Do Elections and Dependence Affect the Growth Forecasts of Central Banks?

Abstract

When the economic forecasts of public institutions deviate from actual outcomes, the deviations may not only be due to error but also to political intentions. A central bank, for example, may publish over-optimistic growth forecasts at the time of election to facilitate the re-election of the government, and it may publish excessively pessimistic growth forecasts during the years prior to the election to justify an expansionary monetary policy before the election. Our sample covers six central banks which have published growth forecasts in the period from 1998 to 2008 with forecast horizons from one to eight quarters. The time series analyses reveal a significant optimism bias in the forecasts published by the Bank of England and the Bank of Thailand in the election quarter and a significant pessimism bias in the forecasts published by the Bank of England in the seven pre-election quarters. Banco do Brasil behaves contrary to our hypotheses. There are no significant electoral effects in the forecasts of Banco Central de Chile, Reserve Bank of New Zealand and Sveridges Rijksbank. In a cross-section analysis, central bank dependence as measured by an updated Cukierman index is significantly associated with excessive growth pessimism. Finally, we present a panel-data analysis with fixed effects for forecast horizons up to four quarters and allow for central bank independence. Using the same index, we find that forecasts are significantly too pessimistic during the pre-election period, thus justifying expansionary monetary policies before the election, even if the central bank is completely independent (but not necessarily non-partisan).