Asset Prices and Current Account Fluctuations in G-7 Economies
  • 1 0000000404811396https://isni.org/isni/0000000404811396International Monetary Fund

The paper analyses the effect of equity-price shocks on current account positions for the G-7 industrialized countries during 1974–2007. It uses a Bayesian vector autoregression with sign restrictions for the identification of equity-price shocks and to test empirically for their effect on current accounts. Such shocks are found to exert a sizable effect, with a 10 percent equity price increase, for example, in the United States relative to the rest of the world, worsening the U.S. trade balance by 0.9 percentage points after 16 quarters. However, the response of the trade balance to equity-price shocks varies substantially across countries. The evidence suggests that the channels accounting for this heterogeneity function both through wealth effects on private consumption and to some extent through the real exchange rate of countries.

Abstract

The paper analyses the effect of equity-price shocks on current account positions for the G-7 industrialized countries during 1974–2007. It uses a Bayesian vector autoregression with sign restrictions for the identification of equity-price shocks and to test empirically for their effect on current accounts. Such shocks are found to exert a sizable effect, with a 10 percent equity price increase, for example, in the United States relative to the rest of the world, worsening the U.S. trade balance by 0.9 percentage points after 16 quarters. However, the response of the trade balance to equity-price shocks varies substantially across countries. The evidence suggests that the channels accounting for this heterogeneity function both through wealth effects on private consumption and to some extent through the real exchange rate of countries.

Current account positions have hardly ever been so dispersed globally as they are today. It is not only that the largest economy, the United States, has been recording a current account deficit in excess of 5 percent for several years, but other industrialized countries, such as the United Kingdom and Australia, and some emerging markets and transition economies have similar or even larger deficits. By contrast, countries such as China, Japan and oil exporters register corresponding large trade surpluses. At the same time, asset prices have gone through a marked cycle over the past decade, with equity markets rising substantially in the second half of the 1990s and in 2002–06 and declining in 2001–02. The financial market crisis of 2007–08 has made the importance of asset prices for the global economy more than apparent. Despite the financial crisis, the role of asset prices for the global economy will most likely increase further as financial markets deepen and emerging economies liberalize and integrate.

This paper analyses the impact of asset (equity) price shocks on the current account. The objective is not only to grasp the magnitude of the effect of asset prices on trade, but also to understand the channels through which this effect materializes. Asset-price shocks affect net exports through a wealth channel as households adjust saving and consumption decisions, and through an exchange rate and terms-of-trade channel, altering the relative prices of domestic and foreign goods. Equally importantly, asset prices may exert different effects across economies, as those with deeper yet more closed financial markets may respond more strongly.

The focus of this paper is on the G-7 industrialized countries and on the role of equity-price shocks during 1974–2007. We use the sign restrictions derived in Fratzscher and Straub (2008), who build an open-economy dynamic stochastic general equilibrium model, in which changes to asset prices influence private consumption through wealth effects. We then employ a Bayesian vector autoregression (VAR), following Canova and de Nicoló (2002), Uhlig (2005), and Peersman (2003), using sign restrictions to test for the effect of asset price shocks in the data. This methodology not only requires imposing a relatively small and intuitive number of identification restrictions, but importantly it also allows us to distinguish asset-price shocks from other types of structural shocks. Our empirical implementation follows closely that of Fratzscher, Juvenal, and Sarno (2007), who test for the effect of equity-market shocks, housing-price shocks and exchange rate shocks on the trade balance of the United States. They show that equity-market shocks and housing-price shocks have been important drivers explaining more than 30 percent of the variation of the U.S. trade balance, whereas exchange rates account for a much smaller share.

Our empirical findings show that asset prices exert a sizable effect on the trade balance of countries. The channels through which equity prices influence net exports are both through wealth effects on private consumption and to some extent through the exchange rate. An increase in asset prices tends to have a positive impact on short-term interest rates and inflation, and leads to an appreciation of the real effective exchange rate (REER) and a sizeable increase in consumption. Moreover, we find a large degree of crosscountry heterogeneity in the impulse response pattern. The U.S. trade balance is among the most sensitive as net exports, on average, decline by 0.91 percentage points after 16 quarters in response to a 10 percent increase in U.S. equity prices relative to the rest of the world. The trade balances of most other countries react substantially less.

The paper is related to three fields of the literature. A first strand focuses on the drivers of the large and persistent global current account imbalances. Several papers emphasize the importance of a saving glut (Bernanke, 2005) in many emerging markets and commodity-exporting countries, partly stemming from the underdevelopment and lack of integration of financial markets in those economies (Caballero, Farhi, and Gourinchas, 2006; Ju and Wei, 2006; Fratzscher, 2008), as well as the increasing role of ensuing valuation effects on gross international asset positions (Lane and Milesi-Ferretti, 2005; Gourinchas and Rey, 2007) and a precautionary motive as a rationale for high saving rates (for example, Chinn and Ito, 2007; Gruber and Kamin, 2007). Other studies to explain the dispersion in current account positions stress the role of productivity differentials (for example, Bussiere, Fratzscher, and Muller, 2005; Corsetti, Dedola, and Leduc, 2006), or link it to the great moderation, which has induced a decline in income volatility and uncertainty (Fogli and Perri, 2006).

As to the second area, a vast amount of literature identifies and measures the effect of price changes in various financial assets on private consumption (for example, Bertaut, 2002; Case, Quigley, and Shiller, 2005). Most of this literature finds a significant effect of both equity wealth and housing wealth on private consumption. However, there is still substantial controversy as to the magnitude and precise functioning of this channel as for instance exemplified by the conflicting results found by Palumbo and Whelan (2006) and Lettau and Ludvigson (2004). The effect of such a wealth channel on the external dimension of countries, in particular the current account and the exchange rate, has so far received little attention in the literature. From a current policy perspective, it has been argued by some that the U.S. dollar decline would have to be very large, as suggested by several studies (Blanchard, Giavazzi, and Sa, 2005; Obstfeld and Rogoff, 2005; Krugman, 2007).

The third area relates to the crucial issue of the structural interpretation of asset-price shocks. Although we can separate an asset-price shock from the standard macroeconomic shocks usually analyzed, it is not clear what asset-price changes represent structurally. One interpretation of an asset-price shock is that of a news shock, along the line of work by Beaudry and Portier (2006 and 2007), in which asset prices adjust because of altered expectations about the likelihood of future outcomes, such as to economic fundamentals. Such changes in expectations should then, in turn, be reflected in today’s asset prices as these represent the net discounted value of all future fundamentals.

This is also related to the work by Engel and Rogers (2006), who show that the large size of the U.S. current account deficit is consistent with expectations of an increasing share of U.S. output in the world. An alternative interpretation is that asset-price shocks reflect rational bubbles, as in Kraay and Ventura (2005) and Ventura (2001). They argue that the sharp increase in asset prices over the past decade may largely reflect a bubble, which is rational because of market expectations that this increase may be persistent. Both interpretations are observationally equivalent to what we understand and see in the behavior of economic fundamentals. We are agnostic about these interpretations; the crucial point is that asset-price shocks reflect factors that function primarily through asset prices. The purpose and intended contribution of this paper is to improve our understanding of how this asset-price channel functions.

I. Methodology

Deriving the Sign Restrictions

This section discusses the set of sign restrictions to identify asset-price shocks. Thereby, we apply the strategy discussed, for example, in Peersman and Straub (forthcoming). In particular, we utilize restrictions, discussed in detail in Fratzscher and Straub (2008), which identify asset-price shocks uniquely and distinguish them from a set of other shocks that are discussed as determinants of current account fluctuations in the literature. Table 1 summarizes the sign restrictions used for the identification in our structural VAR. We associate positive asset-price shocks (that is an exogenous increase in asset prices) with a rise in consumption, inflation, and interest rates. As discussed in Fratzscher and Straub (2008), the rise in current stock market wealth triggers an increase in private consumption. The latter induces a surge in inflation rates, and under the assumption of an active monetary policy rule, an increase in interest rates.

Table 1.

Theoretical Impulse Response Functions

article image

Note that the latter reaction is fundamentally different from the response following technology and monetary-policy shocks. Technology shocks, for example, trigger a rise in consumption and a fall in inflation rates. monetary policy shocks induce a positive response of consumption and inflation and are characterized by a fall in interest rates.1 Note that, although we base our sign restriction identification strategy on the predictions of a theoretical model, we do not have to restrict the response of the current account and the real exchange rate, the main variables of interest. In this respect, we can let the data speak for itself.

A crucial issue is the structural interpretation of asset-price shocks. The identifying restrictions above separate an asset-price shock from the standard macroeconomic shocks usually analyzed (technology, monetary policy, and government spending), without identifying the structural factors behind the asset price increase. Note that other demand-.side shocks such as shocks to time preferences or distortionary taxes might imply, under certain assumptions, similar patterns for the endogenous variables as asset-price shocks.

On the other hand, exogenous changes in distortionary taxes or time-preference rates are unlikely to be an important determinant of business cycles at a quarterly level.

What is our interpretation of asset-price shocks? As discussed above, one interpretation of an asset-price shock is that of a news shock (Beaudry and Portier, 2006 and 2007), in which asset prices adjust because of changed expectations about the likelihood of future outcomes, such as to economic fundamentals; or as in Engel and Rogers (2006), where current account changes are consistent with changing expectations of relative output shares.

Alternatively, asset-price shocks can echo rational bubbles, as in Kraay and Ventura (2005) and Ventura (2001). We are agnostic about these interpretations; the main point is that asset-price shocks reflect factors that function primarily through asset prices. The objective of the empirical exercise is to illustrate how this asset-price channel functions.

Model Specification and Data

Consider the following specification for a vector of endogenous variables Yt:

Yt=α+Σi=1nAiYti+Bεt,(1)

where α is a (n × 2) matrix of constants and linear trends, Ai is an n × n matrix of autoregressive coefficients and εt is a vector of structural disturbances. Identification of the impact of structural disturbances requires imposing n(n—1)/2 restrictions on B, which we achieve by using the sign restrictions shown in Table 1. Our sign restriction approach is based on Canova and De Nicolo (2002), Uhlig (2005), and Peersman (2003), discussed in some detail in the next section.

Our VAR includes six variables: Yt = [EQ, c, i, π, TB, REER], a relatively standard specification as, for instance, also used in Fratzscher, Juvenal, and Sarno (2007), that is, private consumption (c), short-term interest rates (i), inflation (π), equity returns (EQ), as well as the trade balance (TB) and the REER.

Our country sample focuses on the G-7 industrialized countries. The time period for the empirical analysis is 1974–2007, using quarterly data. We use 1974 as the starting point of the analysis as it is the start of the floating exchange rate period after the collapse of the Bretton Woods system. We limit our analysis to the G-7 economies.2

For our empirical estimation we use relative variables, that is, we specify each variable in domestic vs. rest-of-the-world terms. More precisely, consumption c is the difference in log private consumption in the domestic economy and log private consumption in the rest of the world, both expressed in U.S. dollar (using end-of-period exchange rates). Interest rates i are the percentage difference of domestic short-term (money market) rates from those in the rest of the world, whereas inflation π is the corresponding percentage difference in consumer price index inflation. The rest of the world for all three variables comprises the other economies in the benchmark sample with each country being weighted by its GDP share in the sample group.

Our preferred measure of asset prices EQ is the difference between domestic equity returns and foreign equity returns, both measured in local currency terms. We use local currencies to express returns, rather than U.S. dollars, because we want to obtain a measure of asset-price shocks that excludes exchange rate movements.3 Moreover, we use shocks to equity prices, rather than changes to market capitalization, as our preferred measures because our primary interest is in the cross-country heterogeneity in the responses of the trade balance and the exchange rate. The rest-of-the-world group comprises the other countries in the sample, with each of these countries being weighted by their equity-market capitalization. We use equity-market capitalization weights, rather than GDP weights, because equity shocks are likely to affect the trade balance of countries partly through wealth effects, which in turn should be related to the size of financial wealth held by households, which is better proxied by market capitalization than GDP. In the section on the robustness analysis below we will discuss how alternative specifications of asset-price shocks influence the empirical findings.

The trade balance TB is measured as a ratio to domestic GDP. We use the trade balance, rather than the current account, as we are interested in the effect of asset-price shocks on net exports and want to exclude the effect on income. As the final variable, the REER uses trade weights for a broad set of partner countries, and is expressed in logs.

As to the data sources, the trade balance, consumption, inflation and short-term interest rates come from the IMF’s International Financial Statistics (IFS). Equity returns and equity-market capitalization are market indices and are sourced from Bloomberg whereas we took the REERs from the IFS and the Organization for Economic Cooperation and Development.4

Implementation of the Sign Restrictions

Before moving on to the empirical results, it is useful to explain how we implement the sign restrictions in our VAR. For a detailed explanation, we refer to Peersman (2003). Consider Equation (1). As the shocks are mutually orthogonal, E(εtεt)=I,, the Variance-Covariance of the reduce form residuals of Equation (1) is equal to Ω=BB. For any possible orthogonal decomposition B, we can find an infinite number of admissible decompositions of Ω, Ω=BQQ’B’, where Q is any orthonormal matrix, that is, QQ = I. Possible candidates for B are the Choleski factor of Ω or the eigenvalue-eigenvector decomposition, Ω=PDP = BB, where P is a matrix of eigenvectors, D is a diagonal matrix with eigenvalues on the main diagonal, and B = PD1/2. Following Canova and De Nicoló (2002) and Peersman (2003), we start from the latter in our analysis. More specifically, P = Πm,n Qm,n (θ) with Qm,n (θ) being rotation matrices of the form

Qm,n(θ)=[10000cos(θ)sin(θ)010sin(θ)cos(θ)00001].(2)

As we have six variables in our model, there are n(n—1)/2= 15 bivariate rotations of different elements of the VAR: θ = θ1, …, θ15, and rows m and n are rotated by the angle θi in Equation (2). All possible rotations can be produced by varying the 15 parameters θi in the range [0, π]. For the contemporaneous impact matrix determined by each point in the grid, Bj, we generate the corresponding impulse responses

Rj,t+k=A(L)1Bjεt.

A sign restriction on the impulse response of variable p at lag k to a shock in q at time t is of the form

Rj,t+kpq0orRj,t+kpq0.

We impose the sign restrictions for k = 4 lags; choosing a different length, however, does not alter the findings in a meaningful way. Following Uhlig (2005) and Peersman (2003), we use a Bayesian approach for estimation and inference. Our prior and posterior belong to the Normal-Wishart family for drawing error bands. Because there are an infinite number of admissible decompositions for each draw from the posterior when using sign restrictions, we use the following procedure. To draw the candidate truths from the posterior, we take a joint draw from the posterior for the usual unrestricted Normal-Wishart posterior for the VAR parameters as well as a uniform distribution for the rotation matrices, using 1,000 draws. We then construct impulse response functions. If all the imposed conditions of the impulse responses are satisfied, we keep the draw, whereas other decompositions are rejected. This means that these draws receive zero prior weight. Based on the draws kept, we calculate statistics and report the median responses, together with 84th and 16th percentile error bands.

II. Empirical Results

This section presents the empirical results from the structural VAR with sign restrictions, applied to the G-7 economies in the period 1974–2007. We also present various extensions to check for the sensitivity and robustness of the findings.

Benchmark Results

Figures 17 shows the impulse responses of the six variables, for each of the countries in our country sample of G-7 countries, to a 10 percent positive equity-market shock based on our Bayesian VAR model. The shaded areas indicate the 16 and 84 percentiles of the posterior distribution, following the convention in the literature. Table 2 summarizes the point estimates of the impulse responses at various time horizons.

Figure 1.
Figure 1.

United States: Impulse Response Following an Asset-Price Shock

Citation: IMF Staff Papers 2009, 003; 10.5089/9781589068209.024.A008

Note: The figure shows the impulse responses of the six variables, for each of the countries in our sample of G-7 economies, to a 10 percent positive equity-market shock based on our Bayesian vector autoregression model. The shaded areas indicate the 16th and 84th percentiles of the posterior distribution.
Figure 2.
Figure 2.

United Kingdom: Impulse Response Following an Asset-Price Shock

Citation: IMF Staff Papers 2009, 003; 10.5089/9781589068209.024.A008

Note: See note to Figure 1.
Figure 3.
Figure 3.

Germany: Impulse Response Following an Asset-Price Shock

Citation: IMF Staff Papers 2009, 003; 10.5089/9781589068209.024.A008

Note: See note to Figure 1.
Figure 4.
Figure 4.

France: Impulse Response Following an Asset-Price Shock

Citation: IMF Staff Papers 2009, 003; 10.5089/9781589068209.024.A008

Note: See note to Figure 1.
Figure 5.
Figure 5.

Italy: Impulse Response Following an Asset-Price Shock

Citation: IMF Staff Papers 2009, 003; 10.5089/9781589068209.024.A008

Note: See note to Figure 1.
Figure 6.
Figure 6.

Canada: Impulse Response Following an Asset-Price Shock

Citation: IMF Staff Papers 2009, 003; 10.5089/9781589068209.024.A008

Note: See note to Figure 1.
Figure 7.
Figure 7.

Japan: Impulse Response Following an Asset-Price Shock

Citation: IMF Staff Papers 2009, 003; 10.5089/9781589068209.024.A008

Note: See note to Figure 1.
Table 2.

Impulse Response to a 10 Percent Domestic Asset-Price Shock

article image

As to the United States (Figure 1), a 10 percent increase in (relative) U.S. equity prices leads to a substantial worsening in the U.S. trade balance. The effect of the asset-price shock increases gradually over time up to 16–20 quarters, when it reduces the U.S. trade balances by 0.91 percentage points of U.S. GDP. This effect of asset prices on the trade balance appears to stem from two channels, a first one through wealth effects and a second related to the exchange rate. The importance of wealth effects is evident by the strong and quite persistent increase in private consumption, which in turn leads to a higher demand for imports.

The role of the exchange rate channel is underlined by the significant appreciation of the REER after a positive asset-price shock. The real appreciation is likely to be influenced both by the increase in domestic inflation and in domestic interest rates, though both of these responses are more short-lived as inflation and nominal interest rates revert back within 10 quarters. The rise in interest rates and real appreciation of the exchange rate is consistent with the evidence of the presence of a significant forward discount bias found in the literature (for example, Engel, 1996), as well as the more recent evidence stressing the importance of monetary policy or “Taylor-rule” fundamentals for exchange rate determination (Engel and West, 2005; Clarida and Waldman, 2007).5

Figures 27 shows the corresponding impulse responses for the other G-7 countries of the sample. With a few exceptions, the patterns of the impulse responses are quite similar across countries: the trade balance of most countries deteriorates in response to a positive equity-price shock, though the permanence of this response is mostly somewhat lower than that of the United States. Moreover, the real exchange rate and private consumption always increases over the medium run after an increase in equity prices, though again the permanence of this effect differs markedly across countries. The strength of the reaction of private consumption for most countries suggests that wealth effects constitute an important channel through which asset-price shocks affect the trade balance of countries.

Nominal interest rates and inflation also rise in the short run, though recall that we imposed this response for the first four quarters in order to identify equity-price shocks. However, the magnitude and the persistence of the reaction of interest rates and inflation again differ substantially across countries. We also note and show the impulse responses for countries with somewhat puzzling results. For instance, the trade balance for the United Kingdom (Figure 2) improves in response to a positive domestic asset-price shock. We will return to a detailed discussion of these and other crosscountry differences in the subsequent section.

Table 2 illustrates the heterogeneity of the point estimates at different time horizons, after one quarter, eight quarters, and 16 quarters, respectively. The table shows the marked differences in the impulse responses across countries, in the magnitude as well as in the dynamics and timing of the transmission of equity-price shocks. For instance, Italy’s trade balance appears to react relatively quickly to asset-price shocks, with the impulse response reverting back to zero relatively quickly. By contrast, the opposite is the case for the United Kingdom, where the reduction in the trade balance materializes only after several quarters.

Table 2 also nicely illustrates the different channels that are at play in transmitting the equity-price shock to the trade balance of countries. Countries that experience a stronger reaction of their trade balance to the asset-price shock also exhibit a larger response of their REER as well as private consumption. For instance, for the reaction after 16 quarters, it is in particular the United States but also Germany and Canada that see the strongest response of their trade balance, yet also experience a relatively larger sensitivity of private consumption and of their REER to the equity-price shock. Hence, this suggests that both a wealth effect on private consumption and an exchange rate channel are at play in explaining the transmission of an asset-price shock to a country’s trade balance.

Finally, the current financial market turmoil has further increased the focus on the role of monetary policy in addressing asset prices, and in particular asset-price bubbles. What do the impulse responses tell us about the potential role of monetary policy as a channel through which asset-price shocks may be transmitted to the trade balance of countries? In principle, one would expect that an aggressive tightening of monetary policy in response to a positive asset-price shock should dampen the effect of this shock on consumption and thus on net exports through the wealth channel. However, on the other hand, such a tightening may lead to an appreciation of the exchange rate and a worsening of the trade balance. Based on the impulse responses in Figures 17, and the summary of these impulse responses shown in Table 2, there seems to be no clear-cut relationship between the response of interest rates, private consumption, inflation, and the trade balance across countries. This of course is no more than suggestive, and does not necessarily imply that monetary policy is not relevant for influencing the impact of asset prices on the trade balance. However, for instance for the United States these findings suggest that the reaction of U.S. short-term interest rates to asset-price shocks is not systematically lower than that of other industrialized countries.

Robustness and Extensions

How robust are these findings across alternative specifications, country samples, and time periods? We conduct several robustness tests on the benchmark model.

One important issue is how dependent our empirical findings are on the identification, that is, the sign restrictions we impose. Although these sign restrictions seem sensible, it is nevertheless useful to see how the results change when using alternative identification methods. We do so by estimating our six-variable VAR using a Choleski decomposition. More precisely, we estimate the VAR using each possible ordering of the six endogenous variables, and then check the distribution of the resulting impulse response functions. Figure 8 shows the impulse responses of the trade balance to a positive asset-price shock for the case of the United States, the United Kingdom, Germany, and France. The top of the shaded area represents to maximum response coefficient among the different Choleski decompositions, whereas the lower end shows the minimum response at any time horizon. Overall, the findings suggest that the direction of the trade balance response to an asset-price shock is mostly the same when taking the Choleski decomposition as when using sign restrictions. However, the range of possible impulse responses is in several cases very large, underlining that the shape of the impulse responses is strongly dependent on the zero restrictions imposed on the Variance-Covariance matrix.

Figure 8.
Figure 8.

Impulse Response Following an Asset-Price Shock Based on Choleski Decomposition

Citation: IMF Staff Papers 2009, 003; 10.5089/9781589068209.024.A008

Note: The figure shows the results following vector autoregressions using each possible ordering of the six endogenous variables and the corresponding application of a Cholesky decompositions for identifying equity-price shocks. The top of the shaded area represents to maximum response coefficient among the different Choleski decompositions, whereas the lower end shows the minimum response at any time horizon.

As a next step, we use alternative variables and variable definitions to check how sensitive the findings are to such changes.6 First, we use the current account instead of the trade balance, taking into account the fact that the dynamics of both can be considerably different for some countries. Figure 9 shows the impulse responses of this specification for the United States and confirms the basic thrust of the benchmark results as the current account declines considerably after a positive asset-price shock. In fact, the reaction of the current account is somewhat stronger, as one would indeed expect, likely due to the decline not only of the trade balance but also of the income part of the current account.

Figure 9.
Figure 9.

United States: Impulse Response Following an Asset-Price Shock with Current Account

Citation: IMF Staff Papers 2009, 003; 10.5089/9781589068209.024.A008

Note: See note to Figure 1.

Second, we use relative equity-market capitalization,7 rather than equity prices, to define asset-price shocks. Figure 10 shows that the pattern of the impulse responses is unchanged for the United States (as well as for other industrialized countries, which are not shown for brevity reasons).

Figure 10.
Figure 10.

United States: Impulse Response Following an Asset-Price Shock with Equity-Market Capitalization

Citation: IMF Staff Papers 2009, 003; 10.5089/9781589068209.024.A008

Note: The figure shows the impulse responses of the six variables, for each of the countries in our sample of G-7 economies, to a 10 percent positive equity-market shock based on our Bayesian vector autoregression model. The shaded areas indicate the 16th and 84th percentiles of the posterior distribution. In contrast to the benchmark specification where we applied GDP weighting when constructing the rest of the world variables, we weight in this specification each individual country’s equity return by using equity-market capitalization.

Finally, we shorten the time sample to 1990–2007 in order to allow for the possibility that asset-price shocks may have become more important over time as countries have become more integrated financially and through trade. Figure 11 shows that the initial reaction of the trade balance is slightly larger and the response of private consumption significantly larger for the United States, lending some support to this conjecture.

Figure 11.
Figure 11.

United States: Impulse Response Following an Asset-Price Shock with Time Sample, 1990–2007

Citation: IMF Staff Papers 2009, 003; 10.5089/9781589068209.024.A008

Note: See note to Figure 1.

In summary, asset-price shocks appear to have a significant effect on the trade balance of countries, partly through wealth effects on domestic consumption and partly through an exchange rate channel that leads a real appreciation of the domestic currency. Moreover, there are substantial crosscountry differences in the effect of equity-price shocks, with the trade balance of the United States in particular exhibiting one of the largest reactions to asset-price shocks.

III. Conclusions

The paper has analyzed the effect of asset-price shocks on the current account. Its focus has been on the experience of the cross-section of G-7 industrialized countries. We have employed a Bayesian VAR with sign restrictions in order not only to motivate the identifying restrictions for asset-price shocks, but also to ensure that we can distinguish this type of shock from other shocks, such as to productivity, monetary policy, and government spending. The empirical evidence suggests that equity-price shocks indeed exert a significant effect on the trade balance of countries, partly through a wealth channel of private consumption and partly via an exchange rate channel.

One of the central findings of the paper is the substantial cross-country heterogeneity that we detect in the sensitivity of the trade balance to asset-price shocks. In particular the U.S. trade balance seems to be among the most sensitive to relative asset-price shocks, falling by 0.91 percentage points in response to a 10 percent increase in U.S. equity prices relative to the rest of the world. By contrast, other countries’ trade balances appear to be less responsive to asset-price shocks.

What explains this cross-country heterogeneity? Although the paper does not offer a systematic empirical analysis, the findings from the differences in the impulse responses suggest that both a channel via wealth effects as well as a real exchange rate channel appear to be at play in the transmission of asset-price shocks to the current account of countries. Specifically, the evidence suggests that differences in these channels are important for understanding the cross-country heterogeneity in the sensitivity of countries trade balances to equity-price shocks.

Many open questions remain and there are various future avenues for better understanding the importance of asset-price shocks, both domestically and globally. In particular against the background of the financial market turmoil of 2007–08, the role of monetary policy for asset prices remains unclear. Similarly, the focus of the present paper has been only on equity markets. Extending the analysis to housing markets seems particularly relevant in the current financial market context. Another important avenue is to extend the analysis to emerging markets, which are rapidly becoming ever more important players in the global economy and international financial markets. We leave these avenues for future research.

Appendix

Table A1.

Country Sample

article image
Table A2.

Data Definitions and Sources

article image
Note: IFS = IMF, International Financial Statistics; OECD = Organization for Economic Cooperation and Development.

References

  • Andersen, T., T. Bollerslev, F. Diebold, and C. Vega, 2007, “Real-time Price Discovery in Stock, Bond and Foreign Exchange Markets,Journal of International Economics, Vol. 73, pp. 25177.

    • Search Google Scholar
    • Export Citation
  • Beaudry, P., and F. Portier, 2006, “News, Stock Prices and Economic Fluctuations,American Economic Review, Vol. 96, No. 4, pp. 1293307.

    • Search Google Scholar
    • Export Citation
  • Beaudry, P., and F. Portier, 2007, “When Can Changes in Expectations Cause Business Cycle Fluctuations,Journal of Economic Theory, Vol. 135, No. 1, pp. 45877.

    • Search Google Scholar
    • Export Citation
  • Bernanke, B., 2005, “The Global Saving Glut and the U.S. Current Account Deficit,Sandridge Lecture, Virginia Association of Economics, March 10.

    • Search Google Scholar
    • Export Citation
  • Bertaut, C., 2002, Equity Prices, Household Wealth, and Consumption Growth in Foreign Industrial Countries: Wealth Effects in the 1990s, Federal Reserve Board International Finance Discussion Paper No. 724.

    • Search Google Scholar
    • Export Citation
  • Blanchard, O., F. Giavazzi, and F. Sa, 2005, “The U.S. Current Account and the Dollar,Brookings Papers on Economic Activity: 1 (Washington, The Brookings Institution).

    • Search Google Scholar
    • Export Citation
  • Bussiere, M., M. Fratzscher, and G. Müller, 2005, “Productivity Shocks, Budget Deficits and the Current Account,ECB Working Paper No. 509 (Frankfurt, European Central Bank).

    • Search Google Scholar
    • Export Citation
  • Caballero, R., E. Farhi, and P.-O. Gourinchas, 2006, An Equilibrium Model of ‘Global Imbalances’ and Low Interest Rates (unpublished; Berkeley, MIT and University of California).

    • Search Google Scholar
    • Export Citation
  • Canova, F., and G. de Nicolo, 2002, “Monetary Disturbances Matter for Business Cycle Fluctuations in G-7,Journal of Monetary Economics, Vol. 49, No. 6, pp. 113159.

    • Search Google Scholar
    • Export Citation
  • Case, K.E., J.M. Quigley, and R.J. Shiller, 2005, “Comparing Wealth Effects: The Stock Market Versus the Housing Market,B.E. Advances in Macroeconomic, Vol. 5, No. 11, pp. 132.

    • Search Google Scholar
    • Export Citation
  • Chinn, M., and H. Ito, 2007, “Current Account Balances, Financial Development and Institutions: Assaying the ‘World Saving Glut’,Journal of International Money and Finance, Vol. 26, No. 4, pp. 54669.

    • Search Google Scholar
    • Export Citation
  • Clarida, Richard, and Daniel Waldman, 2007, “Is Bad News About Inflation Good News for the Exchange Rate?NBER Working Paper No. 13010 (Cambridge, Massachusetts, National Bureau of Economic Research).

    • Search Google Scholar
    • Export Citation
  • Corsetti, G., L. Dedola, and S. Leduc, 2006, “Productivity, External Balance and Exchange Rates: Evidence on the Transmission Mechanism Among G-7 Countries,NBER Working Paper No. 12483 (Cambridge, Massachusetts, National Bureau of Economic Research).

    • Search Google Scholar
    • Export Citation
  • Engel, C., 1996, “The Forward Discount Anomaly and the Risk Premium: A Survey of Recent Evidence,Journal of Empirical Finance, Vol. 3, No. 2, pp. 12392.

    • Search Google Scholar
    • Export Citation
  • Engel, C., and J. Rogers, 2006, “The U.S. Current Account Deficit and the Expected Share of World Output,Journal of Monetary Economics, Vol. 53, No. 5, pp. 106393.

    • Search Google Scholar
    • Export Citation
  • Engel, C., and K. West, 2005, “Taylor Rules and the Deutschmark-Dollar Real Exchange Rate,Journal of Money, Credit and Banking, Vol. 38, No. 5, pp. 117594.

    • Search Google Scholar
    • Export Citation
  • Fogli, A., and F. Perri, 2006, “The Great Moderation and the U.S. External Imbalance,NBER Working Paper No. 12708 (Cambridge, Massachusetts, National Bureau of Economic Research).

    • Search Google Scholar
    • Export Citation
  • Fratzscher, M., 2008, “U.S. Shocks and Global Exchange Rate Configurations,Economic Policy, Vol. 23, No. 54, pp. 363409.

  • Fratzscher, M., L. Juvenal, and L. Sarno, 2007, “Asset Prices, Exchange Rates and the Current Account,ECB Working Paper No. 790 (Frankfurt, European Central Bank).

    • Search Google Scholar
    • Export Citation
  • Fratzscher, M., and R. Straub, 2008, “The Role of Asset Price Shocks, Financial Integration and Home Bias for Understanding Global Current Account Dynamics” (unpublished; European Central Bank, April).

    • Search Google Scholar
    • Export Citation
  • Gourinchas, P.-O., and H. Rey, 2007, “International Financial Adjustment,Journal of Political Economy, Vol. 115, No. 4, pp. 665703.

    • Search Google Scholar
    • Export Citation
  • Gruber, J., and S. Kamin, 2007, “Explaining the Global Pattern of Current Account Imbalances,Journal of International Money and Finance, Vol. 26, No. 4, pp. 50022.

    • Search Google Scholar
    • Export Citation
  • Hau, H., and H. Rey, 2006, “Exchange Rates, Equity Prices and Capital Flows,Review of Financial Studies, Vol. 19, No. 1, pp. 27317.

    • Search Google Scholar
    • Export Citation
  • Ju, J., and S.-J. Wei, 2006, “A Solution to Two Paradoxes of International Capital Flows” (unpublished; Columbia University, June).

  • Kraay, A., and J. Ventura, 2005, “The Dot-Com Bubble the Bush Deficits, and the U.S. Current Account,NBER Working Paper No. 11543 (Cambridge, Massachusetts, National Bureau of Economic Research).

    • Search Google Scholar
    • Export Citation
  • Krugman, P., 2007, “Will There Be a Dollar Crisis? Economic Policy, Vol. 22, No. 51, pp. 43567.

  • Lane, P., and G.M. Milesi-Ferretti, 2005, “A Global Perspective on External Positions,” in G-7 Current Account Imbalances: Sustainability and Adjustment, ed. by R.H. Clarida (Chicago, Chicago University Press).

    • Search Google Scholar
    • Export Citation
  • Lettau, M., and S. Ludvigson, 2004, “Understanding Trend and Cycle in Asset Values: Reevaluating the Wealth Effect on Consumption,American Economic Review, Vol. 94, No. 1, pp. 27699.

    • Search Google Scholar
    • Export Citation
  • Obstfeld, M., and K. Rogoff, 2005, “Global Current Account Imbalances and Exchange Rate Adjustments,” Brookings Papers on Economic Activity: 1 (Washington, The Brookings Institution).

    • Search Google Scholar
    • Export Citation
  • Palumbo, Rudd, and K. Whelan, 2006, “On the Relationships between Real Consumption, Income, and Wealth,Journal of Business and Economic Statistics, Vol. 24, No. 1, pp. 111.

    • Search Google Scholar
    • Export Citation
  • Peersman, G., 2003, What Caused the Early Millennium Slowdown? Evidence Based on Vector Autoregressions, CEPR Discussion Paper No. 4087 (London, Centre for Economic Policy Research).

    • Search Google Scholar
    • Export Citation
  • Peersman, G., and R. Straub, forthcoming, “Technology Shocks and Robust Sign Restrictions in a Euro Area SVAR,International Economic Review.

    • Search Google Scholar
    • Export Citation
  • Uhlig, H., 2005, “What are the Effects of Monetary Policy on Output? Results from an Agnostic Identification Procedure,Journal of Monetary Economics, Vol. 62, No. 4, pp. 38119.

    • Search Google Scholar
    • Export Citation
  • Ventura, J., 2001, “A Portfolio View of the U.S. Current Account Deficit,” Brookings Papers on Economic Activity: 1 (Washington, The Brookings Institution).

    • Search Google Scholar
    • Export Citation
*

Marcel Fratzscher is a division chief and Roland Straub a senior economist with the European Central Bank. The authors would like to thank the participants at the conference on “Current Account Sustainability in Major Advanced Economies” at the University of Wisconsin, Madison, and in particular our discussant, Ken West, for comments and discussion.

1

In the discussed model a fall in government spending, financed, for example, by lumpsum taxes for simplicity, is associated by a rise in private consumption and inflation, but a fall in aggregate output. As a result, the reaction of policy interest rates depends obviously on the monetary-policy rule. As argued in Fratzscher and Straub (2008), a standard Taylor rule implies a fall in interest rates, as the rise in inflation is relatively small, but the response of output is more pronounced for a wide range of structural parameters.

2

Appendix Table A1 lists the countries included.

3

Hau and Rey (2006) and Andersen and others (2007), for instance, show that there tends to be a negative correlation between equity returns and exchange rate returns in the data for several industrialized countries.

4

Appendix Table A2 lists the variables and their definitions and sources.

5

Moreover, this positive effect of asset prices on the exchange rate is not necessarily inconsistent with the literature that finds a negative correlation between equity returns and exchange rate movements (Hau and Rey, 2006; Andersen and others, 2007) as those correlations are unconditional ones and may stem from other types of shocks.

6

We show here only the corresponding results for the United States, though the conclusions on the robustness checks are qualitatively similar for other countries.

7

Relative equity-market capitalization is measured as the difference in the log domestic-market capitalization and the log rest-of-the-world market capitalization, both measured in U.S. dollars. Using market exchange rates or PPP exchange rates does not change the findings in a meaningful way. More precisely, although the magnitude of the impulse responses may change depending on the specification, the direction and dynamics is very similar across specification.

IMF Staff Papers, Volume 56, No. 3
Author: International Monetary Fund. Research Dept.
  • View in gallery

    United States: Impulse Response Following an Asset-Price Shock

  • View in gallery

    United Kingdom: Impulse Response Following an Asset-Price Shock

  • View in gallery

    Germany: Impulse Response Following an Asset-Price Shock

  • View in gallery

    France: Impulse Response Following an Asset-Price Shock

  • View in gallery

    Italy: Impulse Response Following an Asset-Price Shock

  • View in gallery

    Canada: Impulse Response Following an Asset-Price Shock

  • View in gallery

    Japan: Impulse Response Following an Asset-Price Shock

  • View in gallery

    Impulse Response Following an Asset-Price Shock Based on Choleski Decomposition

  • View in gallery

    United States: Impulse Response Following an Asset-Price Shock with Current Account

  • View in gallery

    United States: Impulse Response Following an Asset-Price Shock with Equity-Market Capitalization

  • View in gallery

    United States: Impulse Response Following an Asset-Price Shock with Time Sample, 1990–2007