Switzerland: Selected Issues and Statistical Appendix
Author: International Monetary Fund
Publication Date:
05 May 1999
eISBN:
9781451807165
ISBN:
9781451807165
Language:
English
Keywords:
ISCR; CR; tax burden; income tax; activistic monetary policy; monetary policy success; feedback period; central bank; SNB's monetary policy formulation; Exchange rates; Monetary base; Inflation targeting; Demand for money
Download PDF (1.4 MB)

This Selected Issues paper and Statistical Appendix compares two alternative time series approaches to analyzing Switzerland’s recent business cycle experience: first, the traditional “smooth-trend-plus-cycle approach,” which envisages observed output growth as fluctuating around a relatively smooth potential output growth path; and, second, the more recently developed “regime change approach,” which views business cycles as shifts between “high-growth” states (expansions) and “slow-growth” states (recessions) of the economy. The paper also examines Switzerland’s monetary policy framework, and describes the challenges to the Swiss tax system.

Abstract

This Selected Issues paper and Statistical Appendix compares two alternative time series approaches to analyzing Switzerland’s recent business cycle experience: first, the traditional “smooth-trend-plus-cycle approach,” which envisages observed output growth as fluctuating around a relatively smooth potential output growth path; and, second, the more recently developed “regime change approach,” which views business cycles as shifts between “high-growth” states (expansions) and “slow-growth” states (recessions) of the economy. The paper also examines Switzerland’s monetary policy framework, and describes the challenges to the Swiss tax system.

II. Switzerland’s Monetary Policy Framework18

A. Introduction and Summary

24. Since the breakdown in the 1970s of the Bretton Woods system of fixed exchange rates, monetary aggregates have played a key role in guiding Swiss monetary policy. In all but one of the last 25 years, the Swiss National Bank (SNB) has established a monetary target as the intermediate objective for monetary policy. Moreover, during this period, the Swiss monetary authorities have remained faithful to narrow monetary aggregates (Ml until 1978 and MO since 1980) as intermediate targets, and it is only very recently that the SNB has increasingly focused on M3.

25. The SNB has, however, been pragmatic in applying its monetary targeting framework. Although the monetary targets were largely met during relatively stable periods, the SNB has allowed monetary growth to deviate considerably from the target path when other concerns were considered more important. In particular, the exchange rate has at times played a decisive role for the stance of monetary policy. This reflects the relative openness of the Swiss economy and the large international financial sector. In the late 1970s, for example, significant upward pressures on the Swiss franc led to a temporary switch in the SNB’s strategy toward setting an explicit floor on the Sw FYDM exchange rate. Moreover, the SNB has often accommodated large unexpected shifts in money demand, without attempting to bring the stock of money back to its medium-term target level.

26. Swiss monetary policy has been successful in delivering a low level of inflation over the past 25 years. Indeed, Switzerland’s average inflation performance is second only to that of Germany among the advanced industrial countries. Moreover, the variability of Swiss real GDP growth has not been larger than the average in comparable countries (Figure II-1). However, the relative performance of Swiss monetary policy does seem to have deteriorated since the second half of the 1980s, reflecting to a large extent increasing volatility of money demand triggered by structural changes in the financial system.

Figure II-1.
Figure II-1.

Switzerland: Monetary Policy Performance in Industrial Countries, 1973-97

Citation: IMF Staff Country Reports 1999, 030; 10.5089/9781451807165.002.A002

Sources: IMF, World Economic Outlook database and staff calculations.

27. The future challenges to Swiss monetary policies are many. First and foremost, continued financial innovation and potential currency substitution are likely to influence money demand in art unpredictable fashion. Due to the more endogenous nature of broader monetary aggregates and their complicated relationship to interest rate developments, it is not certain that the increased use of M3 as an alternative indicator will prove an effective alternative to a monetary targeting framework based on base money. Furthermore, in the early years of the euro, portfolio preferences and the perception of risk could change rapidly. In particular, the Swiss franc may be affected by perceptions of strength or weakness of the euro or any other uncertainties regarding the economic or political prospects in the euro area. Switzerland may also continue to be subject to “safe haven” capital flows in the context of global financial tensions. The concluding section considers the pros and cons of alternative strategies in response to exchange rate pressures or persistent instability of money demand.

B. The SNB’s Monetary Framework19

28. The Swiss monetary policy framework has been described as “pragmatic monetarism,”20 “disciplined discretion,”21 or “a monetary policy rule with an escape clause.”22 All of these expressions try to capture the mixed nature of the Swiss monetary policy framework. On the one hand, monetary targets have played a central role, both as a framework for interpreting the macroeconomic stance of monetary policy and as a tool used to communicate monetary policy intentions to the public. On the other hand, the annual or medium-term target has not been regarded as an ironclad law; the SNB has on occasion allowed monetary growth to deviate by a large amount from the target path, either accommodating a shift in the demand for money or attempting to counteract deflationary pressures from a sharp appreciation of the exchange rate.

29. As already noted, in Switzerland a key limitation of following a strict monetary targeting framework is related to the disturbing effects of large exchange rate movements. Owing to the perception of Switzerland as a safe haven, and to its relatively small size, international portfolio adjustments can result in significant pressures on the exchange rate stemming from capital flows. Moreover, given Switzerland’s relative openness to international trade—the share of exports in goods and services in GDP is about 40 percent—a sharp appreciation of the currency can have significantly negative, and potentially long-lasting, effects on the economy. As a consequence, there have been substantial pressures to limit such large exchange rate movements by using monetary policy.

30. The use of a monetary growth target as a key element in the SNB’s monetary policy strategy dates back to the early 1970s, when the collapse of the Bretton Woods system of fixed exchange rates forced the Swiss authorities to float the Swiss franc. Inspired by the monetarist thinking of the time, the SNB announced at end-1974 a target growth rate for Ml for the following year (Table II-1). In order to achieve the medium-term inflation target of 1 percent annual inflation, the monetary overhang inherited from the fixed exchange rate period was scheduled to be lowered gradually by reducing monetary growth from 6 percent in 1976 to 2 percent per annum in the medium term. In the first three years of operation, the growth targets were missed by only small margins, and the inflation rate dropped quickly, reaching about 1 percent in 1976. Thus, not only in terms of keeping monetary targets but also in terms of the ultimate objective of monetary policy, the early years of the monetary targeting framework were considered a great success.

Table II-1.

Switzerland: Targeted and Realized Monetary Growth, 1975-98

(Annual percent change)

article image
Sources: Rich (1997); SNB; and staff estimates.

M1 Currency, and demand deposits with banks and the postal giro system, held by the nonbank public. For M1 only end-of-month data are available.

MO Monetary base, defined as the sum of banks’ deposits with the SNB and the aggregate bank no circulation. Until the end of 1988, adjusted for the end-of-month bulge in SNB credit to banks; from 1989 onwards, seasonally adjusted. Data represent monthly averages of daily figures.

Arithmetic mean of monthly year-on-year growth rates.

Arithmetic mean of monthly year-on-year growth of seasonably adjusted M1.

Arithmetic mean of annualized monthly growth rates over November 1979 level.

Arithmetic mean of annualized monthly growth rates over level in the fourth quarter of the preceding year.

Growth rate in the fourth quarter over the level in the fourth quarter of the preceding year.

Average of the five annual growth rates shown below.

Average of the four annual growth rates shown below.

31. In the late 1970s, pressures to abandon the monetary targeting framework became very strong after a spectacular appreciation of the Swiss franc; in trade-weighted terms the nominal effective appreciation amounted to some 80 percent (Figure II-2). To a certain extent the effect of this appreciation on competitiveness was partly offset by the relatively better Swiss inflation performance, but even in real terms (corrected for relative inflation differences) the trade-weighted value of the Swiss franc appreciated by about 40 percent, half of which took place in the latter half of 1977 and 1978. This surge in the exchange rate became a primary policy concern. In the fall of 1978, the SNB decided to suspend its monetary target strategy and introduced a temporary floor on the exchange rate’s value against the deutsche mark at 0.80 Sw F/DM. Although the exchange rate target was underpinned by temporary capital controls, the SNB’s interventions led to rapid monetary growth, and the target for 1978 was overshot by a large margin (see Table II-1 and Figure II-3).

Figure II-2.
Figure II-2.

Switzerland: Exchange Rates, 1973-98

Citation: IMF Staff Country Reports 1999, 030; 10.5089/9781451807165.002.A002

Sources: Swiss Institute for Business Cycle Research, data tape; IMF International Financial Statistics database; OECD Analytical Data Bank.1/ All major trading partners, except Korea, Taiwan Province of China, and Hong Kong.2/ Based on relative consumer prices.
Figure II-3.
Figure II-3.

Switzerland: Monetary Aggregates

Citation: IMF Staff Country Reports 1999, 030; 10.5089/9781451807165.002.A002

Sources: IMF, World Economic Outlook database; Swiss National Bank, Monthly Bulletin.1/ One percent per annum from 1989, fourth quarter, and 1994, fourth quarter, respectively.2/ Seasonally adjusted; in billions of Swiss francs.

32. As a result of the sharp monetary expansion and subsequent fears about its inflationary consequences, monetary targets were reintroduced in place of the exchange rate floor in late 1979. The targeted monetary aggregate was changed from Ml to the monetary base (MO), as this was believed to be more directly under the SNB’s control and less influenced by relative interest rate developments. The monetary base was also much closer than Ml to the daily operations of the SNB, since Ml includes sight deposits in commercial banks, which have to be projected on the basis of information on the stock of base money, interest rates, and domestic demand. By contrast, the size of the monetary base can be controlled more directly by the central bank. As in the early years of monetary targeting, the annual growth targets stipulated a gradual reduction of annual monetary growth from 4 percent in 1980 to 2 percent from 1986 and onward. In the event, the target was significantly undershot in 1980 and 1981; in the following six years, by contrast, the deviations from the targeted growth path were only small, coupled with a gradual reduction in inflation and an expanding economy. At the same time, both the real exchange rate and the bilateral exchange rate vis-à-vis the deutsche mark were stable, thus allowing monetary policy to be conducted without an urgent need to counteract pressures on the exchange rate.

33. This second “honeymoon” of the Swiss monetary targeting framework ended abruptly in 1988 with the introduction of the Swiss Interbank Clearing (SIC) system, a new electronic interbank payment system based on “real-time gross settlement.” The introduction of this system—coupled with a modification of bank cash reserve requirements—led to a sharp reduction in banks’ demand for base money. The SNB was well aware of the money demand shift, but obtaining even an approximate estimate of the shift proved difficult to produce ex ante. Thus, other indicators such as the resulting effects on interest rates had to be followed in order to gauge the magnitude of the demand shift, potentially delaying the SNB’s response.

34. This episode was followed by a period of rapid inflation. Views diverge, however, on whether this inflationary spurt was a result of the problems with measuring money demand, or whether the economic boom in the late 1980s was to blame. The role of monetary technicalities should not be overstated: inflationary pressures had been building up for some time before the demand shift, and it was commonly expected that the world economy would turn into recession in 1988. In any case, the introduction of the new clearing system came at a particularly inopportune moment in the business cycle.

35. Since the early 1990s, the demand for base money has shown considerable instability. Partly as a result of the large demand shift in 1988, the SNB switched in 1989 to five-year growth target paths for the monetary base, stressing that it would not attempt to force the monetary base to respect the target path every year. Reflecting increased financial innovations, the medium-term growth target was reduced to 1 percent (from 2 percent previously): the trend increase in velocity was assumed to increase by 2 percent per annum together with a medium-term inflation target of 1 percent and potential real GDP growth of 2 percent.

36. These medium-term target paths have functioned more as yardsticks than strict target paths: deviations from the target path have then been explained in terms of demand shifts or, alternatively, as an indication of monetary policy intentions. Medium-term target paths are quite demanding, however, since there is no annual base drift as in the case of annual growth targets; one year’s overshooting should in principle be compensated by slower monetary growth in the subsequent years.

37. From the introduction of the medium-term target paths in 1988, the SNB has been faced with the need to explain large deviations from the target path. In the first 5 years, the monetary base undershot the target path substantially, reflecting, inter alia, an initial underestimation of the shift in money demand in 1988. Thus after an initial drop in the monetary base, the SNB constrained monetary growth to follow the medium-term growth path in parallel without attempting to return to the medium-term path. With the shift to monetary relaxation in 1996, the situation changed dramatically: both the level of the monetary base and its growth rate have since then significantly overshot the medium-term target. Several special factors, however, appear to explain a large share of the rapid growth in base money: in particular, the increased demand for high-denomination bank notes, the introduction of a repo market, and the merger of the two largest banks all appear to have led to unexpected increases in the demand for base money. At the same time, although it appears possible to trace the large shifts in the demand for base money to specific events, there is a large degree of uncertainty attached to both the durability and the size of these demand shifts.

38. As a result of these large and frequent demand shifts, other indicators of monetary conditions have gained increased importance in the implementation of Swiss monetary policy. In particular, in the December 1997 announcement of the monetary policy intentions for 1998, the SNB signaled an increased reliance on M3 as an alternative monetary indicator, indicating that an annual expansion of M3 in the range of 4 percent was considered to be consistent with medium-term price stability. Research at the SNB suggests that the demand for M3 has been relatively stable: for example, Peytrignet and Stahel (1998) show that M3 is cointegrated with the level of consumer prices, real GDP, and the return on government bonds and that there were no structural breaks during 1977-96. Moreover, the SNB estimates that the mean time lag between an increase in growth of M3 and inflation is about 39 months.

39. It is inevitable, however, that in a situation of large instability in money demand, price indicators—such as short-term and long-term interest rates—acquire a greater importance in the operational implementation of monetary policy23 (Figure II-4). The introduction of a broad monetary aggregate is unlikely to alter the trend to the use of a broader set of indicators. Although broad monetary aggregates appear to exhibit a larger degree of long-term stability than base money, broader monetary aggregates are highly sensitive to relative interest developments (the difference between the “own rate” and other interest rates) and to financial innovation. As conventionally defined, M3, for example, does not include highly liquid money market funds or other mutual funds that may be close substitutes to time deposits. In this context, it is noticeable that the recent unexpectedly slow growth of M3 (only 1 percent in 1988) has been attributed by some observers to the increased prevalence of mutual funds in Switzerland, less than one year after M3 was introduced by the SNB as “supplementary indicator”.

Figure II-4.
Figure II-4.

Switzerland: Interest Rates, 1973-98

Citation: IMF Staff Country Reports 1999, 030; 10.5089/9781451807165.002.A002

Sources: IMF, Surveillance database; IMF, International Financial Statistics database.1/ Three-month euro rate.

C. Monetary Policy Feedback Rules

40. It can be concluded from the above discussion that Swiss monetary policy may not be easily characterized by a simple monetary growth rule, nor can the SNB’s escape clauses be easily identified in ex ante terms. The monetary growth target has been frequently over or undershot by a large margin. While the SNB has carefully explained such deviations ex post, in terms of demand shifts or discretionary monetary policy, the exact conditions that would trigger the use of discretionary monetary policy have not been spelled out in any detail. In the case of large exchange rate movements, reference has been made to exchange rate developments that “threaten to seriously damage economic growth,” without specifying the size or level of the qualifying exchange rate movements, or by how much the SNB would respond. A formal investigation confirms these impressions and indicates that the SNB’s observed behavior as implied by an estimated monetary policy “reaction function” was not significantly different from that of other central banks with less formal commitments to monetary targets.

41. Based on casual observation of the behavior of the U.S. Federal Reserve System, a simple feedback rule was proposed by Taylor (1993). Taylor proposed this rule not so much as a rigid prescription for policy formulation but as a benchmark rule to be used as a starting point for monetary policy decisions, which in practice would take account of a wide set of indicators and also make allowance for “special factors.” Later researchers have shown that in a simple closed economy model, variants of the Taylor rule would be optimal by minimizing the weighted average of inflation and output variability (Ball, 1997 and Svensson, 1997).

42. Figure II-5 depicts the level of the nominal short-term interest rate as predicted by this rule during 1973-98. The rule simply sets the nominal short-term interest rate rt as a function of inflation dpt, half the output gap GAPt, and half the deviation of inflation from targeted inflation IIt (the “inflation gap”), and a constant term equal to the steady state real interest rate (rr*):
(1)rt=dpt+0.5GAPt+0.5Πt+rr*.
Figure II-5.
Figure II-5.

Switzerland: Short-Term Interest Rates and the Simple Taylor Rule, 1974-98

(in percent)

Citation: IMF Staff Country Reports 1999, 030; 10.5089/9781451807165.002.A002

Sources: SNB, Central Statistical Office and staff estimates.1/ Percent difference between actual real GDP and potential GDP.

In Figure II-5, the steady state real interest rate was set at 1 percent (compared with 2 percent in Taylor, 1993, who applied the model to the United States).

43. Comparing the short-term interest rate as predicted by the simple Taylor rule and the actual observed interest rate produces a reasonable fit (Figure II-5). At first sight, this is particularly noteworthy since the SNB does not use the interest rate as the primary operational target in the same way as, for example, the U.S. Federal Reserve System. Furthermore, the SNB has explicitly stated that its ultimate objective is price stability and that stabilizing output is not a primary policy objective.

44. A closer look at the relationship between the Taylor rule and observed short-term interest rates, however, reveals some important differences. In both of the high interest and high inflation periods (early and late 1980s), the Taylor rule would have predicted a sharper increase in the interest rate than actually took place. Furthermore, short-term interest rates would have decreased more slowly in the early 1970s, increased earlier in the late 1970s, and remained much lower since 1974. These differences cannot be easily ignored since they have tended to persist in time (i.e., they are “serially correlated”) and have been as large as 5 percent.

45. In order to examine the circumstances under which the simple Taylor rule is consistent with a strict monetary targeting framework, it is useful to determine whether a variant of the Taylor rule, under certain conditions, can be derived from a strict monetary targeting rule. Assuming that there exists a stable money demand function and expressing the demand for the natural logarithm of real base money as a linear function of real output and short-term interest rates, expressed as deviations from their trend levels:
(2)(mt−mt*)−(pt−pt*)=β(yt−yt*)+γ(rt−dpt)−γ(rt*−dpt).
In the case of a strict linear monetary growth rule, (mt − mt*) = 0 and the equation becomes:
(3)rt=rt*−β/γ(yt−yt*)−1/γ(pt−pt*)

According to equation (3), the short-term interest rate can be expressed as a function of the steady state nominal interest rate, the deviation of prices from its trend level (a price level target), and the output gap, with all variables being contemporaneous. Therefore, the monetary targeting rule differs from the Taylor rule by the fact that—in its strictest interpretation—it implies that the central bank responds to deviations of prices from the long-term price level, as opposed to its change (the level of the inflation rate).

46. The observed reaction function of a central bank following a monetary targeting approach may differ from (3) in several respects, however. First, shifts in the money demand function would warrant a more flexible approach to monetary targeting, using other indicators such as interest rates and the exchange rate to determine the appropriate level of interest rates. Second, the monetary authorities may correct for the effects of tax changes, direct exchange rate related price increases, and supply-side shocks. Third, pragmatic monetary targeting would also permit—in the case of large over or undershooting of monetary target—some “base drift” in the monetary target; not allowing for occasional base drifts would imply accepting periods of significant deflation possibly combined with a sharp economic slump in excess of what would be necessary to bring inflation—i.e., the rate of change of the price level—down to zero.

47. In order to investigate empirically whether the Swiss monetary policy framework can be described as following a monetary policy targeting rule or a Taylor rule representation (3) was reformulated by using the fact that the current price level pt can be expressed as a function of last period’s price level pt-1 and the price change from last period dpt:
(4)pt=dpt+pt−1.
Combining (3) and (4), and noting that Πt=(dpt−dpt*) and GAPt=(yt−yt*) we obtain:
(5)rt=rt*−β/γGAPt−1/γ(pt−1−pt−1*)−1/γΠt.
Thus, assuming that rt* is constant over the estimation period, we can estimate the following econometric model:
(6)rt=α+β1GAPt+β2(pt−4−pt−4*)+β3Πt+εtandεt∼iid(0,σε).

From (5) it follows that β2 = β3 in the case of a strict monetary targeting rule. Thus, the monetary targeting rule can be tested against a Taylor rule specification by testing whether such a restriction can be rejected. The price level target Pt* was assumed to grow by 1 percent per year. The model was estimated using quarterly data over the period 1975Q1-1998Q2 by ordinary least squares (OLS). Since inflation was measured as the average of quarterly year-on-year growth rates, the price level was entered in the regression lagged one year (four periods). The three-month euro rate for Swiss franc deposits (conventionally annualized) was entered as the dependent variable.

48. The regression results can be found in Table II-2. In total, three different versions of the model were estimated; while the two first models AO and Al use the headline consumer price index, model A2 uses an estimate of underlying or “core” inflation produced by the SNB. The “core” price consumer price index differs from the headline index by excluding “extreme components” (defined as the 15 percent largest price increases and 15 percent largest price reductions) from the index and making an ad hoc correction for the 1995 VAT increase. Models Al and A2 differs from the first model AO by the inclusion of the lagged price (level) gap (pt-4 − pt-4*). A dummy was introduced to correct for the large shift in velocity in 1988 with the introduction of the Swiss Interbank Clearing system (SIC).

Table II-2.

Switzerland: Feedback Rule Estimations, 1975.Q1-1998.Q2

(Standard errors in parentheses)

article image
Source: Staff estimates.

A core price index was used in model A2. The core price index was calculated by cutting out 15 percent of the largest negative and 15 percent of the largest positive changes in the consumer price index (two-sided “trimmed mean”). A correction was also made for the 1995 VAT increase.

49. Model A0 is the conventional Taylor rule formulation. It is noticeable that the coefficient on the output gap is very large, while the coefficient on the inflation gap is significantly below the coefficient implied by the Taylor rule (1.5) in equation (1). The low coefficient on the inflation gap is intriguing: the result suggests that an increase in inflation is not fully met by an equal increase in nominal interest rates, potentially reflecting a high degree of inflation credibility; to the extent that inflation surprises do not translate into a change in inflation expectations, there is not necessarily a need to respond to an increase (or fall) inflation by a similar increase (or fall) in nominal interest rates, unless there are underlying inflationary (or deflationary) pressures already captured in the output gap.

50. The inclusion of the lagged price gap in model Al does not significantly alter the estimation results: the estimated coefficient on the output gap was only slightly reduced, while the coefficient on the inflation gap remains significantly below the Taylor rule coefficient. The coefficients on the lagged price gap and the inflation gap are clearly different, however, and their equality can be formally rejected both in model Al and in model A2. Thus, the estimation results support the hypothesis that the SNB has not been following a strict monetary targeting rule over the estimation period, even when allowing for the 1988 episode. It is also noticeable that the use of core inflation results in more plausible estimates of the Taylor rule coefficients: the estimated coefficient on the output gap was 0.2 and coefficient on the inflation gap increased to 1.1.

51. The results reported above are broadly in line with the results from a recent paper by Dueker and Fischer (1996). They characterize Swiss monetary policy as following a feedback rule where monetary growth is set as a function of projected growth of real base money and an inflation target, but with an occasional exchange-rate feedback. The model was estimated using discrete “switching parameters” with a total of four different states: high exchange-rate feedback (with low and high inflation targets); and low exchange-rate feedback (with low and high inflation targets). They find that including a price-level target in the model does not improve its goodness-of-fit; removing the exchange rate feedback, however, would significantly reduce the model’s explanatory power. Only one high exchange-rate feedback period—in 1978-79 (the period when an exchange rate floor was in place)—is clearly identified by the model (see Dueker and Fischer, 1998), but the exchange rate feedback is also significant in the low inflation state.

52. It is clear that none of the feedback rules above fully captures the decision-making process of the SNB. However, it is interesting to note that the feedback rule implied by a strict monetary targeting strategy is clearly rejected. Thus, although monetary targeting has been successfully followed in some periods, the exceptions to strict monetary targeting and the presence of “base drift” have been frequent. The estimations suggest that the practice of the SNB’s monetary policy formulation might then not differ too much from the more recent practice in countries pursuing formal inflation targeting. Rather, the difference lies with the “rhetoric of monetary policy” as well as the communication of the policies to the markets and the public in general.

D. Past Performance of Swiss Monetary Policy

53. Policy discussions often assess monetary policy success solely in terms of the average inflation rate and make only secondary references to the variability of inflation or output. Academic literature by contrast often assesses monetary policy performance as an explicit trade off between output and inflation variability: in these terms monetary policy may be inefficient if by changing the policy rule inflation variability can be reduced without increasing output variability (or the reverse). Given that the policy pursued is efficient, the particular combination of inflation and output variability depends on whether the central bank attaches a high degree of importance to low inflation variability or whether low output variability is more important. The mandates of the major central banks differ on this issue: while the objectives given to the newly created European Central Bank states “price stability” as its sole monetary policy objective, the mandate of the Federal Reserve Board explicitly acknowledges that stabilizing output is an important objective for monetary policy. The present mandate of the SNB states in broad terms that it should conduct a monetary policy “in the general interest of the country.”24

54. Figure II-1 presents in two panels the three dimensional trade-off between average inflation and inflation variability (panel 1) and the variability of output and inflation (panel 2). Although countries might have different preferences on the relative importance of these dimensions, the figures provides the basis for a rough ranking of monetary policy success: countries close to the origin have clearly been successful whereas countries far away from the origin of the figures have been less successful.

55. A closer inspection of the figures reveals that Swiss monetary policy was very successful in keeping inflation low in the period 1973-1985, a period when many industrial countries experienced high and highly variable inflation. None of the other countries in the selection had lower average inflation rate in this period, although the average inflation rate in Germany was only marginally higher. This achievement may, however, to some extent have been obtained at the expense of high output variability (second panel), although three countries—the United States, Sweden and the Netherlands—had higher output variability without having a lower average inflation rate. Inflation variability was about average.

56. The outstanding Swiss inflation performance in the first subperiod seems to have broken down in the second period. Although the average inflation rate was much lower than in the first period, as many as five countries including France and Belgium—had a lower average inflation rate in the second period. At the same time, Swiss inflation variability—measured by its average standard deviation—was only exceeded by Sweden. This result is clearly dominated by the short period of high inflation in the early 1990s, following the large money demand shift in 1988. In response to the increasing inflation, the SNB tightened monetary policy rapidly and succeeded in reducing the inflation rate from approximately 5 percent to close to nil in three years.

E. Future Challenges

57. Swiss monetary policy is likely to face many challenges in the future. The effects of the move to monetary union by its neighbors are far from the only challenge which is likely to face the Swiss monetary authorities. The experience of other countries suggests that the pace of financial innovation is likely to continue with undiminished strength; the introduction of liquid mutual funds is just one example that is blurring the distinction between monetary assets and investment vehicles. Although it is possible to adjust monetary growth for particular demand shifts, the large number and complexity of these demand shocks would threaten the viability of a strict monetary targeting framework. A switch to a broader and more stable monetary aggregate such as M3—which is currently used as an additional intermediate indicator—would be helpful from this perspective. The direct link to SNB’s balance sheet would, however, be lost, and M3 data become only available with a lag. Furthermore, in contrast to the supply of base money, which can to a large degree be directly influenced by the monetary authorities, the supply of broad money is endogenous and tends to vary pro-cyclically.

58. Although the introduction of the euro in January 1999 has not had significant effects on the Swiss franc, the period leading up to the final replacement of national currencies by euro notes and coins in 2002 could still result in volatile demand for Swiss francs. On the one hand, the elimination of several European currencies could result in an increased demand for Swiss francs from institutional investors seeking to diversify their portfolios. On the other hand, however, increased currency substitution by Swiss firms and households, and sharply improved expectations about the stability of the euro could result in unexpected outflows of capital and downward pressures on the exchange rate. Although the Swiss franc has often experienced periods in the past where it has been exposed to the shifting sentiments of international investors, the international demand for Swiss franc assets could become more fragile in the first years after the introduction of the euro. If it were to become important, currency substitution would also add to the instability of money demand, and further undermine the monetary targeting framework.

59. A previous study conducted by the staff discussed in some detail the effects of portfolio preference shifts in favor of the Swiss franc.25 The main conclusion of the discussion was that timely monetary policy action—with monetary expansion temporarily exceeding the monetary growth target—would be effective in reducing the short-term negative effects of an upward pressure on the currency resulting from a portfolio preference shift (both temporary and persistent). Activistic monetary policy is, nevertheless, not likely to completely prevent an initial drop in output, as the contractionary effect of the appreciation of the exchange rate more than offsets the positive stimulus from lower Swiss interest rates. It would also result in a temporary increase in consumer price inflation. A temporary exchange rate floor to counter appreciation would, according to the study, have a stabilizing effect on output, provided that the monetary authorities accommodated some of the initial appreciation of the exchange rate.

60. A radically different way to approach the problem of portfolio preference shifts would be to fix the exchange rate irrevocably to the euro. If credible, this would have the advantage of eliminating currency risk attached to trade between Switzerland and the euro zone. Such a peg would, however, result in the removal of any monetary policy independence: monetary policy would then be effectively determined in the European Central Bank (ECB) without any regard to Swiss economic conditions. Given the decentralized Swiss fiscal structure, it is unlikely that counter-cyclical fiscal policy would compensate for the loss of monetary policy autonomy.

61. Accordingly, by providing the possibility of “leaning against the wind,” the present monetary policy framework could be useful in the face of any financial turbulence attached to the introduction of the euro; the negative upward pressures on the exchange rate could be partly offset by some monetary expansion. It should also be noted that even a constant monetary growth target provides some automatic counter-cyclical stimulus through the negative effect on interest rates from slower economic growth. At the same time, the SNB could, if desired, offset any Swiss-specific domestic demand shocks.

62. The introduction of an inflation targeting framework could be an alternative to both the irrevocable fixing of the exchange rate and the monetary targeting framework. In such a framework, an explicit medium-term inflation projection would need to be produced, conditional on the developments in a range of factors. To the extent that monetary aggregates have provided important information about future inflation, such aggregates would figure prominently in the framework, but other indicators could also be included. This inflation projection could in turn be used to provide an anchor for private sector inflation expectations. In many respects, such a framework would not differ much from the framework that has been used since 1980, when the annual monetary targets were replaced by a medium-term monetary growth path, but it opens the way for the formal introduction of a broader set of indicators. The inflation targeting framework is also more forward-looking, as it focuses on likely inflation over a given future horizon.

63. In spite of its recent popularity, there are several problems attached to the introduction of an inflation targeting framework. Frequent over- or undershooting of the inflation target might undermine policy credibility, even if the central bank were to focus on “underlying inflation.” A dramatic change in fiscal policy stance or other unexpected events, not taken into account in the “special adjustments,” could affect the inflation outcome. Consequently, if a precise target range is established, the strategy could easily prove counterproductive. Finally, the exchange rate could complicate the implementation of an inflation targeting framework in much the same way as it has complicated the pursuit of the monetary targeting strategy. Large exchange rate movements are likely to have significant effects on developments in domestic prices, even when the direct effect of foreign prices is removed: the indirect effects on domestic prices could cause large over- or undershooting of the inflation target. Finally, inflation targeting requires the production of explicit inflation projections with long time horizons. In the Swiss case, this horizon may have to extend to three years due to a particular long estimated time-span from an increase in monetary growth to an increase in inflation (in most other countries the time-span is estimated to be about two years).

64. Given the instability of the demand for base money, monetary base targeting, the classic “low-cost strategy,” has an uncertain future. The introduction of “alternative indicators”—such as M3 and the yield curve—and using these indicators to allow base money growth to stray away from its target path by a significant amount implies, however, the establishment of an alternative assessment of medium-term inflation pressures. Such an assessment requires—at least implicitly—the determination of the relative importance of the different indicators for a range of possible inflation outcomes. Thus, even in the absence of formal inflation targeting, a high-quality interpretation of available economic data has to be ensured. What is different, however, with an inflation targeting framework (e.g., as practiced by the UK, Canada, New Zealand, and Sweden) is the production of explicit inflation forecasts, stipulating clearly underlying assumptions as well as the uncertainties attached to the forecast. Making such contingency forecasts of inflation available to the public, policy transparency could be significantly enhanced, and the public’s uncertainty about policy actions could be reduced. The experience of countries which have pursued inflation targets for some time suggests that the presentation of such a forward looking assessment of inflation pressures has been a key element in improving the policy dialogue between the public and the monetary authorities.26

  • Ball, Laurence, 1997, “Efficient Rules for Monetary Policy,” National Bureau of Economic Research NBER Working Paper Series, No. 5952.

    • Search Google Scholar
    • Export Citation

  • Ball, Laurence, 1998, “Policy Rules for Open Economies,” National Bureau of Economic Research NBER Working Paper Series, No. 6760.

  • Bernanke, Ben S., and Frederic Mishkin, 1997, “Inflation Targeting: A New Framework for Monetary Policy?” Journal of Economic Perspectives, Volume 11, No. 2, pp. 97–116.

    • Search Google Scholar
    • Export Citation

  • Bernanke, Ben S., and Frederic Mishkin, 1992, “Central Bank Behavior and the Strategy of Monetary Policy: Observation from Six Industrial Countries,” National Bureau of Economic Research Working Paper Series, No. 4082.

    • Search Google Scholar
    • Export Citation

  • Camen, Ulrich, Hans Genberg, and Michael Salemi, 1990, “Optimal Monetary Policy and the Revealed Preference Function of the Swiss National Bank,” in Monetary Policy: a Theoretical and Econometric Approach, ed. by P. Artus and Y. Barroux, Kluwer Academic Publishers.

    • Search Google Scholar
    • Export Citation

  • Dueker, Michael, and Andreas M. Fischer, 1996, “Inflation Targeting in a Small Open Economy: Empirical Results for Switzerland,” Journal of Monetary Economics, No. 37, pp. 89–103.

    • Search Google Scholar
    • Export Citation

  • Dueker, Michael, and Andreas M. Fischer, 1998, “A Guide to Nominal Feedback Rules and Their Use for Monetary Policy,” Federal Reserve Bank of St. Louis Review, July/August.

    • Search Google Scholar
    • Export Citation

  • Friedman, Milton, 1960, A Program for Monetary Stability (New York: Fordham University Press).

  • Genberg, Hans, and Ulrich Kohli, 1997, “Recent Developments in Swiss Monetary Policy,” in Economic Policy in Switzerland, ed. by Philippe Bacchetta and Walter Wasserfallen, Macmillan Press, Ltd.

    • Search Google Scholar
    • Export Citation

  • Heller, Daniel, 1997, “Zinskurven und ihr Informationsgehalt für die Geltpolitik der SNB,” Quartalsheft/Bulletin trimestriel, No. 2 (Zürich: Schweizerische Nationalbank).

    • Search Google Scholar
    • Export Citation

  • Hamilton, James D., 1989, “A New Approach to the Economic Analysis of Non-stationary Time Series and the Business Cycle,” Econometrica, No. 57, pp. 357–384.

    • Search Google Scholar
    • Export Citation

  • Laubach, Thomas, and Adam S. Posen, 1997, “Disciplined Discretion: The German and Swiss Monetary Targeting Frameworks in Operation,” Federal Reserve Bank of New York Working Paper (New York: Federal Reserve Bank of New York).

    • Search Google Scholar
    • Export Citation

  • Laxton, Douglas, and Esward Prasar, 1997, “Possible Effects of European Monetary Union: An Analysis Using MULIMOD Simulations,” IMF Staff Country Reports No. 97/18.

    • Search Google Scholar
    • Export Citation

  • McCallum, Bennett. T., 1987, “The Case for Rules in the Conduct of Monetary Policy: A Concrete Example,” Economic Review, Federal Reserve Bank of Richmond, September/October, pp. 10–18.

    • Search Google Scholar
    • Export Citation

  • McCallum, Bennett. T., 1989, Monetary Economics (New York: Macmillan).

  • Peytrignet, Michel, 1996, “Stabilité économétrique des agrégats monétaires suisses,” Quartalsheft/Bulletin trimestriel, No. 3, (Zürich: Schweizerische Nationalbank).

    • Search Google Scholar
    • Export Citation

  • Peytrignet, Michel, and Christof Stahel, 1998, “Stability of Money Demand in Switzerland: A Comparison of the M2 and the M3 cases,” Empirical Economics, No. 23.

    • Search Google Scholar
    • Export Citation

  • Persson, Torsten, and Guido Tabellini, 1990, Macroeconomic policy, Credibility and Politics, (Chur: Harwood Academic Publishers).

  • Rich, Georg, 1992, “Die schweizerische Teurung: Lehrnen für die Nationalbank,” Quartalsheft/Bulletin trimestriel, No. 1 (Zürich: Schweizerische Nationalbank).

    • Search Google Scholar
    • Export Citation

  • Rich, Georg, 1997, “Monetary Targets as a Policy Rule: Lessons from the Swiss Experience,” Journal of Monetary Economics, No. 39, pp. 113–141.

    • Search Google Scholar
    • Export Citation

  • Rich, Georg, and Jean-Pierre Béguelin, 1985, “Swiss Monetary policy in the 1970s and 1980s,” in K. Brunner and others, Monetary policy and Monetary regimes, Center Symposium Series, No. CS-17, Center for Research in Government policy and Business, University of Rochester.

    • Search Google Scholar
    • Export Citation

  • Sutherland, Alan, 1994, “Target Zone Models with Price Inertia: Solutions and Testable Implications,” The Economic Journal, No. 104, pp. 96–112.

    • Search Google Scholar
    • Export Citation

  • Svensson, Lars, 1997, “Inflation Forecast Targeting: Implementation and Monitoring Inflation Targets,” European Economic Review, No. 41, pp. 111–1146.

    • Search Google Scholar
    • Export Citation

  • Svensson, Lars, 1998a, “Monetary Policy Issues for the European System of Central Banks,” Carnegie-Rochester Conference on Public Policy.

    • Search Google Scholar
    • Export Citation

  • Svensson, Lars, 1998b, “Inflation Targeting as a Monetary policy Rule,” submitted to Journal of Monetary Economics.

  • Taylor, John B., 1993, “Discretion versus Policy Rules in Practice,” Carnegie-Rochester Conference on Public Policy, No. 39, pp. 195–214.

    • Search Google Scholar
    • Export Citation
18

Prepared by Ketil Hviding.

19

This section is to a large extent based on Rich (1985, 1997) and Genberg and Kohli (1997).

24

Current proposals to bring Switzerland’s “monetary constitution” in line with current practice would establish price stability as the primary objective.

Same Series

Other IMF Content

Other Publishers

Cover IMF Staff Country Reports
Switzerland: Selected Issues and Statistical Appendix
Author: International Monetary Fund
Volume/Issue:
Volume 1999: Issue 030
Publisher:
International Monetary Fund
ISBN:
9781451807165
ISSN:
1934-7685
Pages:
88
DOI:
https://doi.org/10.5089/9781451807165.002

Switzerland: Monetary Policy Performance in Industrial Countries, 1973-97

Switzerland: Exchange Rates, 1973-98

Switzerland: Monetary Aggregates

Switzerland: Interest Rates, 1973-98

Switzerland: Short-Term Interest Rates and the Simple Taylor Rule, 1974-98

(in percent)