I study the short-term impact of surprise inflation on the primary balance by separating those budgetary items which immediately respond to inflation from non-responding ones. I assume a passive fiscal policy in a one-year horizon; therefore items fully controlled by the central government are treated as non-responding ones. In the case of the private sector and decentralised government, their responses to inflation depend on decisions.

If a 1 percentage point inflation surprise is compensated in the private sector by an increase in wages and consumption, then the nominal increase of the related tax revenue is equivalent to 0.26% of the GDP, i.e. their real value and their proportion to the increase of the nominal GDP remains unchanged. Otherwise, the nominal value of the revenue remains unchanged, resulting in a decrease in its real value and an increase in the deficit-to-GDP ratio by 0.26%. Compensation decided in a decentralised government has the opposite effect, because if it is implemented, then the real value of decentralised expenditure and its proportion to the GDP remains the same through an increase in the nominal expenditure, while if the nominal expenditure is fixed, it results in a decrease of the real expenditure and approximately a 0.13% decrease in the deficit-to-GDP ratio. The fixing of the other nominal expenditure, which does not respond to inflation, results in a 0.08% decrease in the GDP-proportionate expenditure and deficit compared to an increase in the nominal GDP.

Reviewing certain episodes in Hungary, we have found that owing to planning errors, the official inflation forecast usually resulted in a larger ‘surprise’ for the budget than for the private sector. Thus, at the time of the impact of the surprise on the budget, the private sector experienced either no surprise or very little, which if materialising was in most cases immediately compensated for. An exception in this case was the adjustment in 1995 - supported by an inflation surprise - when the ratio of the moderately increasing nominal tax revenue to the GDP and its real value significantly decreased.

The increase of indirect taxes had an adversary effect on the increase in nominal consumption (inflationary compensation) (1995 and 2004). In addition to moderating consumption, the indirect tax increase in mid-2006 can also moderate wages in real terms since it was announced after the usual wage increases. The decentralised government behaved similarly, as indicated by the experiences in the developed OECD countries.

The response of the decentralised government to the moderate nominal increase in central government transfers (i.e. to the decrease in real value of their funds) was to moderate the nominal increase of decentralised expenditure, i.e. real value loss was for the most part not compensated for. The rate of compensation was larger in those years when cheap financing was available (funds from privatisation in 2000), or when the surprise was not sizeable and coincided with the uprising phase of the election-related investment cycle (1998). In 2007 the optimistic inflation projection can result in lower-than-planned central transfers in real terms, which can in turn moderate the nominal increase of decentralised expenditure. Presumably, however, the size of the planning error and its deficit-decreasing effect will be not significant.

Keywords: surprise inflation, inflation sensitivity, decentralised government.

JEL: E31, E65, H61, H71, H72.