This Selected Issues paper takes a close look at the proximate determinants and short-term dynamics of inflation in Switzerland. It identifies salient features of the inflation experience in Switzerland. The paper uses a wage-price model to gauge the sensitivity of inflation to the business cycle and the exchange rate. The findings suggest that underlying inflation of about 1 percent reflects a slightly higher rate for domestic goods and services and low increases in prices of imported goods.

Abstract

This Selected Issues paper takes a close look at the proximate determinants and short-term dynamics of inflation in Switzerland. It identifies salient features of the inflation experience in Switzerland. The paper uses a wage-price model to gauge the sensitivity of inflation to the business cycle and the exchange rate. The findings suggest that underlying inflation of about 1 percent reflects a slightly higher rate for domestic goods and services and low increases in prices of imported goods.

I. Low Inflation in Switzerland1

A. Introduction

1. Switzerland has always been a low inflation country by international standards, but in the past ten years it has come the closest of any industrial country to achieving price stability. Low inflation was institutionalized in 2000 with the adoption of a new monetary policy framework that specifically mandates the SNB to keep long-term inflation below two percent, but positive.2

2. This paper takes a close look at the proximate determinants and short-term dynamics of inflation in Switzerland. The main conclusions are:

  • Underlying inflation of about 1 percent reflects a slightly higher rate for domestic goods and services and very low increases in prices of imported goods.

  • The variability of inflation has also been low, with discreet factors—adjustments in rents, VAT and product market reforms—accounting for much of the variance of inflation within a fairly narrow 0-2 percent range.

  • Low inflation has not prevented substantial shifts in relative prices and in a typical month as many as one third of all prices may be declining.

  • In a low inflation environment there is significant potential for overall inflation to turn negative. To the extent that negative inflation reflected the fruits of productivity increases and product market liberalization, the economy would be presented with an opportunity. But if negative inflation were the result of a prolonged economic slump, policy makers could, as discussed in Chapter II, be faced with a serious challenge.

3. The chapter is structured as follows: Section B identifies salient features of the inflation experience in Switzerland; Section C uses a wage-price model to gauge the sensitivity of inflation to the business cycle and the exchange rate; and Section D concludes.

B. Inflation Close-Up

4. Inflation has been historically low in Switzerland. It has averaged 3.3 percent over the past forty years, the lowest (along with Germany) among OECD countries (Figure I-1). Over the much longer period 1923-2002, average inflation was only 2.3 percent.3 The several inflationary bouts that have occurred have been less pronounced than in other industrial countries as the SNB has been successful in preventing them from becoming entrenched into expectations. This success reflects a tight rein over money growth (Figure I-2).

Figure I-1.
Figure I-1.

Inflation in Selected OECD Countries

(Annual average rates of change in percent)

Citation: IMF Staff Country Reports 2003, 149; 10.5089/9781451807257.002.A001

Sources: ADB; OECD; and IMF staff calculations.
Figure I-2.
Figure I-2.

Inflation and Money Growth in OECD Countries, 1970-2002

Citation: IMF Staff Country Reports 2003, 149; 10.5089/9781451807257.002.A001

M3 growth minus real GDP growth

5. In the past decade, inflation has dropped to a low level by even Switzerland’s standards (Figure I-3). The monetary relaxation in the wake of the oil shocks in the 1970s and the overheating in the late 1980s gave rise to two inflationary surges, followed by rapid disinflations, that kept average consumer price inflation at 3½ percent in 1980-93. Subsequently, average inflation has declined to only 1 percent. Inflation volatility has also been low. Inflation has fluctuated in a 0-2 percent range and the standard deviation of 12-month inflation has been 0.6 percent or one third of its level in 1980-93.

Figure I-3.
Figure I-3.

Headline Inflation

(Annual rate of change, in percent)

Citation: IMF Staff Country Reports 2003, 149; 10.5089/9781451807257.002.A001

6. A picture of stable, low inflation in the past decade is even clearer using non-parametric measures of inflation and correcting for discrete changes (Figure I-4). Non-parametric measures (e.g. median and trimmed means) are robust to noise introduced by outliers that tend to distort the information content of headline inflation.4 Inflation volatility has been accentuated by rent adjustments linked to interest rate changes. In addition, increases in VAT rates, one-off price drops associated with the opening up of the telecommunications sector and agricultural policy reform, and gyrations in oil prices have added to inflation volatility. Adjusted for these factors, the standard deviation of inflation has been only 0.3 percent (Table I-1).

Figure I-4.
Figure I-4.

The Anatomy of Swiss Inflation, 1984-2002

Citation: IMF Staff Country Reports 2003, 149; 10.5089/9781451807257.002.A001

Sources: Federal Statistical Service and IMF staff calculations.
Table I-1.

Switzerland: Consumer Price Inflation, 1994-2002

article image
Sources: Federal Statistics Office; and Fund staff calculations.

VAT changes, telecommunications, and oil prices.

Core inflation I excludes from headline inflation food items, drinks, tobacco, seasonal products and energy.

Core inflation II excludes from Core 1 administered prices.

Relative price adjustments

7. The dispersion of inflation rates of individual items in the CPI has been quite stable. The interquantile range5 has been roughly 3 percentage points, i.e., 50 percent of the price changes have been ±1½ percentage points about median inflation regardless of the level of inflation (Figure I-4). This empirical regularity implies a high incidence of price declines in the current low inflation environment as sizable relative price declines can materialize only through absolute price declines.

8. Thus, low average inflation has not prevented considerable downward price adjustment. At any given month since 1994, at least one quarter of the items in the consumer basket (not always the same items) have registered price declines (middle panel in Figure I-4). Price declines were more frequent when the output gap was widest in 1997—almost half of the items in the CPI basket declined in price during that period. The largest declines were registered in telecommunications and computer hardware, where the cumulative drop reached -60 percent, food items (mainly dairy products), clothing and some services. Overall, absolute price declines shaved 0.35 percentage points off the average annual inflation rate. On the upside, above average inflation rates have been concentrated in services: at any one month since 1994 roughly one fifth of all items increased at annual rates of at least 2 percent.

9. An important relative price change in the CPI basket has been the decline of imported goods prices relative to domestic goods. Prices of imported items have been more volatile than domestic ones reflecting, principally, gyrations in the exchange rate and oil prices.6 They have also tended to increase on average less than prices of domestic goods and services due to the trend appreciation of the Swiss franc and, in recent years, dramatic drops in the prices of information and telecommunications technology products (Figure I-5). In 1994-2002, the average price increase of domestic goods and services was 1.1 percent per year compared to only 0.2 percent for imported goods. This inflation differential has resulted in a trend decline of import prices relative to the CPI (Figure I-6).

Figure I-5.
Figure I-5.

Domestic and Foreign Inflation

Citation: IMF Staff Country Reports 2003, 149; 10.5089/9781451807257.002.A001

Figure I-6.
Figure I-6.

Relative Price Development

Citation: IMF Staff Country Reports 2003, 149; 10.5089/9781451807257.002.A001

Domestic inflation

10. Low productivity growth in sheltered sectors, and the slow pace of liberalization, have underpinned above average domestic inflation. Large non-traded/traded goods inflation and productivity growth differentials are consistent with a Balassa-Samuels on effect. Two measures of the non-traded/traded goods price—the ratios of consumer/producer prices and of the domestic/foreign component of the CPI—exhibit a similar upward trend over the past twenty years. Notwithstanding some divergence in the wake of liberalization efforts in the late 1990s,7 both measures increased by a cumulative 43 percent or 1.6 percent annually (Figure I-7). There are also significant differentials in sectoral productivity growth: in manufacturing, which is most exposed to competitive pressures, labor productivity increased in 1990-98 by 3.8 percentage points per year faster than in the rest of the economy (Table I-2). However, the labor productivity differential tends to overstate the Balassa-Samuelson effect as it abstracts from the impact of capital deepening on labor productivity and variations in relative markups.8

Figure I-7.
Figure I-7.

Traded/Non-traded Goods Prices

Citation: IMF Staff Country Reports 2003, 149; 10.5089/9781451807257.002.A001

Table I-2.

Labor Productivity 1/

article image
Sources: Federal Statistical Office; and Fund staff estimates.

In full time employment equivalents.

11. The productivity differential appears to have increased temporarily in the early 1990s (Figure I-8). This reflects restructuring in the traded goods sector, mainly manufacturing and financial services (Gagales, 2002). According to the Balassa-Samuelson model, differences in sectoral productivity growth put pressure on wages in lagging sectors to increase faster than the rate consistent with productivity growth thus leading to faster price increases in those sectors and, if accommodated by monetary policy, higher headline inflation. With the productivity differential at 1½ percentage points and the sheltered sectors representing 40 percent of the Swiss economy, the Balassa-Samuelson effect could account for 0.6 percentage points of headline inflation.

Figure I-8.
Figure I-8.

Traded/Non-traded Goods Prices

Citation: IMF Staff Country Reports 2003, 149; 10.5089/9781451807257.002.A001

12. Cross-country price level convergence does not appear to have been at work. According to Eurostat estimates (BFS, 2002) the price level remains at roughly 35 percent above the EU average (20 percent above Germany) without signs of convergence as the lower rate of inflation relative to the EU has been offset by exchange rate appreciation (Figure I-9). The price differential is larger in services than in traded goods, where arbitrage opportunities are greater, especially in border regions. The differential reflects Switzerland’s higher standard of living (which boosts prices of non-traded goods and services) and market segmentation (which nurtures inefficiencies and monopoly profits) that more than offset the effect of low indirect taxes. It also reflects the slower pace of liberalization in Switzerland relative to its trading partners.9 In a low inflation environment, and if the trend nominal appreciation of the Swiss franc were to continue, any significant price level convergence would entail declines in non-traded goods’ prices and, most likely, also in headline inflation. However, to the extent that such declines were associated with faster productivity, they would represent a positive supply shock. Falling prices need not be a problem for monetary policy to the extent that higher real interest rates would be consistent with higher productivity growth, although there would remain the danger of generalized deflationary expectations (see Chapter II).

Figure I-9.
Figure I-9.

Relative Price Levels

Citation: IMF Staff Country Reports 2003, 149; 10.5089/9781451807257.002.A001

Sources: Eurostat; OECD; and IMF staff estimates.

Imported inflation

13. Given the openness of the Swiss economy, foreign inflation and the exchange rate are important determinants of inflation. Imports and exports of goods and non-factor services each amount to about 40 percent of GDP—compared to 29 percent in Germany and only 12 percent in the United States. Moreover, imported items carry a weight of 26.7 percent in the CPI basket (non-energy: 22.5 percent; energy: 4.2 percent).10 Foreign prices and the exchange rate filter into domestic prices directly via the distribution and production chains; and indirectly via second round effects on the cost of intermediate products, wages and inflationary expectations.

14. That said, the relationship between the prices of imported items in the CPI and the import deflator appeared to break down in the early 1990s (Figure I-10) A weak correlation between the two is not unique to Switzerland and is related to several factors: compositional differences between the two price indices (e.g., a portion of imports are used primarily as inputs for the production of exports and, thus, have no effect on domestic prices); the significance of domestic value added (distribution costs, import duties, taxes and profit margins) in the consumer price of imported items; monopolistic market structures;11 and the SNB’s low tolerance for inflation.12 However, the disconnect in the 1990s was particularly pronounced and persistent, including for energy products; it also occurred at the producer price level (Figures I-11 and I-12). The breakdown in the relationship between the prices of imported items in the CPI and the import deflator does not appear to reflect a change in the pricing strategy of foreign exporters to Switzerland as the pass-through of exchange rate changes to the import deflator remained fairly stable in the 1990s (Figure I-13.13 A possible explanation is that the prolonged period of economic slack and the enterprise restructuring that took place in Switzerland during that period had a persistent impact on domestic firms’ and distributors’ pricing policies.

Figure I-10.
Figure I-10.

Import Prices and CPI

Citation: IMF Staff Country Reports 2003, 149; 10.5089/9781451807257.002.A001

Figure I-11.
Figure I-11.

Import Prices and PPI

Citation: IMF Staff Country Reports 2003, 149; 10.5089/9781451807257.002.A001

Figure I-12.
Figure I-12.

Energy Prices

Citation: IMF Staff Country Reports 2003, 149; 10.5089/9781451807257.002.A001

Figure I-13.
Figure I-13.

Foreign Prices and NEER, 1980-2002

Citation: IMF Staff Country Reports 2003, 149; 10.5089/9781451807257.002.A001

Change in NEER

C. Proximate Determinants of Inflation

15. In this section a standard price-wage model is used to further analyze the inflation process in Switzerland and, in particular, the sensitivity of inflation to the business cycle and exchange rate gyrations. The approach abstracts from a broader explanation of inflation—which, as in other countries, can ultimately be traced to money growth. Rather, the emphasis is on the proximate determinants of inflation and the conditions under which inflation might be pushed outside the 0-2 percent range it has fluctuated in during the past decade.14

16. Econometric analysis (see Appendix) finds considerable support for a relatively stable long-run relationship between consumer prices, labor costs, imported prices, and the productivity differential between traded and non-traded goods sectors. In the short run, the output gap also has a significant impact on inflation. The main findings can be summarized as follows:

  • In the long-run, there is full pass-through of exchange rate and foreign price changes to the CPI. The coefficient of foreign prices in the long-run price equation, which reflects the direct pass-through, is similar to the weight of imported goods in the CPI (around 25 percent). But the system pass-through is unity because wage inflation sparked by price rises reinforces the initial price increase. Full pass-through implies that the trend nominal appreciation of the Swiss franc does not affect the long-term price level differential against trading partners.

  • However, the speed of pass-through is slow. It takes nearly 5 years before half of the full pass-through, including the effects induced by secondary wage changes, is complete—i.e. by the end of the fifth year, the price level would have declined by 5 percent, ceteris paribus, for a permanent 10 percent appreciation. Even so, this means that a 10 percent appreciation reduces the annual inflation rate by about 1 percentage point on average in the medium term.

  • The productivity growth differential between the traded and the non-traded goods sectors is estimated to have contributed 0.8 percentage points to annual inflation in 1994-2002. This is broadly in line with the earlier back-of-the-envelope calculation of the Balassa-Samuelson effect. In effect, the just under I percent inflation rate accommodated by monetary policy in the past decade was consistent with negligible traded goods inflation—both foreign and domestic.

  • The output gap has a significant impact on inflation dynamics. Its cyclical impact on markups is reinforced via Phillips curve effects on wages.

17. The estimates are reasonably stable, apart from the period 1991Q2-1992Q2 when there was a breakdown in the relationship between import prices and their domestic component. Of course, the model provides only a reduced-form representation of the inflation process and, as such, is potentially hostage to the Lucas (1976) critique. Two types of policy change could affect the stability of the model. First, to the extent that the economy-wide markup is affected by the opening-up of sheltered sectors (agriculture, construction and services) and the integration of the domestic market, the model could be less stable going forward if liberalization were stepped up. Second, changes in the monetary policy framework or the policy rules—e.g. a switch to quantitative monetary easing in the face of an interest rate floor—could affect expectations, price setting behavior, and the monetary policy transmission mechanism in general.

18. The model implies that the business cycle and exchange rate fluctuations account for a significant part of the variability of inflation. Moreover, it would not take major deviations of exchange rates or the output gap from their historical variability to push inflation outside of the 0-2 percent range—especially if the output gap and exchange rate movements reinforced each other given their negative correlation.15

  • Sustained appreciation could cut inflation significantly. The nominal effective exchange rate has been fluctuating considerably around its trend with a standard deviation of 3.4 percent (Figure I-14). The average phase of these fluctuations stretches over two years and the average peak deviation from trend is 5 percent (Figure I-15 and I-16). A spread-out appreciation of this order of magnitude would temporarily shave 0.3-0.4 percentage points from inflation with the maximum impact being felt after 5-6 quarters. A more abrupt 5 percent appreciation that lasts for one year has an even sharper and more immediate impact on inflation: at its peak, in the fourth quarter, it reduces inflation by 0.5 percentage points (Figure I-17).

  • Likewise, a persistent output gap can significantly reduce inflation. The recent business cycle experience has been characterized by abrupt and persistent shifts in real GDP growth (Jaeger, 1999). A moderate slowdown in economic activity would last about three years with output dropping 1½ percent below potential. The model predicts a strong effect of variations in the output gap on inflation. A gradual widening of the output gap over a period of four years to an average of 0.75 percent has its maximum impact at the end of the third year when inflation is 0.6 percentage points below trend. A step increase in the output gap by 1½ percentage points for a period of one year would have its maximum impact at the end of the first year (Figure I-18).

Figure I-14.
Figure I-14.

Exchange Rate Volatility

Citation: IMF Staff Country Reports 2003, 149; 10.5089/9781451807257.002.A001

Figure I-15.
Figure I-15.

Periods of NEER Weakness

Citation: IMF Staff Country Reports 2003, 149; 10.5089/9781451807257.002.A001

Figure I-16.
Figure I-16.

Periods of NEER Strength

Citation: IMF Staff Country Reports 2003, 149; 10.5089/9781451807257.002.A001

Figure I-17.
Figure I-17.

Impulse Response of Inflation to NEER Appreciation

Citation: IMF Staff Country Reports 2003, 149; 10.5089/9781451807257.002.A001

Figure I-18.
Figure I-18.

Impulse Response of Inflation Widening Output Gap

Citation: IMF Staff Country Reports 2003, 149; 10.5089/9781451807257.002.A001

19. In the current context, the pass-through of the earlier widening of the output gap and the exchange rate appreciation has probably not been completed, suggesting that inflation has not yet bottomed out. Assuming the recent easing of oil prices is sustained, the gradual elimination of the output gap over the next four years (baseline scenario in Figure I-19), and no further appreciation of the nominal effective exchange rate, the model predicts non-rent inflation to average ½percent in 2003 and 2004 before rising to 1¾ percent in 2006 (Figure I-20).

Figure I-19.
Figure I-19.

Output Gap Scenarios Figure

Citation: IMF Staff Country Reports 2003, 149; 10.5089/9781451807257.002.A001

Figure I-20.
Figure I-20.

Non-Rent Inflation Scenarios

Citation: IMF Staff Country Reports 2003, 149; 10.5089/9781451807257.002.A001

20. As the model projects inflation to bottom out close to zero, it would not take much to push inflation into negative territory. For example, a slower closing of the output gap, especially if accompanied by additional appreciation (1 percent annually) could easily push inflation down to -½ percent in 2004 (Scenario 1). However, the model predicts that negative inflation would not become ingrained and would eventually return to around 1-2 percent in the medium term. The upside risks appear small: a faster closing of the output gap would raise average inflation to ¾ percent in 2003-05 (Scenario 2).

21. While the simulations suggest a moderate risk that the overall price level could begin to fall in the period ahead, they abstract from a fuller equilibrium assessment of inflation prospects that takes into account recent policy actions. For example, the surge in money supply in early 2003 from the relaxation of monetary policy implies that the forces to close the output gap—or, if the gap is overstated, to put upward pressure on prices—are likely in the pipeline. The lags involved, however, are difficult to gauge. At the same time, negative inflation also contains some risks of engendering deflationary expectations that are not properly captured by the econometric estimates. In essence, the dynamics of deflation could be quite different from those predicted by the experience of the past 10-20 years of low but positive inflation.

D. Conclusion

22. Inflation in Switzerland has been close to zero for almost a decade now. Behind this impressive achievement has been a very low imported component of inflation, on average, with somewhat higher domestically-generated inflation. Close to zero means that many prices often fall. It also means that generalized price declines are never far away. This could come about as a result of liberalization of sheltered domestic sectors—the sharp falls in liberalized prices in the late 1990s are a good example—which will present an opportunity for policy makers. Macroeconomic factors could also conspire to drive inflation into negative territory. Given the openness of the economy, pass-through of exchange rate appreciation could pull down the CPI. Likewise, a prolonged slump could do the same. In this case, monetary policy might face an uphill task of countering negative inflation expectations given the zero floor for interest rates.

APPENDIX: A Price-Wage Model of Inflation

Price equation

23. Following de Brouwer and Ericsson (1998), the price level (P) is determined in the long-run by a trendless markup (k) over total unit costs, including unit labor costs (ULC), intermediate non-energy import prices (PM) and energy prices (PE).

P=(1+k)ULCaPMbPEc(1)

24. The markup may change over the business cycle and is expressed as a function of the output gap (GAP). Under linear homogeneity a+b+c=1 and error correction price dynamics, short-term inflation behavior is given by:

Δp=θ(p1kaulc1bpm1(1ab)pe1)+φgap+φ1Δgap+a1ulc+b1Δpm+c1Δpe+ε(2)

where lowercase letters denote logarithms Δ and stands for the first difference operator. The term in parenthesis denotes the deviation of the price level from its equilibrium level. Money growth, the main determinant of inflation in the longer run, is accounted for only implicitly via its effects on the exchange rate and the output gap; and on the parameters of the model.

25. Given the evidence in Section B that there is a significant productivity growth differential between the traded and non-traded goods sectors, the model is recast in a two-sector framework (traded and non-traded goods sectors are denoted respectively by T and N). Consumer prices, which have a large domestic component, are assumed to depend on unit labor costs in the non-traded goods sector:

p=k+aulcN+bpm+(1ab)pe(1)

In Switzerland, ulcN is not observed directly but can be inferred from the identity:

ulc=λulcT+(1λ)ulcN

where λ stands the share of the traded goods sector in total employment. Denoting by δ the productivity growth differential between the two sectors, ulcN can be expressed as:

ulcN=ulc+λσ

where σ stands for the cumulative productivity differential since the base period. Substituting the above expression in (1’) gives the augmented long-term price equation:

p=k+aulc+αλσ+bpm+(1ab)pe(1)

and the corresponding dynamic equation:

Δp=θ(p1kaulc1aλσ1bpm1(1ab)pe1)+φgap+φ1Δgap+a1Δulc+b1Δpm+c1Δpe+ε(2)

From equations (2) and (2’) it is evident that, if ulc and σ are positively correlated, the exclusion of the latter from (2) biases upwards the estimate of a because it forces it to pick up also the Balassa-Samuelson effect.

26. To estimate the long-run relationship, a system cointegration analysis was conducted for {p, ulc, pm, pe, σ}, the results of which are summarized in Table I-3. The tests support the existence of one cointegrating vector. The estimated coefficients are plausible (with the exception of the coefficient of oil prices, which appears with the wrong sign) and satisfy linear homogeneity. However, the estimate of the cointegrating vector is sensitive to the choice of the sample.

Table I-3.

Inflation Equation—Cointegration Analysis

article image
*(**) denotes rejection of the hypothesis at the 5%(1%) levelTrace test indicates 2 cointegrating equation(s) at the 5% levelTrace test indicates 1 cointegrating equation(s) at the 1% level
article image
*(**) denotes rejection of the hypothesis at the 5%(1%) levelMax-eigenvalue test indicates 1 cointegrating equation(s) at both 5% and 1% levels
article image

27. Equation (2") is estimated with quarterly data for the period 1981:1-2002:4, excluding observations for 1991:2-1992:2 when the relation between import and domestic prices appears to have broken down.16 The final form of the equation is given by (Table I-4:)

Table I-4.

Inflation Equation

article image
Δp=0.15(p1plev+1.510.75ulc10.19σ10.23pm10.025pe1)+1.26Δplev+0.17gap2+0.12Δgap+0.18Δulc1+0.05Δpm+0.01Δpe(3)

28. The inclusion of observations for 1991:2-1992:2 derails the coefficient of the productivity differential but does not affect materially the other coefficients. The equation fits the data equally well in the 1980s and the 1990s. Figure I-21), summarizes the results of recursive estimates. Although some coefficients shift around 2000, when the economy recovered from the protracted recession, these are well within the ±2 standard errors band suggesting that the model is reasonably stable.

Figure I-21.
Figure I-21.

Inflation Equation: Recursive Coefficients

Citation: IMF Staff Country Reports 2003, 149; 10.5089/9781451807257.002.A001

29. The main features of the estimates are:

  • Linear homogeneity is accepted by the data.

  • The speed at which prices adjust to deviations from their long-term level (given by the error correction term) is slow. With only 14 percent of a disequilibrium in the markup being corrected each quarter, the implied half-life is 5 quarters.17 This estimate, however, abstracts from feedback from secondary wage adjustments

  • Imported inflation is an important determinant of consumer inflation. The 23 percent long-term pass-through for non-energy imports and 2.5 percent for energy imports are similar to the weights of these two items in the CPI basket.18 This, however, represents only the direct effect on inflation. If indirect effect through wages are also taken into account, the linear homogeneity condition ensures that pass-though is full (see below).

  • The output gap has a large and statistically significant effect on inflation. Its change (acceleration principle) has an equally large but statistically insignificant effect.

  • Import prices and unit labor costs affect inflation mainly via the error correction term. Their direct short-term effect is quantitatively small, albeit statistically significant.

30. To check the robustness of the equation to errors in measuring the output gap and the possibility of non-linear effects, the equation was estimated also with an HP-based output gap and with a quadratic output gap term.19 The HP-based estimate yielded a slightly higher coefficient for the output gap (a reflection of the smaller amplitude of this series) without affecting materially the other coefficient estimates and the overall fit of the model. The quadratic term was invariably numerically small and statistically insignificant—similar to the conclusion in Valdivia (1999)—and its inclusion was inconsequential for the rest of the coefficient estimates.

Wage equation

31. Wages are assumed to adjust to deviations of the labor share from its trend level, as well as to cyclical conditions (captured by the output gap), trend productivity growth, inflation and the exchange rate. Again, an error correction formulation is used for the dynamics. The estimates (see Table I-5) can be summarized as:

Table I-5.

Wage Equation

article image
Δw=0.10(w1lphp1p1s1)+0.19gap3+1.85Δlphp+0.08(Δp1+Δp2)0.01Δx+0.43Δw1+ξ(4)

where w stands for the wage rate (lacking direct observation, the quarterly series is interpolated from annual figures), Iphp denotes (HP-based) trend labor productivity (high frequency productivity fluctuations do not appear to have a significant effect on wages), s denotes the trend labor share and x stands for the nominal effective exchange rate (IFS based). Unit labor costs, based on smoothed labor productivity, are given by w-lphp. Equation (4) provides conditional nominal wage forecasts for predetermined paths of inflation, the exchange rate, productivity growth and the output gap.

32. Equation (4) is estimated over the period 1988:1-2002:4 which is relatively homogeneous with respect to the cyclical flexibility of the labor supply.20 The overall fit is reasonable, although in 2000-02 the equation predicts a more agile (compared to the actual) response of wages to economic conditions (Table I-5). The coefficient of the error correction term suggests that wages adjust more slowly than prices to deviations from their steady-state path. As in the case of the inflation equation, the output gap exerts a strong and statistically significant impact wage growth. Using an HP-based estimate of the gap does not alter the properties of the equation and, as in the case of the inflation equation, the output gap does not appear to have significant non-linear effects. Recursive estimates do not suggest the presence of structural breaks (Figure I-22).

Figure I-22.
Figure I-22.

Wage Equation: Recursive Coefficients

Citation: IMF Staff Country Reports 2003, 149; 10.5089/9781451807257.002.A001

Underlying inflation

33. Underlying inflation is determined by the long-run price and wage equations, the terms in parentheses in equations (3) and (4), after substituting out ulc = w-lphp from the price equation:

p=k+a(wlphp)+bpm+(1ab)pe+aλσ
w=lphp+p+s

which, after solving for p taking first differences, gives:

Δp=α(1α)1Δs+b(1α)1Δpm+(1ab)(1α)1Δpe+a(1α)1λΔσ(5)

This equation incorporates both the direct and indirect effect (through wages) of imported inflation and the productivity differential on inflation.

  • Because of linear homogeneity in the markup equation, the trend exchange rate appreciation and foreign inflation pass through fully to domestic inflation.

  • The productivity growth differential between the traded and the non-traded goods sectors imparts a positive bias to overall inflation (if accommodated by monetary policy). For α=0.75, λ=0.19 and with the relative non-traded/traded goods price increasing at an annual average rate of 1.4 percent in 1994-2002, model predicts a Balassa-Samuelson effect of 0.8 percentage points; and for a foreign inflation of 0.2 percent during the same period, it predicts also an underlying inflation of 1 percent, which is very close to average inflation (0.9 percent,(Table I-1).

References

  • Aebersold, C. and A. Brunetti, 1998, “Die langfristige reale Aufwertung des Schweizer Frankens gegeniiber der D-MarkAussenwirtschaft, Vol. 53. No III., pp. 365386.

    • Search Google Scholar
    • Export Citation
  • Bundesamt für Statistik, 2002, “Internationale Preisvergleiche und Kaufparitäten”, Neuchâtel, Switzerland.

  • Choudhri, E. and D. Hakura, 2001, “Exchange Rate Pass-Through to Domestic Prices: Does the Inflationary Environment Matter?”, IMF Working Paper 01/194.

    • Search Google Scholar
    • Export Citation
  • De Brouwer, G. and N. Ericsson, 1998, “Modeling Inflation in Australia”, Journal of Business and Economic Statistics, October, pp. 433449.

    • Search Google Scholar
    • Export Citation
  • Faber, S. and A. Fischer, 2000, “Kerninflationsraten in der Schweiz”, SNB Quarterly Bulletin, December.

  • Gagales, A., 2002, “Growth in Switzerland: Can Better Performance be Sustained?”, IMF Working Paper 02/153.

  • Gil-Pareja, S, 2002, “Export Price Discrimination in Europe and Exchange Rates”, Review of International Economics, 10(2), pp. 299312.

    • Search Google Scholar
    • Export Citation
  • Jaeger, A, 1999, “Alternative Models of the Swiss Business Cycle”, Switzerland, Selected Issues and Statistical Appendix, SM/99/31.

  • Jordan, T. and M. Peytrignet, (2001) “Die Inflationsprognose der Schweizerischen Nationalbank”, SNB Quarterly Bulletin, June.

  • Jordan, T., P. Kugler, C. Lenz and M. Savioz (2002), “Inflationsprognosen mit vektorautoregressiven Modellen”, SNB Quarterly Bulletin, March.

    • Search Google Scholar
    • Export Citation
  • Krajnyak, K, 2001, “Monetary Policy Transmission and the New Policy Framework”, in IMF Country Report 1/75, Washington.

  • Laxton, D and E. Prasad, 1997, “Possible Effcets of EMU: an Analysis Using MULTIMOD Simulations” in “Switzerland: Selected Issues and Statistical Appendix”, IMF SM/97/23.

    • Search Google Scholar
    • Export Citation
  • Lucas, R., 1976, “Econometric Policy Evaluation: A Critique” in K. Brunner and A. Meltzer (eds.), “The Phillips Curve and Labor Markets”, Carnegie-Rochester Conference Series, Vol. 1, pp 1946.

    • Search Google Scholar
    • Export Citation
  • Siegenthaler, H, 1996, “Historische Statistik der Schweiz”, Chronos Verlag, Zürich.

  • Stalder, P., 2001, “Forecasting Swiss Inflation with a Structural Macromodel: the Role of Technical Progress and the ‘Mortgage Rate-Housing Rent’ Link”, in BIS Papers No. 3 “Empirical Studies of Structural Changes and Inflation” August.

    • Search Google Scholar
    • Export Citation
  • Stalder, P., 2002, “The Changing Role of Foreign Labor and Female participation: Impacts on wage-price dynamics and unemployment in Switzerland”, paper presented at the Central Bank Workshop on Macroeconomic Modeling at the FRB, Washington, D.C.

    • Search Google Scholar
    • Export Citation
  • Valdivia, V., 1998, “The Tradeoff Between Inflation and Real Activity”, IMF Country Report 98/43.

1

Prepared by Anastassios Gagales.

2

The new monetary policy framework is described in Krajnyak (2001) and Jordan and Peytrignet (2001).

3

Estimate based on spliced series from Siegenthaler (1996).

4

Faber, S. and A. Fischer (2000) find that (weighted) trimmed inflation in Switzerland is better predicted by lagged broad money than headline inflation. The correlation was found to be robust over a wide range (15 to 50 percent) of the trim parameter.

5

The interquantile range is defined as the range between the first and third quartile and contains 50 percent of individual price changes.

6

The standard deviation of imported inflation was 2.1 percent compared to 0.7 percent for domestic inflation (last column in Table I-1).

7

Liberalization has affected mainly telecommunication charges and the prices of certain dairy products, which dropped cumulatively in 1996-2002 by 40 percent and 10-20 percent, respectively.

8

In a two-sector neoclassical model with mobile labor across sectors, the change in the relative price of non-traded/traded goods price is related negatively to the total factor productivity growth differential in the two sectors and positively to the change in relative market power (measured by the relative markups).

9

See Gagales (2002). The slow pace of liberalization can also account for the finding in Aebersold and Brunetti (1998), who estimate that the Balassa-Samuelson effect accounts for ½ of the price level differential between Switzerland and Germany.

10

With GNFS imports at 40 percent of GDP and inputs estimated in the national accounts at ½ of gross output, the weight of imported items in the consumer basket is at par with the share of imports in gross output.

11

The new open economy macroeconomics literature tries to account for this disconnect with imperfect competition and nominal rigidities (e.g., menu costs). Two complementary modeling strategies have emerged: in the first, imported goods are assumed to go through a distribution channel that adds significant local value added; in the second, all imports are treated as intermediate goods, often mixed with domestic goods to produce final consumer goods. Depending on market structure, exporters price their products either in their own or in local currency. Using an optimizing framework, Bacchetta and Wincoop (2002) show that when there is competition in the domestic final goods market, exporters tend to price final goods in domestic currency, in which case the pass-through to consumer prices is zero even when the pass-through to import prices is complete. There is substantial empirical evidence that pricing-to-market behavior is pervasive and tends to be pronounced among European exporters, which represent Switzerland’s main trading partners (Gil-Pareja, 2002).

12

In low inflation environments and in economies with a credible inflation targeting regime, pass-through is smaller as monetary policy generally offsets the effect of imported inflation on inflation to keep inflation within target (Choudhri and Hakura, 2001).

13

The correlation was -0.69 in 1980:2-1990:4 and -0.62 in 1991:1-2002:4. Increases in the nominal effective exchange rate denote appreciation.

14

See Jordan et. al. (2002) and Krajnyak (2001) for VAR models of inflation in Switzerland that account explicitly for the effect of monetary growth; their general conclusion is that broad money aggregates affect inflation with a 2-3 year lag. Other recent attempts to model inflation in Switzerland include Valdivia (1998), who uses a non-linear Phillips curve model; Laxton and Prasad (1997), who use MULTIMOD to examine the responsiveness of inflation to variations in money growth; and Stalder (2001), who estimates a medium-size structural model in which inflation is driven by interest rates and the exchange rate.

15

The correlation between deviations of the NEER from its trend and the output gap was -0.27 in 1980-2002.

16

Rents are excluded from the CPI as their adjustment is governed by longer-term contracts and regulations, including indexation of rents to the mortgage rate. The unit labor cost has been interpolated from annual data and the productivity differential is proxied by the ratio of consumer over producer prices; non-oil import prices are proxied by the weighed average of trading partners’ non-oil export deflators. The cost of energy is proxied by the Swiss franc equivalent of the spot price of oil. The output gap is interpolated from annual estimates based on the production function approach. The estimated equation includes also plev—which controls for changes in VAT rates, the liberalization of telecommunications and retail trade—and seasonal dummies.

17

The half-life is given by ln(0.5)/ln(l-θ).

18

The weight of imported items in the CPI basket has hovered around 25 percent since 1992, down from almost 30 percent in the 1980s.

19

Sources of measurement error in the production function based estimate of the output gap include the capital stock estimates (due to the lack of direct measurements), incomplete data on capacity utilization and problems in estimating capacity utilization in the services sector, and difficulty in measuring the NAIRU in an environment where a flexible foreign labor supply acts as a shock absorber in the labor market. On the other hand, the HP based estimate of the output gap, being mechanical, may fail to accurately capture excess demand pressures.

20

Stalder (2002) finds evidence that cyclical flexibility of the labor supply declined in the 1990s—as foreign labor supply and the participation rate became less sensitive to labor market conditions—thus making wages more responsive to labor market conditions than in the past. The bilateral agreements with the EU are expected to increase the cyclical flexibility of labor supply.

APPENDIX: The Theory of the Liquidity Trap

65. The IS-LM framework is sufficiently rich to discuss the liquidity trap and the related issues of negative real interest rates and deflation. Despite being a staple of introductory macroeconomic courses and sometimes considered overly ad-hoc, Krugman (1999a) demonstrates that first-principle underpinnings for the IS-LM framework can readily be provided: the negative association between income and real interest rates traced out by the IS-curve can be viewed as the result of intertemporal maximization on the part of consumers; the positive association between income and nominal interest rates traced out by the LM-curve can be viewed as the result of a cash-in-advance constraint, including the liquidity trap section where interest rates on bonds are so low that agents hold money even if they do not need it for transaction purposes.

66. Sufficiently large negative demand shocks can put the economy into liquidity-trap territory. Figure II-7 shows the usual IS-LM diagram. Both curves are draw as a function of real interest rates (r). Changes in expected inflation (πe) thus do not shift the IS-curve, as investment and savings decisions depend on real rates rather than nominal interest rates. However, they do shift the LM-curve as money demand is sensitive to nominal rather than real interest rates. Negative demand shocks shift the IS-curve inward, e.g., as the result of the desire to rebuild savings following an unexpected fall of asset prices, depressed investment demand due to past overinvestment, or a deteriorating growth outlook due to demographic aging. Following such a shock the curves cross at less than full-employment output (Y*). If sufficiently large they intersect in the flat part of the LM-curve, as indicated in (Figure II-7). Interest rates are then so close to zero that agents are indifferent between holding money or bonds; monetary injections in the form of open-market operations are simply absorbed through a reallocation of agents’ portfolios toward money at the expense of bonds without further consequences. Empirically, one would expect to see in such a situation an expansion of narrow and broader monetary aggregates, without commensurate reductions in interest rates.

Figure II-7.
Figure II-7.

IS-LM Framework

Citation: IMF Staff Country Reports 2003, 149; 10.5089/9781451807257.002.A001

67. If the economy is in a liquidity trap, full-employment real interest rates are typically negative. As is immediately clear from (Figure II-7), for investment to match savings at full-employment output real interest rates have to be negative if the IS-curve crosses through the flat section of the LM-curve. In other words, expected inflation has to exceed nominal interest rates if real interest rates are to bring investment and savings back into balance and restore full-employment output following a disruption to demand that has put the economy in a liquidity-trap situation. Graphically, πe has to increase sufficiently to shift the LM-curve down enough to intersect the IS’-curve at Y*.

68. The need for negative real interest rates is a less awkward proposition than it might seem. One might be tempted to argue that investment projects with a positive yield are always around and that therefore zero real interest rates are always sufficient to stimulate demand back to full-employment output. But this view is at odds with the experience in Japan and during the great depression and it can also be refuted on theoretical grounds: with spare capacity and/or with asset prices expected to decline, investment projects with a zero rate of return might not be available after all.

69. Instances where the full-employment real rate remains positive despite the economy being in a liquidity trap are possible but unlikely to be very relevant. If agents expect sufficiently large deflation, the economy could end up in a liquidity trap although real interest rates at full employment remain positive. In terms of Figure II-7, this could happen if the IS-curve shifted inward only by a little but deflation expectations shifted the LM-curve up enough so that the intersection occurs in its flat part. This is the scenario that Krugman (1999b) has in mind when arguing that there might be a window of opportunity for monetary policy to prevent a deflationary spiral: the initial small demand shock gives rise to an output gap but does not put the economy in a liquidity trap. At this point, expansionary policy could still stimulate the economy back to full employment. However, if this opportunity is not seized, the persistent output gap might give rise to deflation, shifting up the LM curve and rendering monetary policy impotent.

70. Economies with low equilibrium real interest rates and inflation on the one hand and large exposure to demand shocks on the other hand are particularly prone to liquidity traps. As is clear from (Figure II-7, even relatively small disruptions can put the economy in liquidity-trap territory if the full-employment real interest rate and expected inflation are low. Assessing a country’s susceptibility for liquidity traps therefore is much about evaluating the size of demand shocks relative to the typical level of real interest rates and inflation. Countries are at risk, for instance, if they exhibit large exposure to volatile external and financial markets combined with a track record of very low inflation and a high propensity to save that keeps real rates down.

71. Liquidity traps provide a fertile breeding ground for deflation and deflation tends to aggravate liquidity traps. While a liquidity trap does not necessarily set off deflation, it does involve insufficient demand so that inflation is lower than it typically is. This could spell deflation in a country where inflation is typically low to begin with and demand insufficiency is pronounced. Vice versa, deflation does not necessarily put the economy in a liquidity trap, but deflation tends to lower the expected rate of inflation thus making a liquidity trap more likely.

72. Deflation is a matter of concern because it can set in motion a contractionary spiral. This is widely believed to have happened during the great depression in the 1930s and, in a milder form, in today’s Japan. Both episodes marked periods of pronounced economic underperformance.27 Once deflation gets entrenched into expectations, real interest rates rise depressing private consumption and investment demand. Moreover, deflation increases the real burden of debt of companies and households thus deteriorating their balance sheets. This curtails their access to financing and their willingness to invest or consume; it could even push them into bankruptcy thus weakening the financial sector and its ability to lend. Both effects widen the output gap further and intensify the downward pressure on prices.

73. Ironically, deflation arises as a result of downwardly sticky prices. As Krugman (1997) points out, in a world of perfect price flexibility, following a negative demand shock prices would fall instantaneously by enough to generate the inflation expectation that is needed to restore full-employment output today. However, with sticky prices this does not happen fast and radically enough. An output gap emerges which puts downward pressure on prices. Prices begin to decline slowly. Rather than generating the needed inflationary expectations this gives rise to deflationary expectations.

74. Against this background, it is not clear whether wage flexibility increases or decreases the risk of deflation. On the one hand, perfect price and wage flexibility would eliminate the possibility of a contractionary deflation spiral to begin with. It appears reasonable to conjecture that almost the same can be achieved with nearly perfectly flexible prices and wages. On the other hand, it is often argued that downwardly rigid wages provide a useful safeguard against falling prices. With nominal wages unable to decline, firms’ ability to lower prices is narrowly circumscribed. This dampens deflation and deflationary expectations. Moreover, in a setting where output is constrained by demand the usual benefits of wage declines do not apply.

References

  • Ahearne, Alan, Joseph Gagnon, Jane Haltmaier, and Steve Kamin, 2002, “Preventing Deflation: Lessons from Japan’s Experience in the 1990s,International Finance Discussion Papers, Number 729, Board of Governors of the Federal Reserve System.

    • Search Google Scholar
    • Export Citation
  • Barro, Robert, 1995, “Inflation and Economic Growth,Bank of England Bulletin 35 (May), pp. 111.

  • Bernanke, Ben, 2002, “Deflation: Making Sure “It” Doesn’t Happen Here,” Remarks before the National Economists Club, Washington, DC, November 21, 2002, http://www.federalreserve.gov/boarddocs/speeches/2002/20021121/default.htm.

    • Search Google Scholar
    • Export Citation
  • Buiter, Willem, 2003, “Deflation: Prevention and Cure,NBER Working Paper No. W9623.

  • BIS, 1999, “Monetary Policy in the Advanced Industrial Countries,69th Annual Report, Bank for International Settlements, Basel, http://www.bis.org/publ/ar99e4.pdf.

    • Search Google Scholar
    • Export Citation
  • Feldstein, Martin, 1997, “The Costs and Benefits of Going from Low Inflation to Price Stability,” in: Reducing Inflation: Motivation and Strategy, ed. Romer and Romer, pp. 12356.

    • Search Google Scholar
    • Export Citation
  • Feldstein, Martin,, 1999, “The Costs and Benefits of Price Stability,University of Chicago Press for NBER.

  • Fischer, Stanley, 1993, “The Role of Macroeconomic Factors in Growth,Journal of Monetary Economics, 32, pp.485512.

  • Gehrig, Bruno, 2003, “Geforderte Geldpolitik - Überlegungen zum Konzept und zur aktuellen Entwicklung,Gastvorlesung an der Universität St. Gallen, http://www.snb.ch/d/publikationeii/publi.html?file=referate/ref 030124 bge.html.

    • Search Google Scholar
    • Export Citation
  • Jordan, Thomas and Michel Peytrignet, 2001, “Die Inflationsprognose der Schweizerischen Nationalbank,” Quartalsheft 2/2001, pp. 5461, Schweizerische Nationalbank, Zürich, http//www.snb.ch/d/publikationen/publi.html?file=geld/text geld.html.

    • Search Google Scholar
    • Export Citation
  • Jordan, Thomas, Peter Kugler, Carlos Lenz, and Marcel Savioz, 2002, “Inflationsprognosen mit vektorautoregressiven Modellen,” Quartalsheft 1/2002, pp. 4066, Schweizerische Nationalbank, Zurich, http://www.snb.ch,/d/publikationen/publi.html?file=geid/text geld.html.

    • Search Google Scholar
    • Export Citation
  • IMF, 1999, “Safeguarding Macroeconomic Stability at Low Inflation,World Economic Outlook, October 1999, International Monetary Fund, Washington, D.C., http://www.imf.org/external/pubs/ft/weo/1999/02/1099ch4.pdf.

    • Search Google Scholar
    • Export Citation
  • IMF, 2001, “Switzerland—Selected Issues,Country Report No. 01/75, International Monetary Fund, Washington, D.C., http://www.imf.org/external/pubs/ft/scr/2Q01/cr0175.pdf.

    • Search Google Scholar
    • Export Citation
  • IMF, 2002a, “Monetary and Exchange Rate Policies of the Euro Area—Selected Issues,Country Report No. 02/236, International Monetary Fund, Washington, D.C., http://www.imf.org/external/pubs/ft/scr/2002/cr02236.pdf.

    • Search Google Scholar
    • Export Citation
  • IMF, 2002b, “Switzerland—Selected Issues,Country Report No. 02/107, International Monetary Fund, Washington, D.C., http://www.imf.org/external/pubs//ft/scr/2002/cr02107.pdf.

    • Search Google Scholar
    • Export Citation
  • IMF, 2003, “Deflation: Determinants, Risks, and Policy Response”, International Monetary Fund, Washington, D.C., forthcoming.

  • Kugler, Peter and Georg Rich, 2001, “Monetary Policy Under Low Interest Rates: The Experience of Switzerland in the Late 1970s,” presented at the 32nd meeting of the “Ausschuss fur Geldtheorie und Geldpolitik - Verein fur Socialpolitik,” Frankfurt am Main, February 2001.

    • Search Google Scholar
    • Export Citation
  • Krugman, Paul, 1998, “It’s Baaack: Japan’s Slump and the Return of the Liquidity Trap,” Brookings Papers on Economic Activity, 2:1998, pp.137205.

    • Search Google Scholar
    • Export Citation
  • Krugman, Paul, , 1999a, “Thinking About the Liquidity Trap,” http://web.mit.edu/lmigman/www/trioshrt.html.

  • Krugman, Paul, , 1999b, “Deflationary Spirals,” http://web.mit.edu/krugman/www/spiral.html.

  • Laxton, Douglas and Eswar Prasad, 1997, “Possible Effects of European Monetary Union on Switzerland: A Case Study of Policy Dilemmas Caused by Low Inflation and the Nominal Interest Rate Floor,IMF Working Paper WP/97/23, International Monetary Fund, Washington, D.C., http://www.imf.org/external/pubs/ft/wp/wp9723.pdf.

    • Search Google Scholar
    • Export Citation
  • Mishkin, Frederic, 2001, “Issues in Inflation Targeting,” in: Price Stability and the Long-Run Target for Monetary Policy, Bank of Canada, pp. 203222. http://www.bankofcanada.ca/publications/working.papers/2000/mishkin.pdf.

    • Search Google Scholar
    • Export Citation
  • Summers, Lawrence, 1991, “How Should Long-term Monetary Policy Be Determined?,Journal of Money, Credit, and Banking, Vol 23, No. 3, pp. 625631.

    • Search Google Scholar
    • Export Citation
  • Svensson, Lars, 2000, “The Zero Bound in an Open Economy: A Foolproof Way of Escaping From a Liquidity Trap,NBER Working Paper, no. 7957, National Bureau of Economic Research, Cambridge, MA.

    • Search Google Scholar
    • Export Citation
  • Sarel, Michael, 1996, “Nonlinear Effects of Inflation on Economic Growth,IMF Staff Papers 43, pp. 199215.

  • Taylor, John B., 1993, “Discretion versus Policy Rules in Practice,Carnegie-Rochester Conference on Public Policy, No. 29, pp. 195214.

    • Search Google Scholar
    • Export Citation
21

Prepared by Christoph Klingen.

22

This section draws on IMF (2002a), Chapter I.

23

See chapter I.

24

Econometric analysis by IMF staff largely confirms this rule of thumb (IMF, 2001).

25

A recent cross-country analysis by IMF staff also puts the risk of deflation in Switzerland as moderate (IMF, 2003).

26

In 1981, the year after oil prices reached their peak some 175 percent above levels in 1978, Swiss inflation hit 6½ percent, about the same as in Germany but still less than in the United States.

27

Historically, there have also been episodes where economies grew at a healthy pace despite deflation. In the late 19th century, new technologies increased competition and expanded production capacities which stimulated growth and reduced prices (IMF, 1999 and BIS, 1999).

Switzerland: Selected Issues
Author: International Monetary Fund