A joint CEPR/IGIER workshop on `Monetary Policy and Exchange Rates in Europe' was held in Milan on 24/25 November 1995, organized by Matthew Canzoneri (Georgetown University and CEPR), Francesco Giavazzi (IGIER, Università Bocconi and CEPR) and Axel Weber (Universität Bonn and CEPR). The workshop was hosted by the Innocenzo Gasparini Institute for Economic Research, Milan and supported by the European commission's Human Capital and Mobility programme. The papers presented can be generically divided into two groups. The first group discussed the economics of the exchange rate: interest rate differentials, foreign exchange risk premia and default risk premia. The second group was primarily concerned with the information value of financial asset prices and spreads for monetary policy.
The Uncovered Interest Parity (UIP) is the tool most used in analysing the relation between interest rates and exchange rates. If UIP holds, the differential between two yields on bonds with the same maturity but denominated in different currencies is the best predictor of the future change in the bilateral exchange rate between the same two currencies. There are two hypothesis on which the theory is based: first, agents do not request a foreign exchange risk premium; second, agents have rational expectations and the market is efficient. In `Interest Rate Parity and Foreign Exchange Risk Premia in the ERM' Juan Ayuso (Banco de España) and Fernando Restoy (European Monetary Institute and CEPR) evaluate the applicability of UIP to interest rate differentials in the Exchange Rate Mechanism (ERM) of the European Monetary System. In an international asset pricing model, they evaluate the hypothesis of market efficiency and measure the risk premia. Instead of focusing on a representative domestic investor endowed with the aggregate domestic wealth, in their model they consider an international investor whose endowment is the world wealth and whose relevant price index holds the world portfolio in equilibrium. With this assumption they can abandon the standard hypothesis of PPP (Purchasing Power Parity) implicitly assumed in the current literature on asset pricing model. They are able to reject the risk neutrality hypothesis: estimated risk premia between ERM currencies are low, while risk premia for currencies outside the ERM are larger, which suggests that foreign exchange risk between ERM currencies are diversifiable. They conclude that UIP is a reasonable approximation of the relationship between exchange rates and interest rates within the ERM.
Carlo Favero (Università Bocconi and CEPR) suggested to study the instability of the parameter of risk aversion with recursive methods and to use survey data on exchange rates. Willem Buiter (University of Cambridge and CEPR) disagreed with the use of the rate of return on world portfolio as an approximation of the rate of growth of the world consumption: global portfolio holders are not global consumers.
In `Price Level Targeting vs Inflation Targeting: A Free Lunch?' Lars Svensson (Institute for International Economic Studies, Stockholm University and CEPR) interprets `price stability' in two different ways: as price level stability and as low and stable inflation which allows base drift of the price level. In the literature, the choice between price-level targeting and inflation targeting involves a trade-off between low-frequency price level uncertainty on one hand and high -frequency inflation and output uncertainty on the other. In order to stabilize the price level under price level targeting, higher-than-average inflation must be succeeded by lower-than-average inflation. This should result in higher inflation and output variability than inflation targeting. Svensson departed from the previous literature by using a principle-agent approach and by considering a realistic degree of persistence in unemployment. This hypothesis was crucial to obtain the following innovative result: there is no trade-off between price level variability and inflation variability. Price level targeting is a `free lunch' since it results in lower inflation variability than inflation targeting.
Willem Buiter objected on the specification of the unemployment target level in the loss function of the central bank: it cannot be different from the optimal level of unemployment resulting from the structural equation. Lars Svensson replied that it was more difficult to agree on the natural rate of unemployment than choosing an unemployment target.
More than ever before, yield differentials between government bonds issued by different states reflect the feelings of market participants regarding countries' financial sustainability. In their provocative paper, Carlo Favero (Università Bocconi and CEPR), Francesco Giavazzi (IGIER, Università Bocconi and CEPR) and Luigi Spaventa (Università degli Studi di Roma, `La Sapienza' and CEPR) analyse `High Yields: The Spread on German Interest Rates'. They try to split interest rate differentials into two components: a default risk, which reflects the fears that something may go wrong with the service of the debt, and an exchange risk which includes both exchange rate depreciation and a foreign risk premium. They measure exchange risk in an innovative way using the fixed interest rates on swap contracts denominated in different currencies. They compute the default risk as the difference between the interest rate differential and the exchange risk. There is a close association between the Italian and the Spanish risk premia, while the Swedish exchange risk is the lowest of the three. The authors conclude that membership of the ERM is not sufficient to stabilize expectations. They find that the total differentials and the exchange premia move together with a common trend and that the exchange premia causes the total yield differentials. For Italy and Spain they also found evidence of a common `international' component driving the exchange risk and therefore the total yield differential: this component could be identified with shocks related to the Dollar-Deutsche Mark exchange rate, the ten-year Bund interest rate and the three-month German Euro-rate.
Stefan Gerlach (Bank for International Settlements and CEPR), analysing the relation between the default risk and the gross public debt and between the exchange risk and an index of government durability, suggested that longer lasting governments have a smaller default risk.
Unless one is willing to accept the notion that exchange rate expectations are systematically and significantly wrong, movements in exchange rates and interest rate differentials can only be reconciled by high, variable and serially correlated foreign risk premia. With a realistic parameter of risk aversion, standard capital asset pricing models are not able to reconcile exchange rate movements and interest rate differentials. In `Interest Rate Differentials and Foreign Exchange Risk Premia. A Model-based Empirical Analysis' Fabio Fornari, Carlo Monticelli and Oreste Tristani (Banca d'Italia) present a sensible estimate of risk aversion and find high and variable risk premia. They use a stochastic general equilibrium of two open economies with risk-averse optimising agents. The resulting equations for the interest rate differential are estimated over a sample starting in the early `80s for several couples of currencies. The model accounts for a large portion of the variability of interest rate differentials. Sizeable foreign risk premia emerge even for ERM currencies.
Frank Smets (Bank for International Settlements and CEPR) noted that in the current literature a simple model written by Lucas is used that accounts for the same results.
The existing analysis of the casual relationship among asset returns, real activity and inflation is incomplete in at least two respects: it is available primarily for the US and it concerns simple moments of domestic variables taken in isolation from the rest of the world. In `Stock Returns, Term Structure, Inflation and Real Activity: An International Perspective' Fabio Canova (Universitat Pompeu Fabra, Università di Modena and CEPR) and Gianni De Nicolò (Brandeis University) analyse the empirical interdependence among asset returns, real activity and inflation from a multi-country and international point of view. Three main questions are asked: first, whether and how shocks which affect nominal stock returns are transmitted to bond markets and affect real activity and inflation; second, whether and to what extent shocks to the slope of the term structure signal movements in real activity and in inflation; third how monetary surprises propagate to real activity. The results of the investigation suggest that the answers to these questions are country dependent and that interesting and important international asymmetries exist. Many of the empirical regularities found in the US are specific to that economy: the link between financial markets and real activity was absent in the other countries. In addition, in the US, the slope of the term structure has a mayor predictive power not only for domestic inflation, but also for foreign inflation. While the international analysis demonstrates that integration in financial and real markets is well under way, it is surprise movements in US financial and real variables that leads and predicts movements in foreign financial and real variables.
Axel Weber stressed the strong connection between the results and the restrictions imposed in the estimation of the model: different restrictions imposed on the relationship between shocks (monetary policy shocks, industrial production shocks and term structure shocks) would imply different conclusions.
E Philip Davis and Gabriel Fagon (European Monetary Institute) presented `Indicator Properties of Financial Spreads in the EU: Evidence from Aggregate Union Data'. They suggest that financial spreads may contain useful information about future inflation and output growth and, therefore, could potentially play a useful role in forecasting and assessing the appropriate stance of monetary policy in the Union. The two authors pay particular attention to three indicators: the yield curve, defined as the difference between domestic bond yields and short term market interest rates; the foreign bond yield differential, defined as the difference between the yield on domestic Government bonds and the yield on foreign Government bonds; and the reverse yield gap, measured by the difference between the yield on domestic Government bonds and the dividend yield on domestic equities. The analysis is carried out using estimated EU aggregates for output, prices, the above spreads as well as money and interest rates. The evidence found does not support an the use of weighted-aggregate financial spread variables in the formulation of the monetary policy in the Union. On the other hand the results, which reflect conditions during the sample period 1975-1992, do not rule out the possibilities that spread variables might acquire more predictive value in the future, by an increase in financial integration.
Krister Andersson (Sveriges Riksbank) commented the paper for its very interesting monetary policy implications. He did suggest that it is better to analyse information contained in the spreads for the different countries separately, since aggregation can bias the results.
A related paper `Some Issues in Extracting Information from Financial Asset Prices' was presented by Creon Butler (Bank of England). Butler offers some general thoughts on the kind of information central banks would ideally like to obtain from financial asset prices and contrasted this with the constraints on what they could actually expect to obtain. Information extracted from financial markets can be useful because it embodies more accurate and more up-to-date macroeconomic data than the information directly available to policy-makers. Moreover financial asset prices reflect market participants' expectations of future economic developments and they can be useful to explain market anomalies. However Butler points out some problems related to the process of extracting information from financial markets: the presence of incomplete markets, poor quality and quantity of data and the necessary hypothesis of risk neutrality which might bias the results. He presented a new technique for extracting the complete distribution of agents' asset price expectations for a particular date in the future using call option price. This technique can provide an indication of average future inflation expectations and hence monetary policy credibility. Another possible application might be in identifying speculative bubbles in financial prices: if investors are aware that the bubble may burst at some point in the future, this might be visible in the market's forward looking expectations.