is 0.14. Thus a 10% change in state-level GDP is expected to increase public spending in the states by on average 1.6%. Total sub-central spending also includes debt interest payments, but when they are excluded, primary expenditure is also found to be procyclical, slightly more so than total sub-central spending. (7) Government spending can be funded by revenue (both own sources of government revenue and intergovernmental transfers) and government borrowing. Total government revenues are strongly procyclical, even more so than spending, with a mean coefficient of 0.46. This result might be explained by the fact that some taxes are proportional to income earned (e.g. corporate income tax) and individual income taxes are typically progressive (Sorensen et al., 2001). Subcentral government borrowing is counter-cyclical, implying a degree of prudence to the states' government finances: positive shocks to state GDP stimulate government spending but they also raise government revenues at an ever faster rate, allowing a control on borrowing during economic upturns.

These findings are consistent with the findings of Sorensen et al. (2001) who investigated the cyclicality of state-level fiscal policy in the USA. Over the period 1978-1994, they found that budget surpluses were procyclical, again because of strong procyclical revenues and weak procyclical expenditures. Using annual changes in spending and time fixed effects their results suggest a 100 dollar increase in per capita state GDP will raise the budget surplus by 3.12 dollars. Our results suggest a 10% rise in state GDP will raise the budget surplus by 3.38%, averaged across the four estimates. Lane (2003), in a cross-country study of cyclical fiscal policy, found total government spending for the US was marginally counter-cyclical.

To understand these results and to identify the potential sources of procyclicality, consider the main components of total expenditure and total revenue. Cyclicality coefficients are derived for intergovernmental expenditure and direct expenditure, as well as the components of direct expenditure referred to in table 2.

These new estimates are presented for the [beta] coefficient from (2), using the SYS-GMM procedure (our preferred specification). The presented estimates yield some interesting conclusions:

(i) Only two components of spending are procyclical: capital expenditure and expenditure on wages and salaries. The largest of these two effects is capital outlays, with an estimated coefficient of 0.61 (more than twice the coefficient of spending on wages and salaries). Procyclical capital spending is supported by Sorensen et al. (2001). While they reported that capital outlays are less procyclical than aggregate sub-central government spending, we find the opposite--capital spending's procyclicality is nearly four times greater than that of total sub-central government spending. (8) Current spending is acyclical.

(ii) As expected, insurance benefits and repayments are counter-cyclical (during economic downturns sub-central governments spend more of their budget on welfare benefits and payments as unemployment rises, and this is the reverse when the economy grows).

(iii) State government revenues (both own revenue and intergovernmental revenue) are procyclical with respect to shocks in state-level GDP. (9) Own government revenues are nearly four times more procyclical than federal grants. (10) The result for own revenues is not surprising, but intergovernmental revenues might have been counter-cyclical (given the insurance element of these transfers).





With strong evidence of procyclical fiscal policy, is there a uniform pattern of procyclicality across the individual states, or do fiscal responses to economic shocks vary considerably? Figures 1 to 4 plot histograms of the individual cyclicality coefficients for the main areas of expenditure and revenue where there is procyclicality. These show the significant variability in procyclicality that exists, particularly for capital spending. Here the mean coefficient is 0.22 but the standard deviation of coefficients is 0.67, producing a coefficient of variation of 3. The coefficient of variation for total spending is 0.72; for Wages & Salaries it is 0.86, and for Total Revenue it is 0.54. (11)

4.2 Why is there variability in sub-central government procyclical fiscal activity?

It has been argued that differences in cyclicality might reflect differences in voters' preferences. Within states, competition for votes encourages an alignment of policy with the preferences of the median voter. Across different states, individuals' incentive to 'vote with their feet' means that there are likely to be different median-voter preferences (Tiebout, 1956). The implication is that the pattern of cyclical sub-central government spending across the US states will be correlated with the pattern of voters' preferences across US states. This prediction can be tested by measuring differences in voters' preferences on the scale that separates the ideologies of the Republican and Democratic Parties. Using the panel data set (across states and across time) it is possible to derive separate cyclicality coefficients for Democrat governors and Republican governors. The results are reported in table 3.

Several interesting results emerge from this analysis.

(i) Expenditure is procyclical for the full sample of observations and during periods of Democrat Governorship, but it is acyclical for Republican Governors. (12) As expected, the cyclicality coefficient of total expenditure is higher for Democrat Governors than from the full sample of observations. (13)

(ii) While total spending is acyclical for Republican Governors, there is evidence of procyclicality in Intergovernmental Expenditure, Capital Expenditure, and Wages & Salaries; although in each case the cyclicality coefficient is smaller in magnitude than for Democrat governors.

(iii) Direct expenditure is procyclical for Democrat governors but not for Republican governors.

(iv) Total Revenue is more procyclical for Republican governors: the cyclicality coefficient is more than twice the size for Republican governors than for Democrats. This difference is even greater for the components of revenue. Own revenue is procyclical, with a [beta] coefficient of 0.91 for Republican governors, but the equivalent estimate is 0.33 for Democrat governors. However, intergovernmental revenue is only procyclical for Democrat governors.

Our second hypothesis is that the likelihood of sub-central procyclical government spending depends on the existence of a coincidence between the political affiliations of the incumbent US President and the incumbent state governor, i.e. when the same political party holds the presidency and the state governorship. Table 4 presents cyclicality coefficients for three sub-samples when there is coincidence: (i) for both parties; (ii) for the Democrat Party and (iii) for the Republican Party.

During periods of coincidence, total sub-central government spending is procyclical (but less so than from the full sample). The analysis suggests that there is only procyclicality for periods of Democrat coincidence. In the full sample there is no procyclicality of intergovernmental expenditure, but there is procyclicality in periods of coincidence. It is only during periods of Democrat coincidence that total spending is procyclical. Focusing on the components of direct spending, capital spending is once again strongly procyclical. For the full sample of observations, the estimated cyclicality coefficient is 0.607, but this rises to 0.653 when the same party shares the presidency and state governorship. However, considering both Democrat and Republican coincidence separately, it is evident that capital spending procyclicality is greater during periods of Democrat coincidence: the cyclicality coefficient for Democrats is 0.921, while it falls to 0.515 for Republican coincidence.

Finally, total revenue is more procyclical during periods of coincidence, but this is especially the case when the coincidence affects the Republican Party. Intergovernmental revenue is only procyclical when there is a coincidence for the case of the Republican Party.


The focus of this study has been threefold: i) to investigate the cyclical properties of sub-central government spending and its funding in the US states; ii) to establish whether differences in cyclicality exist between states and over time, depending upon how voters vote; iii) to explain differences in cyclicality through the presence of political networks.

There is evidence that sub-central government expenditures and revenues are procyclical. The evidence indicates that, on balance, there are procyclical fiscal surpluses (as sub-central government revenues are strongly procyclical, but expenditures are weakly procyclical). The procyclicality of sub-central government revenue appears to be financed by procyclical own revenue, rather than by federal grants. When the analysis focuses on sub-central government spending, the key driver is procyclical capital spending.

If there is evidence of differences in the likelihood of procyclical spending, how can they be explained? There is support for two hypotheses. Differences in the cyclicality of state spending are correlated with the ideological differences that span the two main US political parties. Total state government expenditure is more likely to be procyclical if the state governor is a Democrat. If there is a Republican governor there is a greater likelihood of procyclical government revenue.

The availability of different political networks is also important. The evidence is that intergovernmental revenues (including 'grants in aid') are more procyclical if there is a coincidence between a Republican state governor and a Republican president. Spending is procyclical when there is a coincidence between a Democratic state governor and a Democratic president. Republicans are reluctant to spend procyclically even if they might be able to call on this political network. Democrats are inclined to spend procyclically when there is this political network. The existence of this network might prove important for states that are inclined to spend procyclically (even when the intergovernmental revenues are not procyclical) if it means that they are more confident to spend procyclically.

The results in this paper are consistent across the two hypotheses that have been tested (in each case differences in procyclical spending depend on the differences in willingness to spend public money). The results are also consistent with the ideological differences between Republicans and Democrats reported in the 'political business cycles' literature. While the Democratic Party is more disposed to increase public spending when national income increases, the Republican Party is more cautious (Republican state governors are systematically more likely to produce procyclical budgetary surpluses). This distinction is even more marked during periods in which there is a coincidence between the political affiliations of state governors and US presidents.

The normative implications are difficult to assess. Procyclical spending can increase a community's welfare, but this paper focuses on whether procyclical spending can be explained with reference to voters' preferences. The results indicate that procyclical spending is driven by procyclical spending on wages and salaries in the public sector and by procyclical capital spending that might also advantage producers in the private sector. The 'public choice' school identifies the inefficiencies that arise when producer groups press for increased public spending (Mueller, 2003). Procyclical central government spending is also explained with reference to pressure from producer groups (e.g. Lane, 2003). The results in this paper suggest that it is willingness to spend public money that is relevant. Pressure from producer groups appear to be more effective when, ideologically, voters and politicians are more willing to spend public money.

While it is not possible to comment on whether more procyclical spending is more (or less) efficient than less procyclical spending, it is possible to test the robustness of the conclusion that procyclicality depends on willingness to spend. It is possible, for example, to investigate whether these conclusions are relevant when focusing on the role that political parties play in other forums, e.g. when the same party controls both Upper and Lower houses of the state legislature. Also, instead of focusing on the components of expenditure described above, it is possible to consider expenditure by function (e.g. education, health) to establish whether certain types of spending are more likely to be procyclical than others. (14)

With all of these qualifications, this paper describes the pattern of fiscal cyclicality that exists across states in the USA. Differences in the political ideology of voters and politicians prove important when explaining sub-central governments' willingness to spend procyclically.


Abbott, A. and Jones, P. (2011) "Procyclical government spending: patterns of pressure and prudence in the OECD", Economics Letters, 111: 230-232.

Akitoby, B., Clements, B., Gupta, S. and Inchauste, G. (2006) "Public spending, voracity, and Wagner's law in developing countries", European Journal of Political Economy, 22: 908-924.

Alesina, A., Campante, F. R. and Tabellini, G. (2008) "Why is fiscal policy often procyclical?", Journal of the European Economic Association, 6: 1006-1036.

Anderson, G. M. and Tollison, R.D. (1991) Congressional Influence and the Patterns of New Deal Spending. Journal of Law and Economics, 34:161-75.

Arena, M., and Revilla, J. (2009) Pro-cyclical Fiscal Policy in Brazil: Evidence from the States, World Bank: Policy Research Working Paper No. 5144.

Blundell, R. and Bond, S. R. (1998) "Initial conditions and moment restrictions in dynamic panel data models", Journal of Econometrics, 87: 115-143.

Cox, G.W. and McCubbins, M.D. (1986) "Electoral Politics as a Redistributive Game", Journal of Politics, 48: 370-89.

Crouch, J. F. and Shughart II, W. F. (1998) The Political Economy of New Deal Spending, Cheltenham: Edward Elgar.

Cullis, J. and Jones, P. (2009) Public Finance and Public Choice, third edition. Oxford: Oxford University Press.

Gavin, M., Hausmann, R., Perotti, R. and Talvi, E. (1996) Managing fiscal policy in Latin America and the Caribbean: volatility, procyclicality and limited creditworthiness, Inter-American Development Bank, mimeo.

Kaminsky, G., Reinhart C. and Vegh, C. (2004) "When it rains, it pours: procyclical capital flows and macroeconomic policies", in M. Gertler and K. Rogoff (eds.), NBER Macroeconomics Annual, Vol. 19.

Kehmani, S. (2004) The Political Economy of Equalization Transfers, Working Paper 04-14, November, Andrew Young School of Policy Studies: Georgia State University.

Lane, P. (2003) "The cyclical behaviour of fiscal policy: evidence from the OECD", Journal of Public Economics, 87: 2661-2775.

Lane, P. and Tornell, A. (1996) "Power, growth and the voracity effect", Journal of Economic Growth, 1: 155-168.

Larcinese, V., Leonzio, R. and Testa, C. (2006) "Allocating the US Federal Budget to the states: the impact of the president", Journal of Politics, 68: 447-456.

Mueller, D. (2003). Public Choice III, third ed. Cambridge: Cambridge University Press.

Nickell, S.J. (1981). Biases in dynamic models with fixed effects. Econometrica, 49: 1417-1426.

Oates, W. (1972). Fiscal Federalism, New York: Harcourt Brace Jovanovich.

Sorensen, B., Wu, L. and Yosha, O. (2001) "Output fluctuations and fiscal policy: U.S. state and local government 1978-1994", European Economic Review, 45: 1271-1310.

Talvi, E. and Vegh, C. (2005) "Tax base variability and procyclical fiscal policy in developing countries", Journal of Development Economics, 78: 156-190.

Tiebout, C.M. (1956) "A Pure Theory of Local Expenditures", Journal of Political Economy, 64: 416-24.

Tornell, A. and Lane, P., (1999) "The voracity effect", American Economic Review, 89: 22-46.

Wagner, A. (1911) Staat in nationalokonomischer Hinsicht Handworterbuch der Staatswissenschaften. Lexis, Jena.

Woo, J. (2009) "Why do more polarized countries run more pro-cyclical fiscal policy?", Review of Economics and Statistics, 30: 310-328.

Andrew Abbott

University of Hull, Hull, UK

Philip Jones

University of Bath, Bath, UK

(1.) The increase in local expenditures is often greater when citizens receive an intergovernmental grant than when citizens receive an equivalent increase in income. This literature, on the 'flypaper effect', is surveyed in Cullis and Jones (2009).

(2.) Anderson and Tollison (1991) and Crouch and Shughart (1998) focused on the New Deal to report a positive correlation between Roosevelt's share of votes in US states and spending at the state level.

(3.) Larcinese et al. (2006) cite the statement of the Republican governor Mitt Romney (in the Washington Post, November 22, 2004): "For Republican governors it means we have an ear in

the White House, we have a number we can call, we have access that we wouldn't have otherwise had, and that's of course helpful".

(4.) Assistance and Subsidies consists of payments for which no product or service is received in return e.g. cash grants.

(5.) Local government expenditure consists of all spending by county governments; municipal governments; township governments; special district governments; and school district governments.

(6.) Intergovernmental revenues and expenditures represent grants-in-aid and the sharing of tax proceeds, as well as payments, in lieu of taxes. They also represent amounts for services performed by one government for another on a reimbursable or cost sharing basis.

(7.) Further testing indicated that interest payments are acyclical.

(8.) The strong procyclicality of capital spending is also supported by Lane (2003). For a sample of 22 OECD countries, considering spending from all tiers of government over the period 1960-1998, Lane concludes 'The most procyclical component of government spending is government investment (GI): indeed, it is the only category in which a strict version of the voracity hypothesis applies for some countries, with spending elasticities above unity' (Lane, 2003: 2668). Lane suggests a cyclicality coefficient for US government investment of 0.17.

(9.) Given that intergovernmental revenues arise principally from federal grants, they do not constitute a source of revenue raised locally. Thus we may also expect that changes in the amount of intergovernmental revenue offered will also depend upon the state of federal government finances, which in turn depend upon shocks to national income. We also found that intergovernmental revenues are procyclical with respect to national GDP.

(10.) We also confirmed that total tax revenues were procyclical, which in turn was explained by the procyclicality of sales tax and corporate income tax revenues.

(11.) Further investigation of the individual US states aggregated up to regional level, identified the Mid-West as a region where expenditures and revenues are acyclical. A broadly similar pattern of procyclicality for revenues and spending is found across the other three US census regions.

(12.) The estimate is not statistically significant.

(13.) These comparisons focus on government spending. In further research it might be interesting to compare cyclicality in tax expenditures in Democratic and Republican administrations.

(14.) In the long run, differences in short-term willingness to spend public money (as a response to changes in national income) may influence citizens' perceptions of the role of the public sector.
Table 1. Cyclical Sensitivity of Sub-national Fiscal Policy

                      Total           Primary
                   expenditure      expenditure

Fixed Effects           0.147 *         0.160 *
                       (3.18)          (3.18)

SYS-GMM                 0.123 *         0.155 *
                       (2.90)          (3.10)

                      Revenue         Deficit

Fixed Effects           0.418 *        -0.324 *
                       (3.55)         (-4.17)

SYS-GMM                 0.496 *        -0.432 *
                       (5.21)        (-11.44)

Notes: Estimates of the [beta] cyclicality coefficient are
derived by estimating (1) and (2) using the fixed effects
and SYS-GMM estimators. The t-ratios are reported in parentheses,
calculated from robust standard errors.
* indicates significance at the 5% level.

Table 2. Cyclical Sensitivity of Revenue and
Expenditure Components


  Intergovernmental          0.163
  Expenditure               (0.54)

  Direct Expenditure         0.050

   Current Expenditure       0.068

   Capital Expenditure       0.607 *

   Assistance and            0.005
   Subsidies                (0.05)

   Insurance Benefits       -0.886 *
   and Repayments          (-5.75)

   Wages and Salaries        0.293 *

   Interest Payments        -0.024

  Intergovernmental          0.146 *
  Revenue                   (2.95)

  General Revenue            0.609 *
  from own sources          (4.66)

Notes: see table 1

Table 3. Cyclical Sensitivity under periods of Democrat
and Republican Governorships

                         Estimated cyclicality coefficient

                         Full         Democrat    Republican
                         Sample       Governor    Governor


Total Expenditure          0.183 *      0.167 *     0.066
                          (4.22)       (3.10)      (1.39)

  Intergovernmental        0.163        0.469 *     0.287 *

  Expenditure             (0.54)       (2.42)      (3.12)

  Direct Expenditure       0.050        0.092 *     0.018
                          (1.44)       (2.51)      (0.36)

   Current                 0.068        0.100       0.031
   Expenditure            (1.50)       (1.92)      (0.55)

   Capital                 0.607 *      0.745 *     0.489 *
   Expenditure            (4.93)       (4.51)      (3.60)

   Assistance and          0.005       -0.039       0.076
   Subsidies              (0.05)      (-0.28)      (0.47)

   Insurance              -0.886 *     -0.793 *    -1.052 *
   Benefits and          (-5.75)      (-3.74)     (-6.47)

   Wages and               0.293 *      0.326 *     0.293 *
   Salaries               (5.18)       (3.66)      (3.06)

   Interest Payments      -0.024       -0.166       0.163
                         (-0.20)       (1.05)      (0.88)

Total Revenue              0.496 *      0.303 *     0.699 *
                          (5.21)       (4.30)      (3.47)

  Intergovernmental        0.146 *      0.121       0.146
  Revenue                 (2.95)       (2.44)      (1.75)

  General Revenue          0.609 *      0.328 *     0.910 *
  from own sources
                          (4.66)       (3.29)      (3.39)

Notes: See table 1

Table 4. Cyclical Sensitivity when there is a common incidence
of the president and the governor

                         Estimated cyclicality coefficient

                           Full      Overall co-    Democrat
                          Sample      incidence    coincidence

Total Expenditure         0.183 *      0.122 *       0.261 *
                         (4.22)       (2.77)        (3.22)
  Intergovernmental       0.163        0.398 *       0.323 *
  Expenditure            (0.54)       (3.29)        (2.42)

  Direct Expenditure      0.050        0.047         0.183 *
                         (1.44)       (0.98)        (2.11)

  Current                 0.068        0.074         0.161
  Expenditure            (1.50)       (1.71)        (1.93)

  Capital                 0.607 *      0.653 *       0.921 *
  Expenditure            (4.93)       (4.47)        (3.23)

  Assistance and          0.005       -0.115        -0.523
  Subsidies              (0.05)      (-0.77)       (-1.86)

  Insurance Benefits     -0.886 *     -0.962 *      -0.774 *
  and Repayments        (-5.75)      (-6.25)       (-3.55)

  Wages and Salaries      0.293 *      0.204 *       0.107
                         (5.18)       (2.34)        (0.94)

  Interest Payments      -0.024        0.051         0.039
                        (-0.20)       (0.25)        (0.10)


Total Revenue             0.496 *      0.578 *       0.449 *
                         (5.21)       (4.70)        (3.37)

  Intergovernmental       0.146 *      0.102        -0.092
  Revenue                (2.95)       (1.37)       (-0.62)

  General Revenue         0.609 *      0.721 *       0.373 *
  from own sources       (4.66)       (5.06)        (2.37)


Total Expenditure         0.052
  Intergovernmental       0.432 *
  Expenditure            (3.04)

  Direct Expenditure     -0.021

  Current                 0.027
  Expenditure            (0.57)

  Capital                 0.515 *
  Expenditure            (3.36)

  Assistance and          0.041
  Subsidies              (0.22)

  Insurance Benefits     -1.072 *
  and Repayments        (-5.57)

  Wages and Salaries      0.220

  Interest Payments       0.077


Total Revenue             0.646 *

  Intergovernmental       0.186 *
  Revenue                (2.18)

  General Revenue         0.892 *
  from own sources       (4.43)

Notes: see table 1
COPYRIGHT 2012 Southern Public Administration Education Foundation, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2012 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Abbott, Andrew; Jones, Philip
Publication:Public Finance and Management
Article Type:Report
Geographic Code:1USA
Date:Jun 22, 2012
Previous Article:NGOs holding governments accountable: civil-society budget work.
Next Article:Public finance and the extent of the voting franchise: a survey of historical cases and new evidence from early U.S. states.

Terms of use | Privacy policy | Copyright © 2020 Farlex, Inc. | Feedback | For webmasters