Are international economic and financial co-movements characterized by asymmetric co-integration?

Abstract

Purpose – The recent economic crises have attracted attention to the issue of international equity co-movements and correlations. Using data from 1980 to 2010, the authors examine the international co-movements of both real economic activity, as reflected in industrial production and the gross domestic product (GDP), and financial activity, as reflected in equity market returns. While classic symmetric co-integration tests do not reject the hypothesis of no co-integration, the authors find evidence of asymmetric co-integration in these three variables between the USA and the rest of the Group of Seven (G7) countries. The momentum threshold autoregressive (M-TAR) model captures the nature of the asymmetry most effectively and is the most applicable model for adjustment to long-term equilibrium. This model suggests that the path of adjustment to long-run equilibrium is somewhat different when the price differential is decreasing than when it is increasing. These findings imply that the benefits of asset diversification for investors with a long horizon might be limited in scope. Design/methodology/approach – This work is based on the theory of integrated time series. The authors use symmetric and asymmetric co-integration tests to market indices, as well as to monthly industrial production statistics and quarterly data about the GDP. In line with the financial economic literature, the authors select the GDP as a proxy that reflects the real economy and share prices to mirror the financial sector of the economy. Because no monthly data exist about GDP, the authors use instead the industrial production. Both variables cover the period from January 1980 to June 2010. Findings – The overall findings demonstrate that the USA and the rest of the G7 countries are not symmetrically co-integrated with respect to the GDP. Indeed, they are asymmetrically co-integrated. These findings may explain the additional important result that the majority of equity markets are also asymmetrically co-integrated with the USA. Research limitations/implications – The co-movements of the equity markets and real economic activity imply that the benefits of asset diversification for investors with a long horizon might be limited in scope. In the short run, however, portfolio diversification can be more beneficial due to the short-term fluctuations that may derive from the asymmetric correction process. Originality/value – Prior research on co-movements has focused mainly on studying the correlations among international equity markets by analyzing conditional correlations or using symmetric co-integration methods; the authors test the existence of a long-term relationship between economic variables with respect to the USA and the rest of the G7 countries using a threshold co-integration model.