Elections and Monetary Policy: Is There a Political Monetary Cycle for Switzerland?
Indirect political influence on monetary policy is investigated by testing for effects of fiscal policy on monetary policy. Direct political influence is analyzed in the framework of elections, parties and administrations. This chapter addresses the question whether elections have any effect on Swiss monetary policy. While this question has been analyzed for most major central banks, an explicit and systematic test for the SNB is still missing.1 This chapter intends to fill that gap.
KeywordsIncome Autocorrelation Sonal
Unable to display preview. Download preview PDF.
- 1.Electoral cycles for monetary policy have been investigated for the Federal Reserve and most other central banks. Applications to the Federal Reserve include Alesina 1988, Alesina/Sachs 1988, Allen 1986, Beck 1982, 1984 and 1987, Grier 1987 and 1989, Haynes/Stone 1989, Havrilesky 1987, Laney/Willet 1983. For more recent applications to the German Bundesbank see Lang/Welzel 1992 and Berger/Woitek 1995. Other applications to the Bundesbank are cited in Vaubel 1993 49. Multi-country studies are Alesina et al. 1992 and 1993, Alesina/Roubini 1992. Switzerland is also included in some of these time series cross section analyses. Country by country results are also reported for Switzerland for the period 1958–1987. (See Alesina et al. 1992 16, table 8.) They show no signs of political manipulation of monetary policy.Google Scholar
- 2.For a more detailed review of the theory of political business cycles see Alesina/Roubini 1992 section 2. See also Belke 1993, Berthold/Fehn 1994, Gärtner 1994a, and the papers compiled in Frey 1997. For a textbook exposition see Mueller 1989 ch. 15. For a detailed theoretical and empirical analysis of the rational partisan theory see Belke 1996.Google Scholar
- 3.See Grier 1989 387 on early attempts to reconcile the political business cycle with rational expectations.Google Scholar
- 4.See Alesina et al. 1992 15. They account for this possibility by investigating a ‘partisan/opportunistic’ interaction term.Google Scholar
- 5.As noted above, tests for the existence of political business cycles can be carried out for policy outcomes and policy instruments. Pre-electoral fiscal and monetary ‘favors’ to key constituencies are not only electorally very useful. They are also easy to implement, relative to an attempt to increase the rate of growth of GNP, for instance. For a test using the inflation rate (a policy outcome) as dependent variable see e.g. Alesina/Roubini 1992.Google Scholar
- 6.According to Laney/Willett (1983 59) this last scenario may correspond most closely to the Nordhaus cycle for the US, since Nordhaus divides his electoral periods generally in half in this fashion.Google Scholar
- 7.Beck (1987 203) argues that voters are unconcerned with (and not even knowledgeable about) money growth. They care about its consequences, such as unemployment and inflation. Because of the monetary policy lag money growth during the electoral quarter will not show up in the variables that have an electoral impact until after the election. Therefore, it is unreasonable to assume that money will grow most rapidly on election day.Google Scholar
- 8.These cycles „may imply very little or nothing at all for a four-year cycle on employment; thus, they are consistent with the lack of empirical evidence for the Nordhaus’ ‘political business cycle’.“Alesina 1988 36.Google Scholar
- 9.This point was brought to my attention by Manfred Gärtner. See e.g. Gärtner 1994b 438.Google Scholar
- 10.Especially for rational political business cycle models it remains unclear whether this quarter should be considered pre- or post-electoral. See also Alesina 1988 37 for the US.Google Scholar
- 11.Of course, quarterly data are preferred to annual data for other reasons, too. Quarterly data yield more observations and, hence, more robust statistical tests. See section 1.3.2e. As already mentioned, monthly data might be more appropriate to model the Rogoff-type cycle. Unfortunately, production variables used in the reaction function approach are only available on a quarterly basis. And to allow for comparisons, autoregressive model specifications also employ quarterly data.Google Scholar
- 12.The question whether seasonally unadjusted or adjusted data are used may be important. „Haynes and Stone (1987) argue that one reason they found a political business cycle when others did not is that they used unadjusted data. Since the political monetary cycle peaks in the fourth quarter of the election year, it is possible that seasonal adjustment also removes the political cycle.“(Beck 1987 213).Google Scholar
- 13.Autoregressive models are based upon the assumption that money growth is generated by a covariance-stationary stochastic process that can be expressed in autoregressive form. See Alesina/Roubini 1992 668–669.Google Scholar