Retirement, health, unemployment, the business cycle and automatic stabilization in the OECD

Official adjustments of the budget balance to the cycle assume that the only category of gov-ernment spending that responds automatically to the cycle is unemployment compensation. But estimates show otherwise. Payments for pensions, sickness, subsistence, invalidity, childcare and subsidies of all sorts to firms respond automatically and significantly to the cycle as well. In addition, it is fairly common to use official figures for cyclically adjusted budget balances, di-vide by potential output, and use the resulting ratios to study discretionary fiscal policy. But if potential output is not deterministic but subject to supply shocks, then apart from anything else, those ratios are inefficient estimates of the cyclically-independent ratios of budget balances di-vided by potential output. (A fortiori, they are inefficient estimates of the cyclically adjusted ratios of budget balances to observed output.) Accordingly, the paper makes use of detailed data from the OECD’s Social Expenditure database to produce separate estimates of the impact of the cycle on disaggregated components of the budget balance, both in levels and in the form of their ratios to output. In addition, we discuss the relation between the two sorts of estimates. When the focus is on ratios of expenditure and revenue to output, the cyclical adjustments de-pend more on inertia in government spending on goods and services than they do on taxes (which are largely proportional to output). But they depend even still more on transfer pay-ments. Besides calling for different series for discretionary fiscal policy if ratios serve, these results also raise questions about the general policy advice to “let the automatic stabilizers work.â€