Euro Area Policies: Selected Issues

This Selected Issues paper for euro area policies analyzes the product market regulation and benefits of wage moderation. The paper identifies structural shifts in the relationship between wages and unemployment rates—a “wage curve”—in 20 industrial countries. It reviews euro area and cross-country developments in labor costs and their bivariate relationship with unemployment rates and business GDP. The paper also examines aspects of the European Central Bank’s monetary analysis, within the context of their overall two-pillar policy framework, and issues surrounding its use.

Abstract

This Selected Issues paper for euro area policies analyzes the product market regulation and benefits of wage moderation. The paper identifies structural shifts in the relationship between wages and unemployment rates—a “wage curve”—in 20 industrial countries. It reviews euro area and cross-country developments in labor costs and their bivariate relationship with unemployment rates and business GDP. The paper also examines aspects of the European Central Bank’s monetary analysis, within the context of their overall two-pillar policy framework, and issues surrounding its use.

III. House Prices and Monetary Policy in the Euro Area 61

A. Introduction

59. The ECB, as part of its monetary analysis, has stressed the risks of an accommodative monetary policy associated with asset (particularly housing) market developments. Concerns have centered around high credit growth feeding through to high house prices, which in turn has implications for price stability. These issues have loomed large in recent ECB communications. In 2005, the January and February monthly bulletins warned of “unsustainable price increases in property markets”. In its March report, the ECB stated that “demand for loans for house purchases continues to be robust, contributing to strong house price dynamics in some regions of the euro area.” The April bulletin noted that “...strong monetary and credit growth indicates the need to carefully monitor whether risks are building up in the context of strong house price increases in some regions of the euro area.” While the May and June editorials did not mention house prices, they did point to continued upside risks to price stability from strong money and credit growth. So, at a time when the economic analysis confirms little in the way of underlying inflationary pressure, the evidence from the monetary analysis cross-check is dampening this appraisal.

60. There are a number of ways house prices can affect real activity and inflation. For a start, higher house prices could stimulate consumption through a wealth effect. Alternatively, a rise in housing prices could raise the ability of households to borrow when there are imperfections in the credit market, by raising the value of collateral (Bernanke, Gertler, and Gilchrist, 2000). Such a “financial accelerator” model posits a feedback mechanism between credit and housing prices—higher credit leading to higher consumption of goods and services, but possibly also to higher asset prices. However, many argue that the true cost of an unfettered increase in housing prices is not so much the direct inflationary impact of the boom, but the potential detrimental effect of the ensuing bust phase. Asset price busts, accompanied by financial instability and a collateral-induced credit crunch, can be extremely costly in terms of output.

61. Taking a long-term perspective, the pattern of real housing prices differed markedly across countries (Figure III.1). Over the period 1970-2003, real house prices barely budged in Germany, rose modestly in countries like France and Italy, and expanded considerably in places like Spain, the Netherlands, and Ireland (Table III.1). As can be seen from Figure III.1, the sample can be basically divided into two camps: high and low house price growth. The “low growth” cadre comprises the three large countries (plus Finland), while the other four (Ireland, Spain, Belgium, the Netherlands) recorded much faster growth over the long-term.62 One caveat should be noted upfront, however: data on house prices are imperfect at best, suffering from different methodological problems across countries.

Figure III.1.
Figure III.1.

Real Housing Prices in Selected Euro Area Countries

Citation: IMF Staff Country Reports 2005, 266; 10.5089/9781451813029.002.A003

Sources: BIS calculations based on national data.
Table III.1.

Selected Euro Area Countries: Real House Price Growth, 1970-2003

article image
Source: BIS calculations based on national data.

62. Divergences have persisted under economic and monetary union. Table III.2 shows the cumulative growth of credit, real housing prices, and goods prices across eight euro-area countries from 1998–2003.63 As can be observed, while inflation has been muted in most countries, real houses prices have expanded at a much faster clip, except in Germany where they stagnated. Similarly, the growth in credit has been robust, especially in countries like Ireland and Spain. Casual inspection supports a link between high credit and high house price growth countries. Of particular note is the cross-country variability; the standard deviation of real house price growth across countries was almost four times that of inflation, and credit growth was even more variable still. Thus low and stable inflation across countries co-existed alongside very different housing market developments.

Table III.2.

Selected Euro Area Countries: Credit, House Prices, and Inflation

(Cumulative change, 1998-2003)

article image
Source: Eurostat, BIS calculations based on national data.

63. This chapter will explore the inter-relationship between credit, house prices, and inflation in the euro area. A basic theme is that the transmission mechanisms between asset prices and economic activity, and between monetary policy and asset prices, are complicated. In this vein, Section B will argue that the relationship between credit and money growth and house prices depends on a variety of country-specific institutional characteristics. Such heterogeneity in experiences across the area complicates the task of a monetary policymaker attempting to extract clear signals from asset prices. Following this, Section C will show that the link between house price and goods price inflation is also not straightforward. In particular, house prices do not appear to help forecast consumer prices over the short- to medium run. Moreover, there is a tension between the potential inflationary consequences of the boom and the far more serious deflationary consequences of the bust. Finally, Section D will argue that these concerns mean that operationalizing monetary policy to address explicit asset price concerns in the euro area is beset with difficulty. Given this, other policy instruments, especially at the national level, may be more suited to tackling emerging asset price booms. Section E concludes.

B. Credit and House Prices

64. There is a broad literature on the economic determinants of real house prices. Appealing to this literature, European Central Bank (2003) derives a comprehensive list of factors with the potential to affect house price dynamics, including: household income; real interest rates; household formation and other demographic variables; supply side variables; financial market institutions and credit availability; and taxes, subsidies, and other public policies directed toward housing. Income is a key variable, while the effect of interest rates has not been as clearly established, although most results show a negative relationship. After income, the main long-run determinant of house prices is household formation. Other researchers reach similar conclusions (e.g. Borio and McGuire, 2004). Schnure (2005) shows that income, unemployment, and interest rates affect housing prices in the United States. Tsatsaronis and Zhu (2004) point to inflation. Others argue that equity prices play a role (Sutton, 2002).

65. The relationship between house prices and credit and money is not always easy to evaluate. European Central Bank (2003) notes that the relationship between the change in mortgage debt-to-GDP and house prices is not straightforward. Causality is hard to pin down, as rising mortgage debt may be the result of high prices, not the cause, while any co-movement could reflect a common response to third factors such as interest rates or expected future income growth. But some studies do find clear evidence of a role for monetary variables. Giuliodori (2004), for example, shows that house prices are affected by monetary shocks. Borio and McGuire (2004) argue that monetary policy matters when it comes to the emergence of sequential equity and housing price booms; housing booms tend to lag equity booms, with the lag length depending on interest rates. Moreover, housing price peaks are influenced partly by financial imbalances.

66. A baseline model is estimated to analyze the short- to medium-run dynamics in real house prices. The following equation is fitted to the data:

Δhi,t=αi+β1Δhi,t1+β2Δhi,t2+β3Δdi,t1+β4Δdi,t2+β5Δri,t1+β6Δri,t2+β7Δci,t1+β8Δci,t2+ϵi,t(1)

where i denotes a country, t is a time subscript, and, Δ represents the first difference operator. In terms of the variables, h is the log of real house prices, d is the log of real disposable income per capita, r is the real long-term interest rate, and c is the log of real credit. Separately, this equation is estimated replacing c with m, the log of real broad money.64 The αi component represents a country fixed effect. Therefore, the baseline is a panel regression for the eight countries in the sample, estimated using fixed effects (LSDV). But a more simple equation using pooled OLS was also estimated. Moreover, separate regressions were also run for each country, although the relatively short credit series means that these results must be interpreted with a great deal of caution.65 The results are instructive (Table III.3). Not surprisingly, the lagged dependent variable has the most explanatory power, and is significant in most countries, with the notable exception of Germany. In line with other studies, real income per capita is not a major determinant of short-run housing price dynamics in the panel, and is significant only in some countries (Germany, Ireland, Finland). The coefficient on the real interest rate has the expected negative sign, and is significant in the panel. Some have argued that nominal interest rates are also important determinants of house price dynamics. Robustness checks (unreported) show that replacing real with nominal interest rates in the baseline equation yields a statistically significant negative coefficient, but that this result no longer holds when including both variables together (real interest rates remain significant). Short-term interest rates are also significant in some specifications. But the real long-term interest rate is the most dominant variable in this class, with a statistically significant coefficient in every specification.

Table III.3.

Real House Price Equations

(Variables in log differences 1/)

article image
Sources: BIS calculations based on national data; Eurostat; WEO; OECD.

Real interest rate in differences

The credit series in Belgium was too short for meaningful analysis.

= t-statistic significant at 1 percent level;

= t-statistic significant at 5 percent level;

= t-statistic significant at 10 percent level.

67. The econometric results show that credit and money help predict real house prices only in some countries. The coefficients on the real credit growth in the panel regressions are not significant. In the individual country equations, credit seems to matter only in France, Ireland, and Spain. If real credit is replaced by real money, then the results show a significant money coefficient in France, Ireland, Belgium, Finland, and Spain, but not for the panel. One tentative conclusion, therefore, is that credit and money variables have no clear predictive power in explaining short- to medium-term real house prices across the euro area as a whole. The relationship depends on country-specific circumstances.

68. These results are robust to different panel specifications. The use of fixed effects in a dynamic panel equation can be criticized, given the noted bias. But Judson and Owen (1999) show that, based on Monte-Carlo experiments, when the time series is long enough relative to the cross-section dimension, the bias inherent in dynamic panel estimation is not large enough to make alternative estimators more desirable. Indeed, they find that the LSDV estimator performs better than alternatives with 30 or more years of data. Others have argued that when the time span covered by the data is reasonably large (around 22), then the application of IV-type estimators to a first differenced version of the dynamic panel model does not seem necessary, and can even lead to a large loss of efficiency (see Haque, Pesaran, and Sharma, 1999). Nonetheless, to check robustness, the model was also estimated using the Arellano-Bond dynamic panel technique. The results are similar, except that the coefficient on the real credit (but not the real money) variable is now marginally significant.

69. Aside from short- to medium-run dynamics, the long-run determinants of real house prices can also be modeled. The following long-run equation between real house prices and the previous explanatory variables is estimated:

hi,t=αi+γ1di,t+γ2ri,t+γ3ci,t+μi,t(2)

Following this, an Engle-Granger two-step version of equation (1) is then estimated, with the lagged residuals from equation (2) acting as the error-correction variable. Results are shown in Table III.4, both for the baseline model with real credit, and for the real money specification. In the long-run levels specifications, the coefficients of the three key independent variables—real disposable income per capita, real long-term interest rates, and real credit and/or money—are all statistically significant with the expected signs.66 Thus while there is scant evidence that real income, credit, or money matter for short- to medium-run dynamics, they are important determinants of long-run trends. The conclusions relating to the short- to medium-run dynamics do not change with the introduction of a (statistically significant) error correction component. In particular, while helping drive long-term trends, credit and money aggregates appear not to affect short- to medium-run dynamics.

Table III.4.

Real House Prices: Error Correction Model

(Variables in log differences) 1/

article image
Sources: BIS calculations based on national data; Eurostat; WEO; OECD.

Real interest rate in differences

= t-statistic significant at 1 percent level;

= t-statistic significant at 5 percent level;

= t-statistic significant at 10 percent level.

70. The literature shows that the effect of monetary policy and conditions on house prices depends largely on institutional factors. Differences in house price volatility across the area depend to some extent on institutional differences in credit markets between countries. In this context, a number of recent studies have analyzed the extent to which these factors affect house price volatility and the transmission mechanism to consumption. Maclennan, Muellbauer, and Stephens (1999) show that countries with fixed interest mortgage rates, low loan-to-value ratios, high transactions costs, and a smaller owner-occupied sector tended to experience lower house price volatility and smaller consumption effects. France and Germany fit neatly into this category, with Ireland and the United Kingdom at the opposite end of the spectrum. Giuliodori (2004) argues house prices enhance the effect of monetary policy on consumption when mortgage markets are more competitive. In the same vein, Iacoviello and Minetti (2003a) argue that the credit or collateral channel itself depends on these kinds of institutional factors. Tsatsaronis and Zhu (2004) also make a similar point, showing that the impact of credit on housing prices is more muted in countries where lending is conservative and equity withdrawal is rare.

71. Countries differ across a number of institutional mortgage market characteristics. Four aspects of mortgage markets are considered:67

  • Mortgage interest rates are variable instead of fixed (Finland, Ireland, Spain). Variable rates are likely to make house prices more sensitive to short-term interest rates and hence monetary policy.

  • Equity withdrawal is used (Finland, Ireland, the Netherlands). If households can withdraw home equity to take advantage of low refinancing rates and increased house values, then the credit channel of monetary policy could be enhanced, with knock-on effects for both consumption and house prices.

  • Mortgage assets are securitized (Ireland, the Netherlands, Spain). If credit institutions can sell excess exposure in the secondary market, this could lead to lower transactions fees and more flexible mortgage contracts, again bolstering the credit channel.

  • The maximum loan-to-value (LTV) ratio exceeds 80 percent (France, Ireland, Belgium, Spain). Prudential ceilings determine how conservative mortgage lending is, which affects the strength of the credit channel.

72. The evidence is consistent with a role for institutional factors in explaining the relationship between credit and house prices. Table III.5 reports coefficients from the variables of interest, when the panel regression is restricted to countries with certain characteristics. In these various sub-samples, there is a clear relationship between real credit and/or real money and real house prices. The results are borne out in each panel specification—LSDV, pooled OLS, and Arellano-Bond—and are especially strong in Arellano-Bond. In particular, the short- to medium-run transmission from real credit to real house prices is more evident in countries characterized by variable mortgage rates, equity withdrawal, and securitization of mortgage assets. Real money seems to affect house prices in countries with high maximum LTV ratios. Also, the coefficient on long-term interest rates is highly significant in every specification, across every sub-group, and tends to be larger than the coefficient in the broader panel.

Table III.5.

Institutional Factors, Credit, and Real House Prices

(From the baseline regression in Table III.3):

article image
Sources: Table III.3, Tsatsaronis and Zhu (2004); Guiliodori (2004).

First lag only. Interest rate coefficient derived from credit equation.

= t-statistic significant at 1 percent level;

= t-statistic significant at 5 percent level;

= t-statistic significant at 10 percent level.

73. A tentative conclusion, therefore, is that house prices are more sensitive to both interest rates and credit and money among countries with certain kinds of institutions. This is in line with previous research, and backs up the conclusions of Tsatsaronis and Zhu (2004) that more aggressive lending practices strengthen the relationship between house prices and credit, and that countries with variable mortgage rates are associated with larger interest rate effects on house prices. These results are robust to different specifications of the interest rate—real and nominal, short-term and long-term. While the results for each sub-sample could be picking up other factors specific to these countries beyond the trait in question, the use of numerous dimensions to capture institutional effects is reassuring.

74. The affinity for owner occupation can also matter. Table III.5 also isolates countries with owner occupation rates exceeding both 60 and 70 percent—an institutional distinction unrelated to financial markets—on the presumption that the relationship between credit and house prices is more pronounced, the more homeownership is entrenched.68 Owner occupation exceeds 70 percent in Ireland, Spain, and Italy and 60 percent also in Belgium and Finland. It is particularly low (40 percent) in Germany.

75. The different interactions between housing and credit/money variables may be partly related to varying patterns of financial liberalization across the EU, reflectingfinancial catch-up.” There are two potential effects at play. First, the convergence of long-term interest rates across countries in the run-up to EMU could have had an impact on credit and house prices in the countries with previously high interest rates. Second, and in parallel, many countries embarked on extensive financial liberalization over this period. Deregulation began in the early 1980s, and the pace varied markedly across countries. Liberalization typically led to more market-based mortgage markets, increased securitization of mortgage loans, higher loan-to-value ratios and an expansion in mortgage debt. These developments increased the sensitivity of house prices to interest rates (Iacoviello and Minetti, 2003b). They also led to a spurt in credit growth across numerous countries, which could reflect an equilibrium adjustment from repressed to liberalized financial markets.

C. Inflation and House Prices

76. One argument for taking account of asset prices in the conduct of monetary policy is that asset price increases herald future increases in goods and services inflation. If the wealth or credit effect of house prices on consumption is strong, it might herald an uptick in inflation or inflation expectations, at least in the countries characterized by the “right” institutional framework. Such upside risks to inflation will naturally concern central banks. Indeed, some have argued that, while the relationship between stock prices and subsequent output and inflation is weak, house price movements are a much stronger predictor of future goods market trends (Goodhart, 2001).

77. Eye-balling the data suggests a positive relationship between house prices and inflation. Figure III.2 plots the average annual increase in CPI inflation against nominal house price inflation from 1970–2003. A clear relationship is discernible, as those countries with higher housing price growth tend to be those very countries with high goods price growth. Interestingly, the post-EMU picture shows that the relationship has become steeper, as larger volatility in house prices is associated with smaller volatility in inflation. Of course, contemporaneous correlation does not imply that house prices actually drive inflation.

Figure III.2.
Figure III.2.

Inflation and Housing Price Growth in Selected Euro Area Countries

(In percent)

Citation: IMF Staff Country Reports 2005, 266; 10.5089/9781451813029.002.A003

Sources: BIS calculations based on national data; IMF, World Economic Outlook; and IMF staff calculations.

78. But the relationship between lagged asset prices and inflation is not robust in the literature. In a comprehensive study, Stock and Watson (2001) show that for seven countries—including France, Germany, and Italy—asset prices contain little or no predictive power for inflation through two years. Indeed, they find that the only variables that onsistently predict better than simple autoregressions are measures of economic activity, such as the output gap.69 Some have argued that housing prices convey little information that is not captured in other variables, even if statistically significant (Gilchrist and Leahy, 2002; Cecchetti and others, 2000).

79. There are some exceptions, however. Goodhart and Hofmann (2000) find evidence that housing price movements do provide such additional information, while equity prices and yield spreads do not. Using quarterly data on residential property prices for 11 countries, and looking at forecasts up to two years ahead, they find that house prices perform especially well at the two-year horizon. But their sample includes only four euro-area countries (Finland, France, Ireland, the Netherlands) and, within this group, their regressions show housing prices being a significant determinant of inflation only in Ireland.

80. To explore the short- to medium-run predictive power of house prices for goods prices, a simple empirical forecast model for inflation is fitted to the data. Specifically, the following equation is estimated:

ΔPi,t=Ki+δ1ΔPi,t1+δ2ΔPi,t2+δ3Δyi,t1+δ4Δyi,t2+δ5ΔMi,t1+δ6ΔMi,t1+δ7ΔHi,t1+δ8ΔHi,t2+vi,t(4)

where (as before) i denotes a country, t is a time subscript, and Δ represents the first difference operator. In terms of the variables, P is the log of CPI, y is the log of real GDP, M is the log of broad money, and H is the log of nominal house prices. The κi component represents a country fixed effect. Equation (2) is estimated for the panel of eight countries for which house price data are available. The data are annual, and the sample size varies between 26 and 31, depending on data availability. The baseline is a panel regression estimated using fixed effects (LSDV), but, as before, the equation is also estimated using pooled OLS and the Arellano-Bond dynamic panel technique. Country-specific regression results are also reported.

81. This chapter finds scant evidence that lagged house prices help forecast inflation over the short- to medium-run horizon. With these different estimation techniques, there is no significant evidence that house prices feed through to goods price inflation, at either a one or two year lag (Table III.6). In the country-specific OLS estimation, the only country with a positive and significant coefficient on lagged house prices is Italy. Given the potential nonstationarity of inflation, the equations were re-estimated using the differenced inflation as the dependent variable; Goodhart and Hoffman (2000) undertook a similar exercise. Under this specification, there is still no evidence that lagged house prices contain predictive power for CPI inflation. These results are also robust to certain changes in the baseline, including replacing growth with the output gap, and adding unit labor costs and short-term interest rates. Nor does replacing nominal house prices with real house prices make a difference. Some argue that the effects on inflation may be felt beyond the standard two-year horizon, but adding more lags of house prices does not provide further economic information on inflation or differenced inflation. Furthermore, and perhaps surprisingly, the results do not change when controlling for various institutional differences across credit markets and the degree of owner occupation; the kinds of factors that theoretically should determine the link between house prices and the real economy.

Table III.6.

Inflation Equations

(Variables in log differences)

article image
Sources: BIS calculations based on national data; Eurostat; WEO; OECD.

= t-statistic significant at 1 percent level;

= t-statistic significant at 5 percent level;

= t-statistic significant at 10 percent level.

82. But the results do not rule out the possibility of nominal declines in house prices leading to protracted goods price disinflation, or even deflation. House price deflation is a relatively rare event. In the present sample, only five episodes of sustained declines in house prices—defined as three or more consecutive years—stand out (Table III.7). Finland and the Netherlands experienced substantial house price deflation, between 1989–93 and 1978–82 respectively. More limited declines occurred in Belgium, France, and Germany. Table III.7 shows the behavior of goods price inflation, before, during, and after these episodes. In Belgium, Finland, and the Netherlands in particular, there was a sharp fall in inflation in the years following the house price bust. Caution is needed in interpreting these trends, however, given that other factors were clearly at play—the post bust period often coincided with more general cyclical conditions that favored low inflation. Nonetheless, it is striking that countries experiencing the largest busts (Finland and the Netherlands) witnessed significant and prolonged disinflation in the post-bust period; both countries recorded the lowest inflation rates in the sample for the respective time periods. The infrequency of large house price busts makes it difficult to disentangle these effects in the empirical evidence. But the true risks to price stability from asset price boom-bust cycles may well be on the downside.

Table III.7.

Selected Euro Area Countries: Nominal Declines in House Prices and Inflation

article image
Sources: BIS calculations based on national data; Eurostat; OECD.

Average of five years.

D. The Role of Monetary Policy

83. Opinions are divided over how monetary policy should address asset price buildups. At one end of the scale, many would argue that monetary policy should not accord any special role for asset prices, except to the extent that they affect inflationary expectations (the “hands-off” view). On the other side of the debate, some call for monetary policy to take explicit account of asset prices (the “activist view”). This argument comes in different hues. At the most extreme end, some have even called for the inclusion of asset prices directly in the index used to gauge price stability.70 More pertinently, adherents of the activist view have urged central banks to react explicitly to observed asset prices and to “prick” asset price bubbles once they have been properly identified. A third view is more cautious, recognizing both the dangers of action and inaction. Views here have coalesced around the notion that central banks should “lean against the wind” by being tighter than would otherwise be warranted in the face of rising asset prices to contain any bubble developments (see European Central Bank, 2005). As with the activist view, embedded in this approach is a belief that asset prices contain information relevant for price stability, over and above the information contained in the standard indicators.

84. Thehands-offapproach maintains that monetary policy should reflect asset price changes only to the extent that they impinge on expected inflation. Bernanke and Gertler (2001) show that a standard inflation-targeting rule—allowing no additional role for asset prices—stabilizes output and inflation, even when asset prices are volatile. This holds whether the boom is caused by fundamentals or not, and there is no additional benefit in responding to asset prices directly. Gilchrist and Leahy (2002) arrive at similar conclusions. While reasonable in theory, it might be unwise for a central bank to always eschew paying attention to asset prices, particularly if the trend is widespread. This is especially true in light of the potential financial distress and deflation that could result from an asset price bust.

85. Theactivistposition holds that addressing asset price misalignments can deliver superior inflation performance and reduced output volatility. This camp believes that standard inflation targeting is myopic to the extent that it focuses on inflation forecasts at fixed (say, two-year) horizons, whereas the full effects of asset mispricing may take more time to materialize (Cecchetti and others, 2000; Cecchetti, Genberg, and Wadhwani, 2002). Adherents of this viewpoint are careful to note that they do not advocate targeting specific levels of asset prices, or responding mechanically to all changes in asset prices the same way; the key is to isolate non-fundamental changes. Dismissing a frequent complaint, they also claim that measuring misalignments is conceptually no more difficult than estimating potential output or equilibrium real interest rates.

86. Activism faces major implementational hurdles. The problems are manifold:

  • Discerning between fundamental and non-fundamental asset price movements can be trying. The position held by Cecchetti and others (2000) is optimal only when the central bank is certain that the asset price boom is driven by non-fundamentals, and when it will burst—conditions unlikely to be met (Bernanke and Gertler, 2001). Even asset price bubbles are driven partly by fundamentals, being typically associated with real factors like high investment and productivity growth (Detken and Smets, 2004; Filardo, 2004). Moreover, central banks are not privy to private information. A recent survey notes that it is still not possible to isolate asset price bubbles empirically with any degree of clarity (Gurkaynak, 2005). To sum up, in the words of Trichet (2005), “it is very hard to identify them [bubbles] with certainty and almost impossible to reach a consensus about whether a particular asset price boom period should be considered a bubble or not”.

  • Addressing a bubble is fraught with uncertainty. Even if the central bank is reasonably confident that a bubble exists, any misstep with respect to the timing or magnitude of the required tightening could destabilize the economy (Cogley, 1999; Bean, 2004). A large increase in interest rates would probably be needed, with an adverse impact on economic activity. In particular, success depends on a variety of factors, including when the bubble will burst, how protracted the bust will be, and whether it can be defused at low cost.71 Thus, the conditions for using monetary policy to tackle asset price misalignments are highly circumscribed and mistakes can be costly.

87. A key justification forleaning against the windis that it can avoid the build-up of financial imbalances and a subsequent credit crunch (Borio and Lowe, 2002; Filardo, 2004). Such pre-emption has the advantages of activism without the need to cope with the uncertainties surrounding the identification of asset price bubbles. Based on the premise that a negative shock is worse than a positive one, the policymaker is willing to tolerate being below the central bank’s definition of price stability to take out the necessary insurance in the form of lower inflation than would otherwise be justified (Trichet, 2005). Indeed, safeguarding stability of the financial system is an implicit (if not explicit) mandate of many central banks. As noted at the outset, a fall in housing prices could do substantial harm to the health of the banking system and reduce its willingness to extend credit. Deflation that begins in the housing sector could easily become more widespread. Gros, Mayer, and Ubide (2005) argue that the true cost of permitting bubbles to develop comes in the form of a misallocation of resources and economic stagnation in the bust phase rather than inflation in the boom phase. Moreover, all major deflationary episodes throughout the world have been associated with asset price busts (Trichet, 2005), and the association between housing price declines and disinflation in the euro area is documented in Section C. A further advantage of “leaning against the wind” is that the moral hazard created by central bankers responding asymmetrically to shocks is diminished.72

88. Adherents of “leaning against the wind” note that boom-bust cycles tend to be associated with strong growth in monetary and credit aggregates, often in the context of low, stable inflation. Policymakers are thus called upon to pay close attention to money and credit developments and the concomitant build-up of financial imbalances. In this regard, a number of recent empirical studies stand out. First, Borio and Lowe (2002)—based on an analysis of financial crises in 34 countries—argue that the credit gap (deviations of the credit-to-GDP ratio from trends) tends to be the best leading indicator of financial distress. Second, Detken and Smets (2004)—based on a sample of asset price booms for 18 OECD countries since the 1970s—conclude that real money and real credit growth are higher during the early stages of high cost booms. Moreover, real money growth tends to be significantly higher during asset price booms that lead to serious recessions as opposed to those that do not (Trichet, 2005). Thus money and credit growth could set off warnings bells for future price stability, even in a low inflation environment (Issing, 2005). The co-existence of asset price and credit booms with low inflation could be the bane of central bank credibility, or it could reflect favorable productivity developments (Borio and Lowe, 2004).

89. This approach to monetary policy is often geared toward longer horizons, and can be used to justify the ECB’s monetary pillar. Supporters of “leaning against the wind” maintain that liquidity indicators contain information on future output and inflation beyond the standard two-year horizon (Borio and Lowe, 2004). The “horizon” position gels nicely with the ECB’s two-pillar strategy, whereby the monetary pillar concerns itself with longer-term price pressures. In the same context, Jaeger (2003) argues that having an explicit pillar focusing on money and credit could guard against the build-up of area-wide asset bubbles. Trichet (2005) argues that the ECB’s approach is superior to inflation targeting in this regard, even if inflation forecast horizons under the latter are extended beyond the standard 1-2 years. Indeed, the evidence from Section B points to a long-run relationship between real credit/money and real house prices, even in the absence of a clear short-to medium-term one.

90. All in all, thelean against the windposition is attractive, but difficult to operationalize in the euro area.73 Some of the difficulties in applying the activist position successfully also apply here. In particular, for “leaning against the wind” to work, the probability of the bubble bursting soon should be low, and the growth in the bubble should be interest sensitive. Also, as demonstrated earlier in this chapter, the relationship between credit and housing prices in the area is not clearly defined, depending on country-specific institutions. In particular, the predictive power of credit and monetary aggregates for real houses prices over the short- to medium-run horizon appears to be confined to a subset of countries: those with more market-based credit markets, more aggressive lending, and higher levels of owner-occupation. Others have noted that asset price booms (and housing price booms in particular) tend to occur more frequently in small countries, and are particularly rare in France, Germany, and Italy (Detken and Smets, 2004; Bordo and Jeanne, 2002). This alone could diminish the adverse impact of a “contagion” effect of a bust phase across countries. Also, this chapter has shown that there is little evidence of housing prices helping predict goods prices in the short- to medium run. Of course, this does not rule out deflation risks. But the need to keep a watchful eye on both inflationary and deflationary pressures at the same time can prove especially challenging.

91. Given these concerns, a central bank focusing on potential asset price booms could face communications problems. This arises from (i) the imprecise link between monetary policy, asset price cycles, and the real sector, (ii) the complexity of the optimal rule, and (iii) the need to match up instruments with goals to ensure accountability (Mishkin, 2001; Issing, 2003). Disyatat (2005) argues that pre-emption against the build-up of financial imbalances really implies putting financial imbalances in the central bank loss function, and this leads to less transparency and greater uncertainty in communications. Communication is harder when asset price trends are at odds with price stability indicators at the standard horizon, and when the central bank needs to signal both upside and downside risks to price stability.

92. But it would be imprudent to downplay the risks to price stability and economic activity from surging asset prices. First, financial liberalization can increase asset price volatility, and some have argued that the financial deregulation in Europe from the 1980s onwards contributed to an increase in the number of asset price booms (Detken and Smets, 2004). Second, a low inflation environment increases the risk of deflation in the event of nominal declines in house prices. Moreover, while the boom phase may be localized in certain markets, the damage caused by the bust phase could become more widespread, particularly if combined with limited flexibility in factor markets (Gros, Mayer, and Ubide, 2005).

93. However, other instruments appear better suited than monetary policy to address house price developments head-on. Concerns surrounding house price booms may be better addressed at the national level, through fiscal policy measures and financial sector regulation/supervision. In the first instance, policymakers can target the various tax deductions and allowances, as well as subsidies, that provide support to house prices. They could also consider responses such as encouraging fixed-rate mortgages, placing tough prudential upper limits on loan-to-value ratios, and promoting a private rental sector (Maclennan, Muellbauer, and Stephens, 1999). More broadly, Schwarz (2002) argues that capital requirements should be put in place that would increase with the growth of credit collateralized by assets with booming prices. In a similar vein, Borio, Furfine, and Lowe (2001) recommend cyclically sensitive capital requirements (raising them in booms, reducing them in recessions). Using prudential means to control local house prices in a regionally-integrated financial sector, however, raises other problems.

E. Conclusions

94. This chapter has argued that the nexus between monetary policy, credit growth, and house prices across the euro area is far from clear. In particular:

  • House prices have behaved very differently across euro-area countries. Also, the relationship between credit, money, and house prices appears to differ.

  • The short- to medium-run predictive power of credit and monetary aggregates for asset prices is uncertain, depending on country-specific institutional factors. Most notably, the factors that lend themselves to a more robust link between credit, interest rates, and house prices are generally absent in the largest members. Over the long run, however, real money and credit do help predict real house prices.

  • The short- to medium-run predictive power of house prices for goods prices is tenuous. There is little evidence that house price inflation feeds through to goods price inflation. The real risk from the point of view of economic activity might be the consequences of a housing price bust, including deflation—a much rarer, and very harmful, event.

In such an environment, pre-emptive monetary policy is difficult to operationalize for the euro area. Accordingly, other, national, policy instruments—fiscal and financial—might be more appropriate tools to reign in surging asset prices, if deemed necessary.

95. Looking ahead, many of the problems created by differences in local institutions could potentially be solved by fostering more integrated mortgage markets across the euro area. For the United States, Schnure (2005) shows that the shift from bank-based mortgage lending to a system of securitized mortgage finance since the mid-1980s was associated with a reduction in the volatility of credit growth and housing prices across U.S. regions. Convergence was fostered by the integration of banking markets and increasing securitization of mortgage loans, leading, in essence, to a national mortgage market. This offers obvious lessons for the euro area. Despite lower interregional migration, there is more divergence of house prices across EU countries than across the different regions of the United States (see Chapter VI). Integration of mortgage markets through securitization across the euro area could potentially bring about similar convergence, also improving the effectiveness of monetary policy.

References

  • Bean, Charles, 2004, “Asset Prices, Monetary Policy, and Financial Stability: A Central Banker’s View,” unpublished, Bank of England.

    • Search Google Scholar
    • Export Citation
  • Bernanke, Ben S., and Mark Gertler, 2001, “Should Central Banks Respond to Movements in Asset Prices,” American Economic Review Papers and Proceedings, Vol. 91, No. 2, pp. 253257.

    • Search Google Scholar
    • Export Citation
  • Bernanke, Ben S., Mark Gertler, and Simon Gilchrist, 2000, “The Financial Accelerator in a Quantitative Business Cycle Framework,” in J. Taylor and M. Woodford, eds., Handbook of Macroeconomics. Amsterdam: North Holland, pp. 13411393.

    • Search Google Scholar
    • Export Citation
  • Bordo, Michael D., and Olivier Jeanne, 2002, “Boom-Busts in Asset Prices, Economic Instability, and Monetary Policy,” NBER Working Paper, No. 8966.

    • Search Google Scholar
    • Export Citation
  • Borio, Claudio, and Patrick McGuire, 2004, “Twin Peaks in Equity and Housing Prices?BIS Quarterly Review, March, pp. 7993.

  • Borio, Claudio, and Philip Lowe, 2002, “Asset Prices, Financial and Monetary Stability: Exploring the Nexus,” BIS Working Papers, No. 114.

    • Search Google Scholar
    • Export Citation
  • Borio, Claudio, and Philip Lowe, 2004, “Securing Sustainable Price Stability: Should Credit Come Back from the Wilderness,” BIS Working Papers, No. 157.

    • Search Google Scholar
    • Export Citation
  • Borio, Claudio, Craig Furfine, and Philip Lowe, 2001, “Procyclicality of the Financial System and Financial Stability: Issues and Policy Options,’ in Marrying the Macro- and Micro-Prudential Dimensions of Financial Stability, BIS Papers No. 1, pp. 157.

    • Search Google Scholar
    • Export Citation
  • Cecchetti, Stephen, Hans Genberg, and Sushil Wadhwani, 2002, “Asset Prices in a Flexible Inflation Targeting Framework,” unpublished.

    • Search Google Scholar
    • Export Citation
  • Cecchetti, Stephen, Hans Genberg, John Lipsky, and Sushil Wadhwani, 2000, Asset Prices and Central Bank Policy, London: International Center for Monetary and Banking Studies.

    • Search Google Scholar
    • Export Citation
  • Cogley, Timothy, 1999, “Should the Fed Take Deliberate Steps to Deflate Asset Price Bubbles?,” FRBSF Economic Review, No. 1, pp 4352.

    • Search Google Scholar
    • Export Citation
  • Detken, Carsten, and Frank Smets, 2004, “Asset Price Booms and Monetary Policy,” ECB Working Paper, No. 364.

  • Disyatat, Piti, 2005, “Inflation Targeting, Asset Prices, and Financial Imbalances: Conceptualizing the Debate,” BIS Working Papers, No. 168.

    • Search Google Scholar
    • Export Citation
  • European Central Bank, 2003, Structural Factors in the EU Housing Markets, March.

  • European Central Bank, 2005, “Asset Price Bubbles and Monetary Policy,” ECB Monthly Bulletin April, pp. 4760.

  • Filardo, Andrew, 2004, “Monetary Policy and Asset Price Bubbles: Calibrating the Monetary Policy Trade-Offs,” BIS Working Papers, No. 155.

    • Search Google Scholar
    • Export Citation
  • Gilchrist, Simon, and John V. Leahy, 2002, “Monetary Policy and Asset Prices,” Journal of Monetary Economics, Vol. 49, pp. 7597.

  • Giuliodori, Massimo, 2004, “Monetary Policy Shocks and the Role of House Prices Across European Countries,” DNB Working Paper, No. 015/2004.

    • Search Google Scholar
    • Export Citation
  • Goodhart, Charles, 2001, “What Weight Should be Given to Asset Prices in the Measurement of Inflation?The Economic Journal, Vol. 111, pp. F335F356.

    • Search Google Scholar
    • Export Citation
  • Goodhart, Charles, and Boris Hofmann, 2000, “Do Asset Prices Help to Predict Consumer Price Inflation?The Manchester School Supplement, Vol. 68, pp. 122140.

    • Search Google Scholar
    • Export Citation
  • Gros, Daniel, Thomas Mayer, and Angel Ubide, 2005, EMU at Risk.7th Report of the CEPS Macroeconomic Policy Group. Brussels: Center for European Policy Studies.

    • Search Google Scholar
    • Export Citation
  • Gruen, David, Michael Plumb, and Andrew Stone, 2003, “How Should Monetary Policy Respond to Asset-Price Bubbles?,” Research Discussion Paper No. 2003-11, Reserve Bank of Australia.

    • Search Google Scholar
    • Export Citation
  • Gurkaynak, Refet S., 2005, “Econometric Tests of Asset Price Bubbles: Taking Stock,” Finance and Economics Discussion Series, No. 2005-04, Divisions of Research and Statistics and Monetary Affairs, Federal Reserve Board.

    • Search Google Scholar
    • Export Citation
  • Haque, Nadeem Ul., M. Hashem Pesaran, and Sunil Sharma, 1999, “Neglected Heterogeneity and Dynamics in Cross-Country Savings Regressions,” IMF Working Paper, No. 99/128.

    • Search Google Scholar
    • Export Citation
  • Illing, Gerhard, 2001, “Financial Fragility, Bubbles, and Monetary Policy,” unpublished, University of Munich.

  • Iacoviello, Matteo, and Raoul Minetti, 2003a, “The Credit Channel of Monetary: Evidence from the Housing market,” unpublished.

  • Iacoviello, Matteo, and Raoul Minetti, 2003b, “Financial Liberalization and the Sensitivity of House Prices to Monetary Policy: Theory and Evidence,” The Manchester School, Vol. 71, No. 1, pp. 2034.

    • Search Google Scholar
    • Export Citation
  • Issing, Otmar, 2003, “Asset Prices and Monetary Policy,” unpublished, European Central Bank.

  • Issing, Otmar, 2005, “The Monetary Pillar of the ECB,” unpublished, European Central Bank

  • Jaeger, Albert, 2003, “The ECB’s Money Pillar: An Assessment.” IMF Working Paper, No. 03/82.

  • Judson, Ruth A., and Ann L. Owen, 1999, “Estimating Dynamic Panel Data Models: A Guide for Macroeconomists,” Economics Letters, Vol. 65, pp. 915.

    • Search Google Scholar
    • Export Citation
  • Mclennan, Duncan, John Muellbauer, and Mark Stephens, 1999, “Asymmetries in Housing and Financial Market Institutions and EMU,” CEPR Discussion Paper No. 2062.

    • Search Google Scholar
    • Export Citation
  • Mishkin, Frederic S., 2001, “The Transmission Mechanism and the Role of Asset Prices in Monetary Policy,” NBER Working Paper, No. 8617.

    • Search Google Scholar
    • Export Citation
  • Orphanides, Athanasios, and Simon van Norden, 2002, “The Unreliability of Output Gap Estimates in Real Time,” Review of Economics and Statistics, Vol. 84, pp. 569583.

    • Search Google Scholar
    • Export Citation
  • Schnure, Calvin, 2005, “Boom-Bust Cycles in Housing: the Changing Role of Financial Structure,” unpublished, International Monetary Fund.

    • Search Google Scholar
    • Export Citation
  • Schwarz, Anna J., 2002, “Asset Price Inflation and Monetary Policy,” NBER Working Paper, No.9321.

  • Stock, James H., and Mark W. Watson, 2001, “Forecasting Output and Inflation: the Role of Asset Prices,” NBER Working Paper, No. 8180.

    • Search Google Scholar
    • Export Citation
  • Sutton, Gregory D., 2002, “Explaining Changes in House Prices,” BIS Quarterly Review, September, pp. 4655.

  • Tetlow, Robert J., 2004, “Monetary Policy, Asset Prices, and Misspecification: the Robust Approach to Bubbles with Model Uncertainty,” unpublished, Board of Governors of the Federal Reserve System.

    • Search Google Scholar
    • Export Citation
  • Trichet, Jean-Claude, 2005, “Asset Price Bubbles and Monetary Policy,” unpublished, EuropeanCentral Bank.

  • Tsatsaronis, Kostas, and Haibin Zhu, 2004, “What Drives Housing Price Dynamics: Cross-Country Evidence,” BIS Quarterly Review, March, pp. 6578.

    • Search Google Scholar
    • Export Citation
61

Prepared by Anthony Annett (EUR).

62

The euro area sample (eight countries) is dictated by data availability on real house prices. Austria, Greece, Luxembourg, and Portugal are not included.

63

Here, and throughout this chapter, “credit” means “credit to total residents granted by monetary financial institutions (consolidated).” This was the only historical series available on a consistent basis for all countries from Eurostat.

64

The money series is longer than the credit series, spanning about 30 years, instead of only 20 for credit.

65

Indeed, Belgium was omitted altogether, for data availability reasons.

66

To capture demographic effects, the long-run equation was also estimated with the log of population as an explanatory variable. This variable did not yield a significant coefficient over the period analyzed; not did it affect any of the other coefficients or standard errors in any significant way. As population varied little in most countries over time (with the exception of Ireland), differences in population would likely be captured by the country fixed effects.

67

The source of this categorization is Tsatsaronis and Zhu (2004).

68

The source of these data is Guiliodori (2004).

69

Note that output gaps are notoriously difficult to measure in real time, which could lead to inappropriate monetary policy (Orphanides and van Norden, 2002).

70

See Goodhart (2001) for an exposition of the issues, and European Central Bank (2005) for a detailed description of the conceptual and implementation difficulties that would surround such a proposal.

72

By loosening in the bust phase but not tightening in the boom, monetary policy can foster excessive risk-taking on the part of investors (Illing, 2001).

73

A similar challenge in operationalizing the ECB’s monetary pillar is discussed in Chapter II.

Euro Area Policies: Selected Issues
Author: International Monetary Fund