Key Considerations Governing the Choice of Monetary Policy Strategy

  • Aerdt C. F. J. Houben
Part of the Financial and Monetary Policy Studies book series (FMPS, volume 34)


At an abstract level, the choice of monetary policy strategy is determined first by the relationships underlying the economy and second by the desired reaction function of the central bank to economic developments. The former is a function of the structural characteristics of the economy and of the nature of disturbances to which the economy is likely to be subjected. The latter depends on the authorities’ prioritisation of economic objectives in the face of different transitory disturbances. And the interrelation between the two is driven by political economy variables, which inter alia determine the credibility of the authorities’ monetary policy commitments. This chapter elaborates on these general considerations determining the monetary strategy choice, drawing on the large body of related theoretical literature. The objective is to provide a normative backdrop against which to evaluate the development of European monetary strategies in practice.


Exchange Rate Interest Rate Monetary Policy Central Bank Money Demand 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Poole’s original analysis is based on a closed-economy Hicksian IS-LM model and is cast in terms of an interest rate or money supply rule. Notwithstanding its limitations, the straightforward and intuitively appealing set-up of the model provides an attractive theoretical basis to gauge the implications of adopting different monetary policy targets. The framework is also applicable to open economies: in the absence of restrictions on capital flows, an interest rate rule is comparable to an exchange rate rule with the interest rate fixed by the external anchor. In this sense, the standard framework amalgamates the instruments, operational targets and intermediate targets of monetary policy; to avoid ambiguity, these are communally referred to as ‘targets’.Google Scholar
  2. 2.
    For an overview of the main issues see B. Friedman (1990). Various extensions are discussed in Argy and de Grauwe (1990); early refinements are presented in Turnovsky(1976).Google Scholar
  3. 3.
    It goes without saying that under exceptional circumstances different conclusions may apply regarding the optimal monetary target. Goodhart (1989, pp. 340-342) presents a hypothetical case where an increase in incomes raises investment more than savings (such that the IS-curve is positively rather than negatively sloped); however, as long as the interest rate is more potent in stabilising the real sector than the monetary sector (i.e., the slope of the IS-curve is less steep than that of the LM-curve), the standard conclusions on the stability of the different monetary targets continue to apply.Google Scholar
  4. 4.
    In principle, the evaluation should take account of all macroeconomic issues of concern to the authorities including, next to output and inflation, objectives for such variables as unemployment, real consumption and the balance of payments. However, this would lead to an excessively complex and unwieldy definition of the authorities’ loss function; this is reflected in the literature, where the optimality criterion is most commonly specified in terms of the stability of real output and prices.Google Scholar
  5. 5.
    At a technical level, the ambiguity is further complicated by the measurement question of whether the stability of output should be assessed relative to its old or new equilibrium.Google Scholar
  6. 6.
    This point is forcefully made by M. Friedman (1953) in an early article on the benefits of exchange rate flexibility.Google Scholar
  7. 7.
    This is argued, for example, by Aghevli, Khan and Montiel (1991).Google Scholar
  8. 8.
    This line of reasoning is also followed by Argy (1990, pp. 36-37), with one exception. He claims that the superior regime is ambiguous under most shocks originating abroad, but that flexible rates are likely to be more stabilising in the case of foreign money demand disturbances. This latter conclusion is, however, questionable. It is based on the assumption that the foreign authorities will adjust interest rates in reaction to the money demand shock; thus, the foreign money demand disturbance ends up looking more like a foreign expenditure disturbance. The opposite conclusion (i.e., that a fixed rate system is likely to be superior) would be drawn if it were assumed that the foreign authorities would fully accommodate the monetary shock (as they should, to promote both price and output stability).Google Scholar
  9. 9.
    Marston(1982).Google Scholar
  10. 10.
    Calmfors and Driffill (1988) provide empirical evidence for this hypothesis; their survey includes 10 of the current European Union members. The hypothesis stressing the importance of ‘corporatist’ wage bargaining is further spelt out in De Grauwe (1997). In a critique, Soskice (1990) questions Calmfors and Driffill’s classification of countries, arguing that actual wage setting often differs from the formal structure governing wage bargaining and that the weaknesses of decentralised systems become apparent when other factors (such as lower incentives to invest in human capital) are taken into account. A further amendment of the literature on corporatism is provided by Garrett and Way (1995) who review the influence of ‘corporatism’ in terms of whether or not the trade union movement is dominated by unions in the tradables sector. They find that if the latter is the case (as in Austria and Finland), the benefit of closer regional monetary and economic integration will be much larger than in countries where this is not the case (as in Sweden). The arguments that link these different views on labour market structure to the monetary strategy choice are essentially similar to those related to the centralised/decentralised case.Google Scholar
  11. 11.
    This argument becomes more complicated when hysteresis effects are taken into account. But even then, it is doubtful whether continuous accommodation of excessive nominal wage demands increases welfare. Rather, the existence of such effects should prompt all parties concerned to make the transitional period of structural adjustment as short as possible.Google Scholar
  12. 12.
    Bertola (1989) illustrates the importance of uncertainty using a microeconomic model incorporating uncertainty and adjustment costs related to resource reallocation. His debatable conclusion that stable exchange rates result in greater income variability and thus reduce labour mobility is dependent on the specification of his model and the assumption that the economy is subject to asymmetric terms of trade disturbances. He also disregards issues related to the eventual reconversion of foreign earnings into national currency.Google Scholar
  13. 13.
    Bini Smaghi and Vori (1993, p. 91) reject migration as an adjustment instrument by emphasising the “social and economic costs that economists generally ignore”.Google Scholar
  14. 14.
    Bayoumi and Prasad (1997) analyse sectoral output, employment and productivity trends, and find that these latter are dominated by country-specific factors in the EU and by industry-specific factors in the United States. Their results also suggest no significant structural changes affecting the degree of integration of European labour markets between 1970 and 1987. Eichengreen (1993) compares labour mobility in Britain and Italy with that in the United States, and finds the elasticity of migration with respect to inter-regional wage differentials is more than five times larger in the United States. The study by Thomas (1994a) analysing labour market flexibility in France, Germany, Italy, the United Kingdom and the United States over a quarter of a century, corroborates these findings.Google Scholar
  15. 15.
    Kenen (1969) introduced this element into the optimal currency area debate.Google Scholar
  16. 16.
    See Bayoumi and Eichengreen (1996), who also suggest that the smaller the country, the tighter the constraint on borrowing, and the more limited the scope for stabilisation policies. In empirical research with clear parallels to the European context, Bayoumi and Masson (1998) investigate fiscal stabilisation across Canadian provinces and find significant Ricardian effects: budgetary stabilisers at the local level are estimated to be only one-third to one-half as effective as those at the national level. Overall, however, they recall that studies on the size of federal fiscal stabilisers have generally found these to be modest, offsetting at best 20 to 30 percent of the initial reduction in income.Google Scholar
  17. 17.
    At a national level, an exception is provided by Denmark, which has actively used budgetary policy in the 1990s to compensate for asymmetric economic developments.Google Scholar
  18. 18.
    Polak (1988) provides a lucid overview of the generally unsatisfactory experience in the 1970s and early 1980s with fiscal policies aimed at economic stabilisation.Google Scholar
  19. 19.
    Reflecting this ambiguity, Aghevli, Khan and Montiel (1991) and Gros and Steinherr (1997) take the view that openness per se does not allow any general conclusions on the preferable exchange regime.Google Scholar
  20. 20.
    This is the position initially set out in the seminal article by McKinnon (1963); more recent supporters include Guitián (1994a), De Grauwe (1997), Masson and Taylor (1992), and Tavlas (1993).Google Scholar
  21. 21.
    Frankel and Rose (1996) find a strong positive link between the degree of trade integration and the correlation of business cycles using panel data for 20 industrial countries over a 30-year period. Similarly, the European Commission (1990) finds a significant inverse relationship between the extent of trade barriers and the symmetry of sector-specific shocks.Google Scholar
  22. 22.
    European Commission (1990, p. 43). In an insightful earlier essay on the monetary policy objectives of the European Community, Padoa-Schioppa (1988) speaks of an “inconsistent Quartet” of full trade integration, complete mobility of capital, fixed exchange rates, and autonomous monetary policy; clearly, the first element can be dropped without losing the inconsistency.Google Scholar
  23. 23.
    For an overview of the movement towards full capital liberalisation in Europe, see Bakker (1996).Google Scholar
  24. 24.
    In the early 1980s, the G-7 Working Group on Exchange Market Intervention chaired by Jurgensen (1983) already concluded that sterilised intervention did not appear to be an effective instrument to influence exchange rates in the face of significant market pressure; domestic policy adjustments, in particular in the field of monetary policy, were viewed as indispensable to counter such pressure.Google Scholar
  25. 25.
    At a global level, there is clear evidence that progress with capital liberalisation has been associated with a declining popularity of exchange rate targeting; see Cottarelli and Giannini (1997).Google Scholar
  26. 26.
    On the risk of self-fulfilling currency attacks in the context of capital mobility, see Eichengreen and Wyplosz (1993) and Obstfeld (1995).Google Scholar
  27. 27.
    Once market participants become concerned that the exchange rate may be adjusted, the interest rate will need to react vigorously to offset such concerns. For instance, if a portfolio manager takes the view that there is a trifling one-in-ten chance of a currency being devalued by just 5 percent in the next week, the interest rate differential will need to be 30 percent on an annual basis in favour of the suspect currency to compensate the relatively mild devaluation fears. A vivid description of the difficulty of sustaining a fixed-but-adjustable regime in the face of full capital mobility, based on first-hand experience, is provided by Crockett (1994a).Google Scholar
  28. 28.
    Kenen (1969), in first highlighting the importance of diversification, correctly took the view that this may be more relevant than the issue of labour mobility.Google Scholar
  29. 29.
    See European Commission (1996, pp. 71-77). An earlier discussion of the degree of product market integration, based on a study for 1987, is provided in European Commission (1990, p. 142).Google Scholar
  30. 30.
    See Bini Smaghi and Vori (1993), Helg et al. (1994) and Masson and Taylor (1993).Google Scholar
  31. 31.
    Bayoumi and Prasad (1997), however, find broadly similar degrees of relative specialisation in Europe and the United States, implying that industry-specific disturbances are likely to have a comparable impact in both economic areas.Google Scholar
  32. 32.
    For example, the monetary policy transmission mechanism is not included in Tavlas’ (1993) list of relevant characteristics identified in the literature on optimum currency areas; neither is it included in the costs and benefits of monetary integration discussed in European Commission (1990) and De Grauwe (1997); nor is it mentioned in the optimum currency area criteria set out in Bayoumi and Eichengreen (1996), Frankel and Rose (1996), and Masson and Taylor (1993).Google Scholar
  33. 33.
    Although insight into the different channels of monetary transmission has been significantly improved by numerous studies during the past decade, the words of King (1994) remain largely valid: “The transmission mechanism of monetary policy is one of the most important, yet least well understood, aspects of economic behaviour.”Google Scholar
  34. 34.
    Borio (1995) provides a comprehensive overview of the structure of credit to the nongovernment sector in fourteen countries, including nine EU members. His findings confirm the popular classification that, in general, Anglo-Saxon countries are characterised by a relatively high share of credit to households, with a comparatively large share of adjustable rate debt (except in the US), and widespread use of real estate collateral. This difference between Anglo-Saxon and continental European countries is set out in more general terms in Boonstra and Eijffinger (eds.) (1997) in a survey of developments in financial systems and monetary policy in nine industrial countries, including seven EU members.Google Scholar
  35. 35.
    Using the United Kingdom as an illustration, Eijffinger (1996) stresses that a high short-term interest rate sensitivity may contribute to a perverse incentive structure for the central bank.Google Scholar
  36. 36.
    These studies have produced very different results and a consensus opinion on the degree of differences in monetary policy transmission has not emerged — not even at a general level (see Kieler and Saarenheimo (1998)). In part, this reflects the diverse methods used in the empirical literature: the studies have included straightforward structural analyses (focusing on differences in financial structures), large macroeconometric models (constructed on a single-and multi-country basis), small macroeconomic models, and structural VAR models. The greatest cross-national differences appear in the studies based on estimates of large, single-country macroeconometric models (see for instance Dornbusch, Favero and Giavazzi (1998)), but the statistical comparability of the results is limited by the variation in model specification. By contrast, the multi-country model approach suggests relatively small differences in monetary transmission, but this may be attributed to the similar structure that such an approach imposes on all the countries. Given their more aggregated nature, small macroeconomic models may likewise do insufficient justice to cross-country diversity in economic structure. In turn, studies using structural VAR models may be questioned on account of the uncertainties involved in the structural identification schemes. Moreover, even within this specific approach, the results have not been uniform: while Ramaswamy and Sloek (1997) find output responding more slowly but cumulatively more strongly in some EU countries than in others, Barran, Coudert and Mojon (1996) also identify differences in the output response to monetary policy changes (as well as in the response of the final demand components), but find a reasonably similar pass-through in terms of timing. The varying strength of individual transmission channels in European countries is analysed in De Bondt (1999) and Kakes (2000).Google Scholar
  37. 37.
    In Goodhart’s (1984, p. 96) own words, “any observed statistical regularity will tend to collapse once pressure is placed upon it for control purposes”.Google Scholar
  38. 38.
    Based on a global study of 100 countries during the period 1970–94, Cottarelli and Giannini (1997) establish that the average volatility of broad money in countries that targeted this aggregate was one-half of that in the other countries. While it is unclear whether this lower volatility is a cause or a consequence of the policy focus, it would a priori seem to indicate that the impact of Goodhart’s Law is not very powerful on an aggregated basis.Google Scholar
  39. 39.
    Hochreiter and Winckler (1995) present evidence for Austria that the exchange rate peg to the Dmark led to adjustment processes which enhanced real wage flexibility and eventually made the country part of an optimal currency area with Germany, even though it clearly was not part of this area at the time the peg was first adopted. Similarly, Kremers (1990) investigates changes in Irish wage formation after the inception of the ERM and finds that inflation expectations in Ireland followed expected price movements in the United Kingdom until 1979 and shifted sharply to the expected price behaviour in the ERM partners thereafter. More generally, while Egebo and Englander (1992) argue that ERM participation per se does not provide a direct credibility bonus, they conclude that it does seem to improve policy discipline. Since such discipline is likely to impact labour market behaviour with a considerable lag, credibility benefits in labour markets will only emerge over time.Google Scholar
  40. 40.
    While its study focused on the consequences of economic integration rather than specifically on exchange rate stability, the European Commission (1990) argues that the increasing product market integration in Europe has tended to be of the intra-industry type, notably in the manufacturing sector; thus, integration reduces the likelihood of asymmetric shocks. Similarly, Frankel and Rose (1996) find that intensified international trade links tend to lead to more closely correlated business cycles between countries. On this basis, they emphasise that since Europe is pursuing trade integration, countries adopting exchange rate links (or joining EMU) may satisfy optimal currency area criteria after the policy switch, even if this was not the case in the past. However, Bayoumi and Eichengreen (1996) present tentative evidence that increasing economic integration in Europe during 1971–1987 has been accompanied by a measurable increase in regional industrial specialisation, thereby reducing the correlation of underlying disturbances.Google Scholar
  41. 41.
    An early IMF study (1984) on exchange rate volatility and world trade concluded that empirical evidence of a systematic significant link could not be found. In its paper on the benefits of a single currency for the European internal market, the not wholly unbiased European Commission (1990) also recognises the absence of strong empirical proof. A small effect is however found by Gagnon (1993), who estimates that the switch to flexible exchange rates after the collapse of the Bretton Woods regime may have reduced the volume of world trade by approximately 1 per cent. In a comprehensive study, Frankel and Wei (1995) also find a small and declining impact. The decline of this impact over the past three decades is arguably linked to the increasing availability and use of financial instruments to hedge exchange rate risks — although it should, of course, be emphasised that the use of these instruments is not available for all currencies and maturities and is certainly not without cost. As an exception, Stokman (1995) finds significant effects of exchange rate risk on intra-EC trade: on the basis of a disaggregated product/sectoral analysis he estimates that the combined exports of five major EC countries would have been 3 percent lower if exchange rate risk during 1987–90 had been at its level in the years preceding the EMS. However, as he only focuses on intra-EC trade, his estimates include both trade creation and trade diversion and may thus overstate the total effect.Google Scholar
  42. 42.
    Bhandari and Mayer (1990) show that saving-investment correlations during 1975–1987 were substantially lower between the countries participating in the EMS and Austria, than between non-EMS countries. The relationship between saving and investment was also weaker in the post-1982 period (when the EMS hardened) than in the pre-EMS period.Google Scholar
  43. 43.
    More specifically, based on observations for 1983 and 1993, the cross-country comparison in Borio (1995) does not provide evidence of convergence in the structure of credit between the Anglo-Saxon and the other industrial countries.Google Scholar
  44. 44.
    Buttiglione, Del Giovane and Tristani (1997) analyse the impact of changes in central bank rates on the term structure of interest rates during 1987–95. They find that central bank rate increases led to decreases in long-term forward rates in Germany, the Netherlands and Belgium, to no change in France, and to rises in Italy, Spain, Sweden and the United Kingdom. While they link this to the past inflation records and indirectly to the credibility of the long-term anti-inflationary commitment of monetary policy, there is a clear-cut relationship with the frequency and size of past adjustments to the central exchange rate parities.Google Scholar
  45. 45.
    Given the implicit consequences for the trade-off between near-term output and price objectives, it is worth noting that monetary strategies involving a regime commitment (such as an exchange rate arrangement or inflation target) are often decided in the political realm-i.e., by the Government and not by the central bank.Google Scholar
  46. 46.
    Based on a comprehensive study involving 136 countries over a 30-year time-span, Ghosh et al. (1995) find that fixed exchange regimes are associated with lower inflation, but more variable output and employment than flexible regimes. Overall growth performance is not found to differ between exchange regimes. These results are materially unchanged when allowance is made for possible endogeneity of the regime choice. Moreover, no anti-inflation benefit is found to accrue to exchange regimes that are purportedly pegged, but show frequent parity changes in practice.Google Scholar
  47. 47.
    Papademos and Modigliani (1990, pp. 476–487) provide an extensive comparison of the trade-offs between different money and credit aggregates.Google Scholar
  48. 48.
    Rogoff (1985) provides a formal analysis of this trade-off. Fischer (1994) presents evidence that, during the period 1960–1992/93, inflation was lower and less variable in Germany than in the United States, but output growth was more variable. (The average growth rates of the two countries were identical during this period.) He attributes this to the greater independence of the German Bundesbank. He acknowledges, however, that his hypothesis is not supported by aggregate data for the industrialised countries.Google Scholar
  49. 49.
    Whether central bank independence facilitates maintaining price stability is a different question from whether such independence reduces disinflation costs. In this context, De Haan, Knot and Sturm (1993) present evidence that central bank independence promotes a lower level of inflation, but does not reduce the (transitory) disinflation costs of arriving at that lower level. In fact, as discussed in Eijffinger and De Haan (1996), studies suggest central bank independence may even increase disinflation costs.Google Scholar
  50. 50.
    In an empirical survey of 18 OECD countries in the period 1960–87, Alesina, Cohen and Roubini (1992) find statistically significant evidence of a political monetary cycle of moderate intensity involving an expansionary monetary policy in election years; however, this cycle is not very strong and does not occur at every election.Google Scholar
  51. 51.
    To a large extent these qualifying arguments are also applicable to the previously mentioned output variability-inflation variability trade-off underlying a direct inflation targeting strategy. But the prioritisation of objectives is less tightly preordained when an independent central bank is given a general price stability objective, than when a central bank is explicitly mandated to aim at a specific, time-bound inflation target. Thus, the short-term trade-off is more likely to favour inflation stability at the cost of output variability under the inflation targeting approach, than under an independent central bank broadly oriented at price stability.Google Scholar
  52. 52.
    Grilli, Masciandaro and Tabellini (1991), De Haan and Sturm (1992), as well as Alesina and Summers (1993) all arrive at the same conclusion: central bank independence promotes price stability, but has no measurable impact on real output growth nor on real output variability. These studies used industrial country samples that are relevant to the European context. Although Cukierman et al. (1993) find a negative relationship between central bank independence and real economic performance in a broader 63-country sample, they find no such significant relationship in separate estimations for the industrial country and developing country sub-groups, suggesting the outcome for the full sample is capturing other differences between the two sub-groups. In a contrary view, Forder (1998) questions the procedures for the measurement of central bank independence employed by Alesina and Summers (1993).Google Scholar
  53. 53.
    Sargent and Wallace (1981).Google Scholar
  54. 54.
    Keynes(1936, p. 203).Google Scholar
  55. 55.
    An overview of the debate is provided in Fischer (1990). See also Barro (1986), McCallum (1987), Goodhart (1989) and Crockett (1994b).Google Scholar
  56. 56.
    This is essentially the position taken by M. Friedman (1968), who suggests following a passive, fixed money growth rule. McCallum (1987) also subscribes to this view, but proposes an activist rule that targets the monetary base towards a non-inflationary expansion of nominal aggregate demand; simple simulations are shown to produce good results for the United States over a period of several decades.Google Scholar
  57. 57.
    The breakthrough was made in an article by Sargent and Wallace (1975), who postulated that, with experience, a rational public would fully anticipate monetary measures and that, given this anticipation, such measures only translate into higher inflation and do not influence the real economy.Google Scholar
  58. 58.
    It has been argued that the central bank will recognise the lack of sense of inflation surprises and that the public will understand this, leading to an equilibrium around price stability. However, with discretionary central bank policy, a risk-averse public that does not know the authorities’ true inflation objective will nonetheless build an insurance premium into its inflationary expectations.Google Scholar
  59. 59.
    McCallum (1995) argues that a truly rational central bank recognises the futility of continually exploiting temporarily given expectations and thus does not behave in a time inconsistent manner. However, this assumes a sufficiently long time horizon and low discount rate, which would not seem to hold in the case of a dependent central bank governed by discretion.Google Scholar
  60. 60.
    Using a general equilibrium model, Barro and Gordon (1983a) show that reputation can to some degree substitute for policy rules. However, while their reputational outcomes are better than those under a discretionary system, they are inferior to the ‘ideal’ rules-based results of the model.Google Scholar
  61. 61.
    This is proposed by Rogoff (1985) as a way to mitigate the time-inconsistency costs of discretion, while maintaining policy flexibility to deal with unforeseen shocks.Google Scholar
  62. 62.
    Two renowned activist rules are McCallum’s (1987) rule, which limits the expansion of the monetary base to the non-inflationary growth rate of nominal demand, and Taylor’s (1993) rule, which sets the interest rate to minimise deviations from both inflation and real GDP targets.Google Scholar
  63. 63.
    Flood and Isard (1989) propose a mixed strategy involving a simple rule that can be overridden in exceptional circumstances; institutional mechanisms are put forward as a way to penalise central banks that exercise discretion in non-exceptional circumstances.Google Scholar
  64. 64.
    These first two costs are incorporated by Giavazzi and Pagano (1988) in a formal statement of how countries can gain credibility by pegging the exchange rate.Google Scholar
  65. 65.
    Drazen and Masson (1994) argue that credibility depends on both observing a tough policy line and not incurring too high unemployment costs, as this will undermine the viability of the strategy. Given this dual concept, a devaluation may on occasion actually enhance credibility. They draw a parallel with a person seeking to lose weight, who may establish credibility more effectively by at times taking small meals, than by skipping meals altogether. They find support for their hypothesis in the development of interest rate differentials in the ERM, notably between Germany and France. Masson (1995) finds similar evidence during the United Kingdom’s ill-omened ERM participation.Google Scholar
  66. 66.
    Fischer (1994) defines this autonomy as ‘instrument independence’ and contrasts it with ‘goal independence’; the latter applies to central banks that have an imprecisely defined objective and thereby enjoy greater policy latitude.Google Scholar
  67. 67.
    On this ground, McCallum (1995) contends that the beneficial effects stemming from central bank contracts are overestimated.Google Scholar
  68. 68.
    In a stark critique of the claim that institutional strengthening improves economic performance, Forder (1996) raises three methodological pitfalls: the threat of a legislative reversal of independence may undermine the purported benefits of that very same independence; established practice may determine that even a central bank with statutory independence predominantly complies with the government’s wishes; and an independent central bank may not actually behave in line with the objectives underlying its independence since it may be subject to perverse incentives (for instance, to use monetary policy to advance the electoral prospects of a party that advocates less-inflationary fiscal and public sector wage policies). While these contrarious arguments have some validity, implying that in theory statutory independence may not always promote specified central bank objectives, they do not carry much weight when account is taken of practical considerations such as the difficulty of adapting central bank legislation and the value that central bankers generally attach to their reputation.Google Scholar
  69. 69.
    De Haan and Kooi (1997) present evidence that, in practice, especially instrument independence is important for inflation performance.Google Scholar
  70. 70.
    Eijffinger (1996).Google Scholar
  71. 71.
    Stella (1997) argues that central banks with relatively weak capital bases invariably experience a decline in operational independence and abandonment of price stability as a primary policy goal.Google Scholar
  72. 72.
    The evidence suggests inflation has an adverse, non-linear effect on growth which is difficult to establish when inflation is relatively low (see Fischer 1996). In a study of inflation in 87 countries (including 13 of the current EU member states) during 1970–90, Sarei (1996) finds a structural break at 8 percent. Above this level, inflation has a very powerful negative influence on growth. Similarly, using a data set comprising 145 countries over the period 1960–1996, Ghosh and Phillips (1998) establish a negative relationship between inflation and growth that is both statistically and economically significant and survives a battery of robustness checks. They suggest the relationship is non-linear in two senses: first, at very low rates the relationship is positive (a kink is tentatively estimated at about 2½ per cent) and, second, although the cumulative effect of inflation on growth rises as inflation increases, the marginal effect becomes smaller.Google Scholar
  73. 73.
    Taylor (1981) and Chowdury (1991) provide evidence of a positive relationship between the level and the variability of inflation. Beyond this, Judson and Orphanides (1996) present tentative calculations, based on a sample of 119 countries over a 30-year period, indicating that the level and the volatility of inflation have independent, significant influences on growth. In particular, they find that inflation volatility is significantly negatively correlated with income growth across time, type of country and level of inflation.Google Scholar
  74. 74.
    Committee for the Study of Economic and Monetary Union (1989, pp. 25–26). This requirement of institutional independence had not been specified in the earlier proposal for an economic and monetary union in Europe embodied in the Werner Report finalised in October 1970.Google Scholar
  75. 75.
    A comprehensive overview of the theoretical literature can be found in Cukierman (1992). Outcomes and reviews of the empirical literature are provided in Alesina and Summers (1993), Cukierman et al. (1993), De Haan and Sturm (1992), Eijffinger and Schaling (1993), Eijffinger and De Haan (1996), and Grilli, Masciandaro and Tabellini (1991). Of course, these studies are subject to the caveat that empirical correlation need not imply causation: it has been suggested that the correlation may, for instance, primarily reflect a third factor such as the inflation aversion of the population; see Posen (1993). On a more specific point, Forder (1998) has challenged the measurements of independence, and therefore the results, presented by Alesina and Summers (1993).Google Scholar
  76. 76.
    See Eijffinger and De Haan (1996) and Bini Smaghi (1998). However, the notion that accountability and central bank independence can be two sides of the same coin is at odds with the empirical finding of Briault, Haldane and King (1996) that these elements are actually negatively related. Using a more refined accountability index, De Haan, Amtenbrink and Eijffinger (1998) have confirmed this inverse relationship, establishing that less independent central banks tend to have heavier communication obligations and, especially, are subjected to greater disciplinary mechanisms vis-à-vis the government and Parliament. But this seems to reflect the fact that accountability and transparency have been pursued more vigorously by central banks with weak track records and low credibility-these have typically also been the less independent central banks. In these cases, accountability has therefore largely reflected an initial mistrust of the monetary authority. Over the course of time, this accountability and openness is likely to elicit commensurate policy independence for the central bank, and the inverse relationship may be expected to disappear.Google Scholar
  77. 77.
    Formal analyses of the optimal contract for a central bank governor are presented in Walsh (1995) and Persson and Tabellini (1993).Google Scholar
  78. 78.
    Cukierman and Meltzer (1986) highlight the other side of the coin, showing that greater ambiguity makes it easier for the policy-maker to engage in positive surprises when the need for stimulation is seen to prevail and to postpone negative surprises for periods in which inflation concerns dominate. Given the greater uncertainty and higher inflation (expectations) that such ambiguity would entail, as well as the difficulties it would create for central bank accountability, this would be an ill-advised way of increasing monetary policy flexibility.Google Scholar
  79. 79.
    See in particular Persson and Tabellini (1993). The key role of preannouncement in backward-looking assessments of monetary policy is also emphasised by Bernanke and Mishkin (1992) and Crockett (1994b).Google Scholar
  80. 80.
    The gradual tightening of the money and credit objectives could theoretically be associated with accelerating money velocity — and thus not with declining inflation — so that the lowering of the target did not genuinely reflect greater policy ambition. However, in practice broad money velocity has declined in most EU countries over the past quarter century.Google Scholar
  81. 81.
    In 1986 and 1987 the German nominal effective exchange rate appreciated by an average 9 and 6 per cent, respectively. As a result, it was possible to increase the upper end of the money target range by one-half of one percent in each of these two years, but to base this on a lower implicit inflation target (set at 2 percent for both 1986 and 1987, against 2½ percent for 1985).Google Scholar
  82. 82.
    Tobin (1983) advocates simple monetary policy rules on complementary grounds, emphasising the political appeal of simplicity. He draws parallels with the attractiveness of a balanced budget norm and with the powerful imperative of gold convertibility at historic parity.Google Scholar
  83. 83.
    See also Bernanke and Mishkin (1992).Google Scholar
  84. 84.
    An example of transparency that would likely cause confusion is if, on account of differences of opinion within the policy setting body, the central bank published more than one inflation forecast under a direct inflation targeting strategy. This is not far-fetched fiction: the Bank of England’s Monetary Policy Committee has indicated that it will report separate inflation forecasts if its individual members disagree about the validity of certain aspects underlying the forecasts; see Budd (1998, p. 1793).Google Scholar
  85. 85.
    This point is made by Goodfriend (1986) in a comprehensive review of standard central bank arguments for secrecy in the implementation of monetary policy. His assessment is based on the US Federal Reserve Bank’s defence of secrecy as argued in the 1975–81 court proceedings against the Federal Open Market Committee under the Freedom of Information Act. On balance, he finds the theoretical arguments for secrecy to be at best inconclusive.Google Scholar
  86. 86.
    Even the minutes of the meetings of the Bank of England’s Monetary Policy Committee (renowned for its transparency creed) do not attribute individual contributions to policy discussions. This is done in order to ensure that individual members do not feel restricted in advancing arguments that they may not personally support, but do wish to have fully explored. See Budd (1998, p. 1789).Google Scholar

Copyright information

© Springer Science+Business Media Dordrecht 2000

Authors and Affiliations

  • Aerdt C. F. J. Houben
    • 1
  1. 1.De Nederlandsche BankAmsterdamThe Netherlands

Personalised recommendations