Long‐Run and Short‐Run Budgeting: Theories and Empirical Evidence for the Canadian Provinces

Abstract

This paper tests public budgeting to ascertain if it has both a long‐run equilibrium and short‐run incremental process. In the model, the decision‐maker strives to achieve budgetary balance over the long‐run but is constrained in the short‐run and follows incremental decision‐making. The interaction between expenditures and revenues, along with several control variables, is tested for each of the Canadian provinces using data between 1961 and 2000. The results show that in the long‐run, expenditures force the budget toward balance in all the provinces with the exception of British Columbia. In that province, there was a fiscal synchronization of revenues and expenditures working in combination. In the short‐run, incrementalism occurs in nine of the ten provinces.

Keywords:

• Budget
• Decision making
• Incrementalism
• Provincial government

Notes

^aBalance over the long‐run is associated with the tendency of the budget, in either surplus or deficit, to head toward balance.

^bAn earlier version of this paper was presented at the Southern Political Science Association 2001 Meeting.I would like to thank the anonymous referees for their helpful comments on revising this manuscript.

^cAlthough we do not specifically model the rational forms of budgeting such as program and performance based budgeting. We argue in this paper that decision‐making is both forward looking, which the rational techniques advocate and incremental.

^dTheir interpretation of incrementalism will be discussed later, specifically as it relates to the size of the increment.

^eSee Alesina et al.[35]for an extensive literature review of PBCs. Note that we take traditional adaptive expectations view of PBCs.

^fAdmittedly, 39 observations is not a large sample size. We could have used quarterly data, but this would artificially boost the sample size. Budgets are determined on an annual basis, not a quarterly basis. Spending is allotted on a quarterly basis to ensure that government agencies and departments do not overspend during the fiscal year. Therefore, quarterly data provides a misrepresentation of true annual budgetary decision‐making outcomes.

^gThe following authors in political science have used or advocated the use of real data: Davis et al.,[11]Gist,[42]LeLoup,[43]Baumgartner et al.,[12]Su et al.,[33]Beck,[44]Berry and Lowery,[45]Kamlet and Mowery,[46]White,[47]and Kiewiet and McCubbins.[48]

^hThis is not simply the deficit for each of the provinces. It represents the long‐run relationships between revenues and expenditures for the provinces.

ⁱThe econometric software used to perform the tests was EViews 4.0.

^jWe use one lag to correct for present spending impacting past spending.

^k Table 2indicates that there is cointegration of revenues and expenditures. We notice that for Alberta, budget revenues and expenditures had a λ_tracevalue of 16.94, well above the critical value of 12.53 at the 5% level. It is also apparent that the λ_maxvalue for Alberta is 14.73, which is above the critical value of 11.44 at the 5% significance level. Similar test results are reported in Table 2for the rest of the provinces indicating a rejection of the hypothesis of no cointegration between revenues and expenditures. Since each series is cointegrated, the next step is to apply the error correction model for each of the provinces. We do this by using the Vector ECM shown in Table 3and demonstrated in Eqs 1.1and 1.2.