This article appears in the following Oxford Review of Economic Policy issue: CAPITALISM AND INEQUALITY [View the issue table of contents]
Monetary policy and European unemployment
* Schumpeter School, Department of Economics, University of Wuppertal, e-mail: schettkat{at}wiwi.uni-wuppertal.de
** Schumpeter School, Department of Economics, University of Wuppertal, e-mail: sun{at}wiwi.uni-wuppertal.de
 |
Abstract |
|---|
In the long history of rising and persistent unemployment in Europe, almost all welfare-state institutions—employment protection legislation, unions, wages, wage structure, unemployment insurance, etc.—have been alleged to have caused and found guilty of causing this tragic development at some point in time. Later, welfare-state institutions in interaction with external shocks were identified as more plausible causes of rising equilibrium unemployment in Europe. Monetary policy has managed to be regarded as innocent. Based on the assertion of the neutrality of money in the medium and long run, the search for causes of European unemployment has shied away from the policy of central banks. But actually the institutional set-up regarding monetary policy is very different between the Federal Reserve System (Fed) and the Bundesbank and its successor, The European Central Bank (ECB). We argue that the interaction of adverse shocks and tight monetary policies may have been the major—although probably not the only—cause of unemployment in Europe remaining at ever higher levels after each recession. We identify the monetary policy of the Bundesbank as asymmetrical, in the sense that the Bank did not actively fight against recessions, but it dampened recovery periods. Less constraint on growth would have kept German unemployment at lower levels.
Key Words: production • employment • unemployment • monetary policy • central banks and their policies
Earlier versions of this paper were discussed at the LoWER conference (‘Institutions, markets and European unemployment revisited: What have we learned?’, Amsterdam, 18–19 April 2008), seminars at the University of Hagen, University of Wuppertal, and IAB, Nuremberg. We are indebted to Frank den Butter, Friedrich Kißmer, Joachim Möller, Robert Solow, Heinz-Peter Spahn, Helmut Wagner, and an anonymous referee for their invaluable comments. Susanne Hochscherf provided excellent research assistance. All remaining errors are ours.
1 For a recent elaboration see Mitchell and Muykens (2008).
2 Solow refers to the ‘natural’ rate of unemployment as the ‘neutral’ rate of unemployment because ‘natural’ is a misnomer in that the level of unemployment, neutral to inflation, is influenced by institutions rather than naturally given. We use ‘natural’, ‘neutral’, ‘NAIRU’, and ‘equilibrium rate of unemployment’ as synonyms in this paper.
3 Mankiw (2001, C48) classifies the finding of non-neutrality by Bernanke and Mihov (1998) ‘surprising’, because the ‘the paper purports to provide evidence for the opposite conclusion—long-run monetary neutrality.... But if one does not approach the data with a prior view favouring long-run neutrality ... the data's best guess is that monetary shocks leave permanent scars on the economy’.
4 In the economic policy debate, usually NAIRUs (or natural rates) and/or potential output are used as proxies. Which one is used is largely a matter of taste, because both are close relatives although not identical twins. Blinder and Reis (2005), for example, analyse the Greenspan legend in terms of NAIRUs.
5 Staiger et al. (1997, p. 40), however, mention that the slope of the Phillips curve (the Phillips coefficient) seems to be quite stable, but that the intercept with the x-axis is imprecisely estimated.
6 In addition, the confidence-band around the Kalman-filtered NAIRU is, according to the OECD (Richardson et al., 2000, p. 67), about 0.7 percentage points in both directions (1 Standard error), i.e. the unemployment gap could be around 5 percentage points or two-thirds of the actual unemployment rate in 1982. Using the Okun gap formula (the Okun Gap describes the difference between the actual and the ‘neutral’ unemployment rate as a difference in unemployment rates as a function of GDP growth rates) we estimate that
- – 1 per cent difference in the unemployment rate results in about 2.5 per cent loss in GDP in Germany before 1989;
- – 1 per cent difference in the unemployment rate results in about 2.0 per cent loss in GDP in the USA before 1989.
- – 1 per cent difference in the unemployment rate results in about 2.0 per cent loss in GDP in the USA before 1989.
This translates into a 10–12.5 per cent loss in GDP. Not peanuts!
7 It could be evidenced by the remark of the former president of the Bundesbank, Otmar Emminger: ‘price stability is not everything, but without price stability everything is nothing’ (Deutsche Bundesbank, 2008).
8 See Spahn (2009, pp. 218–19) for a discussion of the formal equivalence of targeting monetary aggregates and inflation rates.
9 Such a monetary response function is a simple interest-rate rule proposed by John Taylor (1993), which states that the central bank should set the short-term interest rate taking into consideration the variability of inflation and output. Taylor (1993) specifies a monetary response function that fits the US data pretty well for the period from 1987 to 1992:
|
|
is inflation rate, and (y–y*) is the percentage deviation of the real output from its target.
10 The data sources are the Bundesbank's statistics and the OECD's Economic Outlook: West German data for 1975–90 and German data for 1991–8. Inflation, , is based on CPI. The implicit inflation targets by the Bundesbank (*) are as shown in Table 1. The real GDP trend is created with the HP filter (with 
equals 1,600). Then (y–y*) is calculated as percentage deviation of real GDP from its trend.
11 We use output of the entire economy (GDP) rather than manufacturing output used by other researchers (e.g. Surico, 2003).
12 Introducing two lags is sufficient to reduce the serial correlation in the residuals of an OLS regression and it reflects the fact that central banks care about financial market stability and thus hesitate to make overly abrupt changes in interest rates (see Blinder and Reis, 2005).
13 The negative coefficient for (y – y*) is not significant. Otherwise it would indicate a pro-cyclical monetary policy, i.e. rising interest rates when the production potential is underused. We find this coefficient to be significant when we use generalized method of moments (GMM) rather than OLS. Other considerations (such as exchange-rate stabilization, or an underestimation of the potential (see the discussion in section II)) may rationalize that result. Adverse supply shocks may be another explanation for the negative coefficient but, in all three recessions analysed, both inflation and output declined, although inflation rates were—especially in the 1970s and 1980s—above their (implicit) targets.
14 Wald-test under the null hypothesis of equality of the coefficients for (i) inflation deviations and (ii) output gaps in 2(a) and 2(b):
| ||||||||||||||||||||||||||||||