IMF Survey: Why have you chosen to focus on regional unemployment within countries?
Mauro: Regional differences have costs of their own and give us hints as to the underlying reasons why overall unemployment rates are so high. Basically, we looked at this issue for three reasons. First, there is the likelihood that large and persistent regional disparities in unemployment may contribute to a higher national nonaccelerating inflation rate of unemployment (NAIRU). Inflationary pressures are more likely to arise in the initial stages of a recovery in a country with large differences in its regional unemployment rates. If you have 5 percent unemployment in some regions and 25 percent unemployment in others, inflationary pressures will start rising early in the 5 percent unemployment areas. Inflation will rise more slowly in a country with a geographically uniform unemployment rate of 15 percent.
Second, there are large social costs associated with regional unemployment disparities. When you have 15 percent unemployment throughout a country, it’s a situation in which I may be unemployed but my neighbors, family, and friends are likely to still be working. Imagine, instead, that you have 5 percent unemployment in some regions and 25 percent in others. In regions where the unemployment rate is 25 percent, many more people will be unemployed and families are much less likely to be able to pool their resources.
But third, and the biggest reason why we look at this, is that the persistence of large differences in unemployment rates within a country suggests that something is wrong with the adjustment mechanisms in that economy. If you have large parts of the country with high unemployment rates for a very long time, and people don’t migrate out of the high unemployment areas and firms don’t move into the high unemployment rate areas either, something is going wrong.
Prasad: We view regional disparities, especially persistent disparities, as a symptom of inappropriate or inefficient labor market adjustment. Aggregate shocks may hit an entire economy, but region- and sector-specific shocks have different effects on different regions and usually cause an adjustment to take place. A large body of literature indicates that in the United States, adjustment occurs through labor migration from a negative shock area to a positive one. Migration of capital is an alternative channel of adjustment.
IMF Survey: Your study examines regional unemployment rates in 14 European countries and the United States. Where were the disparities most severe and what do the disparities tell you about the possible reasons for high regional unemployment?
Mauro: Clearly, Italy and Spain stand out as the countries with the largest regional disparities. Italy, in particular, has striking regional differences. In Spain, Andalusia, in the agricultural south, has an unemployment rate of more than 30 percent. In Navarra, unemployment is 10 percent. But Spain has a higher nationwide unemployment rate. What makes Italy so striking is that the lower national unemployment rate coexists with an equally large regional difference—5 percent unemployment in the north and more than 22–23 percent in the south. In the United States, the regional differences are much smaller—ranging from 3 percent to 8 percent.
When we looked at countries with high regional differences—Italy and Spain in particular—we also looked for clues as to what institutions and policies seemed to be associated with these large regional disparities.
Prasad: We wanted to examine how wide the dispersion was in European countries as compared to the United States and how persistent these differences were in Europe relative to the United States. In the United States, for example, asymmetric shocks produce temporary increases in regional unemployment. The United States has had many episodes of this over the past 20 years, with fairly large increases in regional disparities in unemployment. But in Europe, these differentials have persisted over long periods of time, suggesting that this disequilibrium is probably perpetuated by labor market and other institutions. In the United States, adjustment mechanisms seem to allow these differentials to be mitigated fairly quickly.
IMF Survey: What is the impact of regional disparities on national economic performance? And which institutional arrangements or policies seem most closely linked to persistent regional unemployment disparities?
Prasad: Regional unemployment disparities tend to create welfare disparities. If governments attempt to reduce these welfare disparities through social transfers, they will introduce additional distortions into the economy. For instance, if the government chooses to tax residents in the more productive and low-unemployment areas and to distribute these resources to high-unemployment areas to make the welfare comparable, it will have further distortionary effects on labor supply and on other economic decisions, thereby reducing overall economic welfare.
Spilimbergo: On your second question, if you have region-specific shocks, you need flexible wages to accommodate local labor market needs. In a centralized bargaining system, the representatives or trade unions look at the overall unemployment rate. Since local wages are not flexible, you don’t give the right price incentives for an adjustment to take place.
Prasad: Typically, the region with low unemployment tends to take the lead in wage negotiations. This can exacerbate problems, however, if the initial shock that led to higher unemployment in one region resulted from lower productivity, but wages are then set on the basis of the region with lower unemployment and higher productivity. This will further widen the disparity between productivity and wages in the region hit by the negative shock. A centralized bargaining system that does not allow wages to adjust across the regions increases unemployment in the region that is doing poorly and worsens the imbalances.
IMF Survey: Does centralized bargaining lead to regional disparities or does it exacerbate disparities where they already exist?
Prasad: The literature suggests that completely centralized or completely decentralized wage bargaining leads to better outcomes than an intermediate degree of centralization. Given that EMU members will have fewer policy instruments to deal with shocks and given other institutional constraints, going halfway in terms of decentralized wage bargaining could be the worst of all possible worlds. This is currently the case in some countries, particularly Italy.
Mauro: Our approach is probably more forward looking. We see countries with large regional differences and are trying to see what can be done about these differences. When economists wonder why quantities don’t adjust—in this case, unemployment rates—we first look at why prices (wages) don’t adapt to ensure that those quantities return to the higher-employment equilibrium level. When we see persistent large differences in unemployment rates, we ask what can be done to the wage-setting mechanism to reduce those differences.
IMF Survey: Your study examines Italy and Spain in detail. Describe the workings of these economies and the reasons for the regional disparities there.
Prasad: In Italy, there is a clear segmentation of markets between the north and the south. In the north, the unemployment rate is about 5 percent, and, during the recent cyclical recovery, unemployment has been declining. In the south, the unemployment rate has been rising and is well over 20 percent. There is also a clear divide between the north and the south in labor force participation rates and in youth unemployment. People in the south also tend to have much longer spells of unemployment.
Apart from some minor leeway for firms to change wages at the plant level, Italy essentially has a nationalized bargaining system within each sector. A firm in the south has to pay its workers the same as workers in the north, though productivity levels in the south might be lower because of lower human capital or poorer infrastructure. The government has tried to encourage the movement of capital to the south by providing wage subsidy incentives, but this comes at a cost to the public exchequer, with overall distortionary effects on the economy.
Other institutional features that have also led to declines in migration from the south to the north include a rather inflexible housing market and an employment intermediation system dominated by public agencies. With regard to housing, for example, large taxes are imposed at various points in the sale and purchase of a house, and there isn’t much of a rental market. Also, public agencies have been very inefficient in providing employment intermediation for those who want to move from the south to the north. A larger role for private sector employment agencies could be very important in enhancing migration.
Recently, the government has tried to provide temporary wage flexibility in the south, but the effectiveness of these measures has been constrained, because investment decisions tend to be driven by longer-term considerations. And these measures haven’t been formulated in the context of an overall labor market reform strategy that would have addressed other rigidities, such as hiring and firing costs. The absence of a comprehensive reform agenda has raised concerns about the credibility of the reforms that have been attempted.
Spilimbergo: In Spain, the pattern of regional differences is not as clear as it is in Italy. Oversimplifying a bit, Spain has high unemployment in its agricultural south and a “rust belt” area of the north that has relied on traditional industries such as ship-building. But Spain’s policies have been similar to those in Italy—its policy responses do not favor the absorption of the local shocks and have included centralized wage bargaining, minimum wages, and unemployment benefits that discourage workers from seeking jobs in other regions. There, too, you have a relatively illiquid market for houses and rentals are infrequent.
Mauro: In Spain, as in Italy, we contrasted high- and low-unemployment areas. We found no differences in real wages or unit labor costs, which suggests there are no incentives for workers to migrate away from high-unemployment areas or for firms to migrate toward high-unemployment regions. In principle, Spain has a nationwide round of bargaining followed by regional, provincial, and in some cases firm-level bargaining. In practice, however, there is usually no downward adjustment at the regional or local level because the nationwide round is viewed as a minimum. We’ve also seen no differences in wage increases. Trade unions and others in Spain argue that the bargaining system is not as centralized as it seems, but we found no evidence of wages responding to local conditions.
IMF Survey: What institutional steps or policy changes could ameliorate these disparities?
Mauro: We have already argued that decentralized wage bargaining would allow wages to fall in areas of high unemployment and thus attract firms to move there. In theory, centralized bargaining systems have opt-out clauses, but the conditions currently attached to these clauses make it unlikely they will be used. Less stringent opt-out clauses might be one means of decentralizing wage bargaining.
I don’t see minimum wages or unemployment benefits as a major cause of regional disparities. Certainly, though, greater flexibility in the housing market could help. In Spain, for example, rental contracts are rarely for less than five years.
An additional step is to ensure that the level of education is adequate for workers in all parts of the country. Of course, if the labor force in areas of high unemployment is highly skilled, firms are more likely to move there. Therefore, from the perspective of the firm, a high level of education throughout the country is more helpful. In principle, both Italy and Spain aim to provide this; in practice, however, there are still some differences in the education levels.
Antonio and I, in a separate study, looked at workers’ migration and their education level. More highly educated workers are more likely to migrate in response to shocks. For less educated workers, the opportunity costs of remaining unemployed in their home region are relatively low. Without a great incentive to move, they prefer to stay and collect unemployment benefits.
Prasad: In Italy and Spain, a wage structure that is flexible and allows for sufficient differentiation of wages across sectors and regions based on productivity differentials seems to be the key for encouraging the optimal allocation of labor and for reducing these persistent differences in regional unemployment. In the south of Italy, even workers with high levels of human capital have difficulty finding a job. This suggests that employment intermediation mechanisms need to be improved, especially for younger workers making school-to-work transitions.
IMF Survey: What impact will entry into the European Economic and Monetary Union (EMU) have on regional unemployment disparities?
Spilimbergo: The EMU has benefits, but it will entail the loss of some policy instruments and could result in problems. One problem is that since nominal wages will be defined in terms of a common currency, there will be the temptation to equalize them across Europe. That would worsen regional disparities.
Mauro: The EMU can draw lessons from the experiences of Italy and Spain. Centralized bargaining in these countries had exacerbated regional differences. You can extrapolate that, if a European-wide centralized wage-bargaining system emerges, low-productivity countries might have much higher unemployment than the high-productivity countries.
Spilimbergo: Second, in the past, inflation has increased the flexibility of real wages. Since the euro area will strive to be a low-inflation zone, this mechanism can no longer be used to address regional unemployment differences. And, third, the exchange rate will no longer be available to address country-specific shocks and their asymmetric effects on different regions.
Prasad: Regional unemployment disparities are essentially a symptom, but a very important symptom, of overall labor market rigidities. Given the constraints on fiscal and monetary policy instruments under the EMU, it is more important than ever that labor market rigidities in European countries be dealt with effectively and quickly.
Copies of IMF Occasional Paper No. 177, Perspectives on Regional Unemployment in Europe, by Paolo Mauro, Eswar Prasad, and Antonio Spilimbergo, are available for $18.00 (academic rate: $15.00) each from IMF Publication Services. See page 180 for ordering information.