بررسی اثرات معرفی یورو بر نوسانات بازارهای سهام اروپا

Have convergence of European economies and introduction of the euro produced some effects on European stock markets? Theory suggests that stabilization of fundamentals should decrease variance of stock returns for historically unstable stock markets. We test this proposition with daily data for the period January 1988–May 2000 and apply a three-regime Markov switching model for the variance-covariance matrix among several stock indices, including the UK and the US. The analysis shows that introduction of the euro, after an initial burst of volatility common to all European stock markets, has indeed stabilized the Spanish and Italian stock markets.

One of the most interesting experiments of capital market integration has taken place in Europe during the last 20 years. The euro has been introduced at the end of a long convergence process beginning in 1979 with the creation of the European Monetary System (EMS). In the few years before the introduction of the euro there has been a dramatic acceleration in the convergence process. Italy and Spain have decreased their annual inflation rate by several percentage points in a short time. Italy has been able to introduce a fiscal stabilization plan which has brought the primary surplus to about 5% of gross domestic product and has stopped the increase in the ratio between public debt and domestic product. The short-term nominal interest rate in Italy has decreased from 11% to 3% in only two years. Economic convergence has therefore left concrete signs in the macroeconomic structure of some European countries. The rules included in the 1998 Stabilization Pact are aimed at forcing even more convergence of fiscal deficits and reductions of public debt over the next years. This paper explores the consequences of the convergence process associated with the introduction of the euro on European stock markets. Has convergence of the European economies been associated with some form of convergence of European stock returns? Economic theory predicts that stock prices should reflect expectations of future dividends, interest rates and risk premia. Therefore the unconditional variance of stock returns should depend on variances and covariances of such fundamentals. It follows that both first and second moments of returns should be affected by the convergence process and the introduction of the euro to the extent that these phenomena affect fundamentals and expectations thereof. In particular, the introduction of the euro should be associated with a reduction of volatility of macroeconomic fundamentals of the historically unstable European economies like Spain and Italy, due to convergence of the stochastic process of fundamentals to that of the more stable northern European economies. The stabilization of some European stock markets could be studied with standard econometric methods only allowing for the existence of structural breaks. While we will test for a reduction in the unconditional variance of stock returns after the date of the introduction of the euro by means of an F test and of a test regarding the constant in a GARCH(1,1) model, we believe that such tests are insufficient due to the forward-looking nature of investors who may have set (in theory, should have set) prices with an eye towards the future introduction of the euro even before the end of 1998. It follows that econometricians have no knowledge of the exact date when investors have started to incorporate the euro factor in stock prices. A viable econometric alternative is provided by models where the variance switches across states. In such a model the introduction of the euro may be interpreted in terms of a stabilization of the stochastic processes of fundamentals of some European countries. A theoretical model for stock price determination suggests then that such a stabilization should be followed by a reduction in the variance of returns. Notice that simply allowing for the existence of different regimes would not be enough in terms of empirical analysis if we believed that the introduction of the euro has created a new economic regime. However we believe that this is not the case. We do not claim that the euro has allowed a transition of the Italian and Spanish economies towards a volatility regime similar to the one historically prevailing in Germany. Instead we assume that each single European stock market may move across three regimes characterized by different perceptions of the volatility of fundamentals. The stabilization of fundamentals brought about by the introduction of the euro should have increased the frequency of visiting medium and low volatility regimes for traditionally unstable European stock markets. In our empirical work we have identified three different volatility regimes, that have been denominated low, medium and high volatility regimes, in a sample composed of France, Germany, Spain, Italy, UK and USA. To anticipate the main empirical results, the structure of the estimated variance processes suggests that only weak similarities exist among the volatilities of the different stock markets. One can reject the restriction of perfect correlation in the timing of the switching across volatility regimes in the different markets. The best model is obtained when estimating each equation separately. We also find important changes in the volatilities across the three regimes and a general increase of correlations during periods of high volatility. There are clear signs of a large increase in volatility around the introduction of the euro and a stabilization of the traditionally unstable European stock markets in the early months of 1999. After this introduction the paper is organized as follows. In Section 2 we discuss theoretical models connecting stock returns and fundamentals. In Section 3 we estimate the effect of the introduction of the euro by means of a GARCH model with a dummy variable, argue that such an econometric methodology is insufficient from various points of view and introduce a Markovian volatility model to investigate the existence of states and the probability of the various markets to belong to the different states over time. Section 4 discusses the estimation results. Finally, we conclude in Section 5.

In this paper we have analyzed the variances and covariances of five major European markets and the US. The aim of the paper is to evaluate the changes in volatility brought about by the introduction of the euro. Theory suggests that the macroeconomic convergence process associated with the introduction of the euro should decrease the variance of the traditionally unstable European stock markets like Italy and Spain. For this purpose we have compared the dynamics of the conditional volatility process estimated by two models, a standard GARCH and a Markov switching three-regime model characterized by varying levels of variances and correlations. We have found that the implications of these two models are very different. The GARCH model finds no effect of the euro on the volatility of European stock markets, while the Markov model suggests the existence of some effects, lasting until the end of our sample period, in May 2000. In particular, we have found that a three regime model is a good characterization of the data and that the correlation coefficients are significantly different among the three regimes. There is a general increase in correlations in the high volatility regime. As far as the convergence process is concerned, there are signs of a stabilization of the Italian and Spanish stock markets. In particular the Italian stock market suffers from a large volatility for the first part of 1998, but then joins the low volatility regime after the introduction of the euro, with a decline in volatility which is larger than that of Germany and France. At the end of 1999 and in the second quarter of 2000 the volatility of the Italian stock market increased again, but this is not a proof that the euro failed to stabilize the Italian stock market. In those periods stock volatility increased all over the world, usually more than in the Italian stock market. The analysis suggests that the volatility of the Italian market relative to the volatility of other stock markets decreased after the beginning of 1999. Given that statistical specification tests suggest rejection of a one-state model for volatility, which is implicit in the GARCH model, we are inclined to stand with the three-regime model and conclude that the euro has had a long-lasting negative effect on the volatility of the Italian and Spanish stock markets. It is interesting to notice that the US has been largely unaffected by the introduction of the euro. These results seem to be consistent with the theory outlined in Section 2 of this paper, which suggests that the effect of the euro on stock markets is not due to a modification of the currency risk premium, but mainly to the stabilization of fundamentals for traditionally unstable European countries like Italy and Spain. Convergence of interest rates to German levels and strict fiscal discipline associated with the Maastricht Treaty and the Stability Pact have been taken into account by investors in the way predicted by the theory. Without the euro, periods of high international volatility in stock markets, e.g. March–April of 2000, would have affected in a much stronger way the stock markets of Italy and Spain. The results have various implications. From the point of view of stock market efficiency, European stock markets have indeed reacted in the way predicted by a fundamental valuation model in terms of volatilities, even though the large jumps in the levels of the stock prices around the introduction of the euro do not seem to be consistent with efficiency in strict sense: the event was expected and should have been discounted in stock prices before 1 January 1999. From the point of view of the European stock market as a whole, the euro seems to have brought a net benefit in terms of reduced volatility due to the discipline that economically strong European countries like Germany have been able to impose on economically weaker countries like Spain and Italy. The results seem to be relevant also for investors. On the one hand the successful modeling of volatility as a multi-regime Markov process suggests that the standard practice of estimating variance–covariance matrices on the basis of historical data, often going back five years into the past, is inappropriate, and may be particularly so after the introduction of the euro. On the other hand the sensitivity of stock markets volatility to general macroeconomic events suggests that macroeconomic variables are important risk factors for stock prices. It may therefore premature to declare the death of country diversification within Europe to support the introduction of sector diversification. Analysis of the events taking place in the Summer of 1998 suggests that future stock market volatility may be very sensitive to general macroeconomic conditions. Investors would therefore be better off in diversifying their stock holdings across European (and non-European) countries in order to avoid being locked in a stock market of a country which finds itself in trouble adapting to the new rules imposed by a fixed exchange rate (within Europe) and those enforcing reductions in the levels of public deficit and debt.