ABSTRACT
Studies of discretionary fiscal policy often centre on the ‘cyclically-adjusted budget balance’, or the budget balance following an adjustment for the part that depends on an automatic response to events. It is also often assumed that all of the adjustments to the cycle come from taxes and unemployment compensation. In addition, following cyclical adjustment, the analysis of discretionary fiscal policy frequently concerns the ratio of the cyclically adjusted government balance to output or potential output rather than the level. I shall put forward two criticisms of this procedure. First, many of the automatic responses to events result from other transfer payments besides unemployment compensation, including payments for pensions, sickness, subsistence, invalidity, childcare and subsidies of all sorts to firms. Second, if the issue is the ratio of the cyclically adjusted budget balance, the cyclical adjustment should be for the ratio rather than the level. Otherwise, the estimates of the cyclical adjustment are inefficient. According to both arguments, the usual estimates of the series for discretionary fiscal policy are often incorrect. The first criticism always applies when there is recourse to official sources for figures for the ‘cyclically-adjusted budget balances’ since those figures are constructed on the assumption that taxes and unemployment compensation are the sole elements of the budget that respond to the cycle. The second criticism follows whenever the subsequent analysis focuses on ratios.
