EMU and Trade Revisited: Long‐Run Evidence Using Gravity Equations

Abstract

Abstract. In this article we present evidence of the long-run effect of the euro on exports for the twelve initial EMU countries for the period 1967-2008 from a double perspective. First, we pool all the bilateral combinations of export flows among the EMU countries in a panel cointegration gravity specification. Second, we estimate a gravity equation for each of the EMU-members vis-à-vis the other eleven partners. Whereas the joint gravity equation provides evidence on the aggregate effect of the euro on intra-European exports, by isolating the individual countries we assess which of them have obtained a larger benefit from the euro. Moreover, this strategy permits to check the robustness of the aggregate results and to find possible asymmetries. Finally, we repeat both the aggregated and individual analysis for the bilateral exports of EMU members to third countries. From an econometric point of view, we apply panel cointegration techniques based on factor models that account for cross-dependence and structural breaks.

Evaluating the effect of the introduction of the euro on tourist flows: A synthetic control approach

“…The EMU represents an interesting and relevant historical case that can be investigated in relation to the impact of a common currency on trade. Beginning with the seminal paper of Rose (), the extensive research on this topic finds that the effect of the euro on intra‐area trade has been positive, on average (Camarero et al., ; Chintrakarn, ; Micco et al., ). However, the measure of such an effect still represents a critical point.…”

“…Many steps have been taken in terms of applying the gravity equations to panel data and controlling for the endogeneity of EMU participation. Progress has also been made in accounting for the statistical characteristics of the time series of the trade flows, such as the presence of trends and cointegration (Bun & Klaassen, ; Camarero et al., , ), structural breaks (Camarero et al., ), time and spatial inertia (dynamic models, as De Nardis & Vicarelli, ; and spatial econometric techniques, as Kelejian, Tavlas, & Petroulas, ) or controlling for relevant omitted variables, such as inward and outward FDI (Camarero et al., ). All of these empirical refinements point to one main result: when controlling for the persistence of trade in time and space, the impact of the EMU is reduced or insignificant.…”

“…When restricting the analysis to the effect of the euro in the Eurozone, estimates are still positive but lower. For example, Micco, Stein, and Ordonez () indicated an impact that ranges between 8% and 16% in the short‐to‐medium term, while the long‐term estimates provided by Bun and Klaassen (, ), and Camarero, Gòmez, and Tamarit (, ) range from 3% to 40%.…”

“…To this purpose, they estimated the impacts by using gravity models (Frankel, ; Micco et al., ; Rose, ) and controlled for changes in variables, such as gross domestic product (GDP) per capita, relative prices and other relevant determinants of bilateral flows. When country or year‐specific effects of the euro were estimated through gravity equations, it emerged that the direct effect on intra‐European Monetary Union (EMU) flows has been generally positive; in some cases and periods, it was negative or insignificant (Camarero et al., ; Herwartz & Weber, ; Santana‐Gallego, Ledesma‐Rodrìguez, & Pérez‐Rodrìguez, ). The presence of heterogeneous effects between countries has raised issues related to the conclusiveness of the findings, as a negative sign of the partial/direct effect is difficult to justify.…”

“…The presence of heterogeneous effects between countries has raised issues related to the conclusiveness of the findings, as a negative sign of the partial/direct effect is difficult to justify. For this reason, these parametric estimates may be affected by the presence of relevant but omitted variables such as inward and outward foreign direct investment (FDI) (see Camarero, Gòmez, & Tamarit, ; Camarero et al., ). Furthermore, the use of parametric econometric techniques may raise specific issues when applied to the analysis of the effect of the euro; this is because the choice to join a currency area is not random: It involves a self‐selection bias (Alesina, Barro, & Tenreyro, ; Persson, ).…”