BODE EXPIATÓRIO

EMU, inequality, and finance
Giuseppe Bertola

Feelings of economic insecurity play an important role in slowing down the European integration process. In the early 2000s, Europeans were certainly troubled by the worldwide cyclical downswing, by competition from newly industrializing trading partners, by slow adoption of new technologies. But public opinion and voters in European Constitution referenda have found it easiest to blame them on the most novel and most apparently avoidable aspect of recent experience: “the euro” and, more generally, deeper and wider economic integration in the European Union (EU).
It is standard to claim that international economic integration, especially across similar countries, improves efficiency, and need not have any impact on inequality and risk. Adopting a single money (and delegating monetary policy to a non-political independent Central Bank) fosters stability, investment, growth – with no necessary impact on inequality. All this is true. But if this was all, it would be hard to explain why all countries have not yet dismantled all barriers to trade and factor mobility, and have not yet adopted one and only one planetary currency.
In fact, extending the reach of markets across countries’ borders offers new freedoms not only to exploit trade opportunities, but also to escape each country’s regulation and taxation. Thus, integration makes it more difficult for each country to implement policies meant to interfere with market outcomes. This may be a good thing if policy is shaped by rent-seeking political interactions: then, economic integration improves efficiency not only directly, but also by fostering policy reform. Again, theory predicts positive production effects of economic integration.
But many government policies do not simply waste and redistribute output. In every country, social policy offers insurance against job-market and other life-shaping risks that markets are poorly equipped to deal with. Fears of a race to the bottom in social insurance, just at the same time as economic integration introduces new sources of risk in each country’s labour market, are an important obstacle to full liberalization of international markets. If integration kills social policy, and financial markets remain imperfect and incomplete, increasing inequality and insecurity may well more than offset aggregate efficiency gains.

What do the data say about the inequality and social policy impact of EMU? The figure illustrates the evolution within-country inequality for each of the EU15 countries, for the EU15 aggregate, and for the Eurozone 12 aggregate. The points plotted are ratios to country-specific period averages of yearly quintile ratios, published by Eurostat and defined as the share of disposable income adjusted for household size accruing to the richest 20% of the population, divided by that accruing to the poorest 20%; observations larger than unity indicate that, in the year considered, inequality was higher than its 1995-2005 average for the country or aggregate considered (the average inequality levels are very heterogeneous across countries, which is very interesting but need not concern us here).
These statistics, like all statistics, are unavoidably imprecise measures of very complex phenomena. But it is interesting to see that just as the Eurozone countries began to enjoy full and irreversible economic integration, inequality increased very sharply in the EU15 and more sharply in the 12 Eurozone countries, and that a similar U-shaped path is followed by many individual countries. A large variety of factors is relevant to these and other developments. But it is possible to use simple statistical techniques to try and detect in noisy data the relationship (if any) of measured inequality to arguably more precisely measured economic variables, such as income and unemployment, and to EMU.
If EU15 countries are sufficiently similar and similarly influenced by other events, it is possible to attribute to EMU what appears difference across “ins” and “outs” before and after EMU, and assess the statistical significance of such “differences in differences”. Detailed results may be found in a paper I wrote. EMU does appear to improve economic performance (both in terms of per capita income and in terms of unemployment) and the intensity of international transactions (especially as regards foreign direct investment flows). But it also appears to be associated with higher inequality, and with lower social spending. In fact, inequality variation associated with EMU is fully accounted for by changes in social policy expenditure (excluding pensions) as a share of GDP, and in GDP and unemployment (both of which are of course likely to be influenced by integration as policies, as well as by global cyclical and technological development).
This evidence is intriguingly consistent with theoretical mechanisms. What is ambiguous in theory (the inequality impact of integration per se, controlling for policy factors) is statistically indistinguishable from zero in the data. And what is unambiguously predicted by theory is confirmed by the data: integration of markets should improve efficiency both directly and by making it more difficult for policy to interfere with markets, and in the data it does increase GDP, decrease unemployment – and increase inequality. In data, economic integration’s inequality effects are mediated by (comparatively, in comparison to pre-EMU and non-EMU) less generous social policy, and some of the apparent increase in country output may reflect smaller inefficiency losses from redistribution’s effects on effort incentives.
Whether such developments should be viewed as good news depends on the side of redistribution budgets one finds himself on, and on whether one views redistribution as a suitable or a misguided tool for pursuing goals that markets should in principle but might in practice fail to achieve. Financial market development can indeed fulfil some of the needs addressed by social policy in theory, and another recent paper I wrote documents its negative association with lower social spending in cross-country data. When governments cannot smooth income shocks, demand for insurance and savings certainly increases.
But can the supply side of many Eurozone countries’ sclerotic and uncompetitive banking and finance sectors accommodate this demand and reduce the welfare impact of economic insecurity? In the data, indicators such as the ratio to GDP of bonds, credit, and stock market valuation do not appear to be any higher in EMU countries, after adoption of the euro, than in the comparison group. This is worrisome, may justify many European citizens’ distrust of ‘the euro’, and should induce Eurozone governments to build the supervisory and antitrust infrastructures needed for markets to supply transparent, inexpensive, and efficient financial instruments.