Reducing Spanish unemployment under the EMU*

by Olivier J. Blanchard Juan F. Jimeno

Enero, 1999

* Paper presented at the conference El euro y sus repercusiones sobre la economía española, San Sebastián, 26 and 27 November 1998. We are grateful to Samuel Bentolila, Ramón Caminal, Juan J. Dolado, Jordi Galí, Jaume Ventura, and our discussants, Javier Andrés and Gilles Saint-Paul, for helpful comments.

** MIT.

*** Universidad de Alcalá and FEDEA.

Los Documentos de trabajo se distribuyen gratuitamente a las Universidades e Instituciones de Investigación que lo solicitan. No obstante están disponibles en texto completo a través de Internet: http://www.fedea.es/hojas/publicaciones.html#Documentos de Trabajo

These Working Documents are distributed free of charge to University Department and other Research Centres. They are also available through Internet: http://www.fedea.es/hojas/publicaciones.html#Documentos de Trabajo

Olivier J. Blanchard and Juan F. Jimeno

January 1999

1 Introduction

Spain enters the EMU with an unemployment rate roughly ten percentage points higher than the Euro average—19% versus 9%. Can it reasonably hope to eliminate this differential and join its Euro partners in further lowering unemployment? If so, how long will it take? And what will it take?

We first visited these questions in 1995 (see CEPR 1995). At the time, we argued that a two-handed approach, with reforms in the labor market on the supply side, and the active use of monetary policy on the demand side could reduce unemployment to 5% within a decade. Since then, two important developments have taken place.

First, the Spanish economy has grown at close to 3% per year, and the unemployment rate has fallen by 4 points, from 23% in 1995 to 19% in 1998. This has given hope to those who think that unemployment can indeed be lowered at a steady pace.

Second, Spain has just gone on the Euro. The macroeconomic benefits of the relative decrease in the Spanish interest rate, coming from the commitment to the Euro and the implied decrease in the risk premium, are now largely in the past. Spain will now have the same interest rate as its Euro-partners. And the margin of maneuver on the other macroeconomic instrument, fiscal policy, will remain very limited. Spain will have to decrease its unemployment rate, with one of its two hands tied behind its back.

Under these conditions, will Spain be able to continue its steady reduction in unemployment? In this paper, we address this question by performing an accounting exercise. We derive the macroeconomic path required to reduce the unemployment rate to 5% by 2005, seven years from now. Starting from the target path of unemployment, we work back to the required growth rate of output, and then to the composition of output to maintain balanced growth and balanced trade. From this exercise, we conclude that Spain faces a tough challenge:

*Paper presented at the conference El euro y sus repercusiones sobre la economía española, San Sebastián, 26 and 27 November 1998. We are grateful to Samuel Bentolila, Ramón Caminal, Juan J. Dolado, Jordi Galf, Jaumie Ventura, and our discussants, Javier Andrés and Gilles Saint-Paul, for helpful comments.

MIT.

Universidad de Alcalá and FEDEA.

On the one hand, Spain has to grow much faster than its Euro-partners. The reason is obvious: it has to decrease its unemployment rate much more than they do. We estimate that Spain will have to sustain a growth rate of about 4.5% for the next seven years.

On the other hand, Spain has to achieve lower inflation than its Euro-partners, perhaps even deflation. The reason here is the need for external balance. Growth in excess of its Euro-partners requires steady real depreciation to maintain a balanced current account. Given the common currency, real depreciation can only be achieved through lower inflation. We estimate that, absent shifts in exports and imports, Spain will have to achieve an inflation rate roughly 3.5% below that of its Euro partners. This may well mean deflation.

Behind these conclusions lie assumptions about productivity, participation rates, export and import relations, the need for current account equilibrium, which can all be challenged. The path may not require deflation. Even if it does, achieving high growth and deflation may not be impossible. One can conceive of sufficient strength in aggregate demand and sufficiently fast reforms in the labor market that the decrease in unemployment is consistent with deflation. But the least which can be said is that it may not be easy. There lies the challenge facing Spain under EMU.

The structure of the paper is the following. In Section 2 we compute the rate of output growth associated with our target path for unemployment. In Section 3, we look at the implications of internal balance for the evolution of investment, consumption and government expenditures. In Section 4, we compute the evolution of the real exchange rate needed to maintain a balanced current account along the path. This in turn translates into the inflation differential Spain has to maintain vis a vis its Euro-partners. In Section 5, we turn to the labor market and derive the evolution of wages and equilibrium unemployment consistent with the target path for unemployment and current account equilibrium. In Section 6 we further discuss two of the assumptions underlying our basic scenario, the stability of export and import equations, and the relevance of the external constraint. Section 7 states our conclusions.

2 Unemployment decline and output growth

We take as a target path for unemployment a path where the unemployment rate falls from 19% in 1998 to 5% in 2005. This path implies a reduction of 2 points in the unemployment rate each year for 7 years, or about 300,000 people a year—a total of 2.1 million. On the one hand, 7 years is a long time. On the other hand, a 14% reduction in unemployment is a large reduction, and this target path may be seen a priori as too ambitious. But “too ambitious” should never be assessed a priori. The purpose of our paper and the exercise we carry out is precisely to find out what such a path would involve, and then to assess whether and why it may indeed be too ambitious.

Our goal in this first section is to derive the rate of output growth required to achieve this target path for unemployment. We go at it in two alternative ways. The first is based on the historical relation between output growth and the change in unemployment rate, Okun's law. The second is based on forecasts of productivity, labor force growth and participation rates.

2.1 The implied path for output: A first pass

The historical relation between the change in the unemployment rate and the growth rate of output is known as Okun's law. Let u be the unemployment rate, and and being GDP growth and potential GDP growth, respectively. Okun's law takes the form:

\[\Delta u _ {t} = - \alpha (g _ {y t} - \overline {{{{g _ {y t}}}}})\]

The relation is typically estimated using quarterly or annual changes in unemployment and growth rates. The results of estimation using annual data are given in Table 1a. For our purposes however —namely forecasting the effects of sustained growth in excess of potential growth for many years— it is better to use changes in unemployment and growth rates over longer time spans. For this reason, Table 1b gives the results of estimation using (overlapping) three-year changes in unemployment ( ) and growth rates ( ). In each of the two tables, the results are given for two different specifications, and three samples.

The two specifications capture two different assumptions about . In the first, potential GDP growth is assumed to be constant throughout the period. The second allows potential GDP growth to vary over time. Potential GDP growth for 1979-1985, 1986-1991, and 1992-1997 is constructed as the sum of average labor force growth and the average rate of technological progress over each subperiod. The numbers for labor force growth are 0.5%, 1.4%, and 1.1% respectively. The numbers for the rate of technological progress (computed as the Solow residual divided by the labor share) are 2.4%, 1.4% and 1.1%. These imply numbers for potential GDP growth of 2.9%, 2.8% and 2.2% respectively.

The three samples all start in 1978 but end in 1991, 1994 and 1997 respectively. The purpose is to show how adding recent years affects the estimated value of .

The estimated values of are between 0.8 and 0.97 in Table 1a, and between 0.9 and 1.05 in Table 1b. Considering the emphasis given to the importance of firing restrictions in Spain, one would have expected employment smoothing and thus a lower value of , especially in Table 1a. The results appear robust however, and we speculate that a partial explanation is the incidence of fixed-term employment contracts (about 10% of employees up to the mid-eighties and about 30%, on average, since 1987). The fact that the coefficient increases as the 1990s are included in the sample gives some support for this hypothesis.

In our target path for unemployment, the unemployment rate goes down by 2 points per year. Taking the value of to be roughly equal to one, this requires a GDP growth rate of two percentage points above potential GDP growth. Taking the average potential growth during the sample period, 2.5%, this gives a required GDP growth rate of about 4.5%.

2.2 A second pass. I. From unemployment to employment

A more ambitious way to derive the path for output under our target path for unemployment is to look inside the Okun's law blackbox, and go from the path of unemployment to the path of employment to the path of output.

The relation between the evolution of unemployment and the evolution of employment depends on both the evolution of the population of working age, and the participation rate. Table 2 gives forecasts of both population and participation rates in 2005, by age and sex.

The forecasts of working age population are based on Fernández-Cordón (1998). These forecasts assume that immigration flows remain constant. The striking feature of these numbers is the modest increase in population in working age over this period, by about 1,1 million people over the period, or about 140,000 people a year. This reflects a continuous fall in the Spanish fertility rate which has decreased from 2.8 in 1975 to 2.2 in 1980 and to 1.4 in 1990 (Spain now has the second lowest fertility rate in Europe after Italy.)

The forecasts of participation rates are ours. We simply extrapolate past trends for each age and sex category. Note that this ignores any increase in participation rates in response to the decrease in unemployment along the target path. Based on existing trends, participation rates are forecast to decline for all categories (excluding females in the 25-54 age group). But the change in the composition of the labor force, toward groups with high participation rates still leads to an overall increase in participation, from 61.2% to 62.5%—reflecting an increasing participation rate for females as a whole, from 48.0% to 49.1%, and a more or less constant participation rate, from 76.0% to 75.9%, for males as a whole.

The forecasts of population and participation rates imply in turn a growth in the labor force of about 850,000 people from 1997 to 2005, or about 110,000 people a year. Thus to decrease unemployment by an average 300,000 people a year, employment must increase at about 410,000 workers a year. This in turn represents an annual rate of growth of employment of 2.7%.

This implies an annual rate of growth for working age population of about 0.5%. We may be overestimating this increase in the working age population for two reasons. First, other sources, for instance, the Spanish Statistical Office, give a lower forecast of the working age population in 2005 (around 26,5 millions). But, as explained in the following footnote, taking the evolution of working age population according to the Spanish Statistical Office does not substantially change the estimation of the labor force by the year 2005. Second, our starting number for population in 1997, which we take from the Labor Force Survey for that year, is lower than the projections by both Fernández-Cordón (1998) and by the Spanish Statistical Office for that year.

Had we taken the forecasts of the working age population by the Spanish Statistical Office, the labor force in 2005 would have been only 100 thousands lower, since the composition effect which raises the participation rate is larger in this case.

2.3 A second pass. II. From employment to output

To go from employment growth to output growth requires a forecast of labor productivity. Labor productivity depends both on the rate of technological progress, and the evolution of the capital-output ratio. There lies an important conceptual issue. Should we think of the growth path required to decrease unemployment as a balanced growth path, with capital, output and labor in efficiency units all growing at the same rate? Or should we expect unbalanced growth, and substantial changes in the capital-output ratio?

A rough answer to the question can be given by looking at the recent evolutions of the capital-output ratio and of the rate of return to capital. Figure 1 shows the evolution of the capital-output ratio (computed from the stock of capital given by Mas, Pérez and Uriel, 1996 ), the rate of return of capital, and the capital share (from the OECD business sector data base) since 1970.

The capital-output ratio, which had increased a lot in the 1970s and early 1980s, has roughly stabilized since at a level around 2.9. It has been very stable since the mid 1990s. The initial increase surely reflected the large increase in the cost of labor in the 1970s and early 1980s. The stabilization reflects the fact that things have improved since then, and that the rate of return on capital has increased, from around 12% in the early 1980s to around 17% in the 1990s.

Evidence from Tobin's q and of the user cost of capital also support the conclusion that the return on capital is now more than adequate to maintain the capital-output ratio. As shown in Figure 2, Tobin's q (taken from Escribá and Ruiz-Tamarit, 1992) increased very quickly in the second half of the 80s, and stood in 1992 (the last year for which it is available) at 1.2/1.3. And, since 1992, the user cost (taken from the database of the regional version of MOISSES, the macroeconomic model of the Spanish Ministry of Economics —see Dabán et al., 1998) has declined, making the comparison between the rate of return and the cost of capital even more favorable.

We read this evidence to say that, at the current factor prices, firms are willing to maintain the current capital-output ratio, and that it is reasonable to work under the assumption of balanced growth in the future.

The next step is thus to derive what this balanced growth rate may be for Spain at this point. A bit of algebra is useful here. Decompose the growth of output as :

\[g _ {y} = (1 - \alpha) g _ {n} + \alpha g _ {k} + \theta\]

where , and are the growth rates of output, employment and capital respectively, is the share of capital and is tfp growth—or equivalently the

The data on the stock of capital are provisional for the years 1993-94. We have computed the data for the 1995-97 period by using data from the National Accounts information on gross fixed investment and using a depreciation rate of 5%.

5%

For a more formal analysis of these mechanisms, and an interpretation of the evidence not only in Spain but in European countries in general, see Blanchard (1998).

Solow residual. Then under the assumption of balanced growth , the rate of growth of labor productivity is given by:

\[g _ {y} - g _ {n} = \frac {\theta}{1 - \alpha}\]

To construct the right hand side-call it the rate of technological progress, one must construct the Solow residual and divide it by the share of labor. We construct the Solow residual in two ways. First, by using conventional measures of the number of workers employed, and the capital stock. Second, by correcting employment for the number of hours worked, and the capital stock for capacity utilization. The two series are shown in Figure 3. Both measures show a noticeable deceleration of the rate of technological progress. Since 1992, the average rate has been equal to about 1.1% for the unadjusted series, 1% for the adjusted series. This low rate surely reflects in part the well known sensitivity of the constructed Solow residual to cyclical conditions. In the expansion of the 1980s (1986-1990) for example, the average rate of technological progress was 1.7% (1.8% for the adjusted series). This may be a better estimate of the rate of technological progress required for our computations.

Assume therefore a range of 1.5 to for the average rate of technological progress for the next seven years. Combining this range with the required rate of employment growth derived in the previous subsection gives a required rate of output growth of 4.2 to , just a bit below the obtained using the rough Okun's law computation. We shall use the round number in what follows.

2.4 Working hours and the 35-hour week

By focusing on employment measured in terms of workers rather than total hours worked, our computations implicitly have assumed that, whatever the evolution of hours worked per worker had been in the past, it would continue over the next seven years. Given the current discussion of potentially large reductions in the workweek, we need to look at this assumption more closely.

In Spain, as in most other European countries, working hours have decreased over time, although the decrease has been small in the 1990s: a 5% decline since 1986 according to the Earnings Survey, a 2% decline since 1991 according to the Earnings Survey and the Employment Situation Survey, and only a 0.8% decline since 1990 according to the Labor Force Survey (This decline has been due to the reduction of the working hours of full time employees and, to a minor extent, to a modest an increase of part-time employment—from 5% in 1987 to 8% today.)

Standard annual hours for full-time manufacturing workers are now about 1,770, or 38 hours per week over 46.5 working weeks a year. Following the lead of France, there has been some talk of moving to a 35 hours week. This raises the question: If such a measure was to be implemented, how would this affect our computation of the required growth rate?

In this respect, Spain is roughly at the middle of the ranking of EU countries (see Hunt, 1998).

A first computation assumes that hours worked and bodies are perfect substitutes and that such a decrease in working hours would have no effect on productivity. It then goes as follows. A reduction from 38 to 35 hours implies a decrease of 8.2% of the number of hours worked, or 1.2% per year over the next 7 years. This would represent about a decline of 1.0% above trend (recall from above that the annual decline in the 1990s has been around 0.1 to 0.2%). For given output, this would in turn allow for an increase in the number of workers of 1.0% per year. To achieve the same decrease in unemployment, output growth would need to be only 4.5%-1.0% = 3.5%.

But this computation is misleading. Research done in the context of the passage of the French 35-hour law suggests that the effects on employment and thus on the required output growth would be substantially smaller. First, firms would increase overtime. Second, hours and bodies are not perfect substitutes. Third, there would be some gains in productivity. Best estimates—which are not much more than guesses—are that a 1% reduction in hours would come with a gain of productivity of 0.3%. Based on French estimates and guesses, a decrease in 1.0% of hours per year above trend might translate in a decrease in the required output growth of about one third of this decrease, i.e. 0.3%. (This assumes no increase in the cost of labor, which would then lead to further effects on output and employment. Research by Hunt (1998) suggests that the reduction in standard working hours is unlikely to be neutral with respect to labor demand, labor supply, and wages.)

When we started this paper, we thought that the required rate of growth of output would be turn out to be much higher than the estimate we have just derived—around 4.5%. In a deep sense, that Spain can only achieve 4.5% output growth even if it reduces unemployment by 2 percentage points a year is bad news: it reflects in particular a very low rate of technological progress. But in the slightly Ubuesque world in which having much higher growth than your Euro-partners creates a number of macroeconomic problems, this is good news indeed. Given a forecast of about 2.5% growth for Euroland, Spain's growth differential will have to be only 2%.

3 Implications of internal and external balance

Assume that output increases at the required rate to meet the unemployment target described above. Let's now ask: what must be the behavior of consumption, investment, government spending, exports and imports such that both internal balance and current account equilibrium are achieved? Internal balance, namely the condition that demand equals supply, and therefore also grows at 4.5% a year better be satisfied. Requiring a balanced current account requires more justification: clearly a country can run a current account deficit for a while, if not forever. However, it seems natural to look first at the path along which the current account is balanced, and this is what we do here. We consider the implications of relaxing this assumption and allowing for delayed adjustment in Section 6.

See the report of the Conseil d'Analyse Economique by Dominique Taddei, Documentation Francaise, 1998, for further discussion

We shall focus mainly on the implications of external balance. But before we do so, we briefly look at the implications of internal and external balance for consumption, investment and government spending. If the current account is balanced, then the sum of consumption, investment and government spending must grow at the same rate as output. Let's see what this implies. (As a reference, Table 3 gives the composition of output and the evolution of the current account in the 1990s)

3.1 The path of investment

Our assumption that Spain decreases employment along a balanced growth path implies that the capital-output ratio must remain at its current level along the path. This in turn implies that the investment to output ratio—equivalently, the share of GDP devoted to investment—is given by:

\[{\frac {I}{Y}} = {\frac {K}{Y}} (g _ {y} + \delta)\]

where is the depreciation rate. From Figure 1 the capital-output ratio stands at about 2.9. The depreciation rate of capital is around 5%. If output grows at 4.5%, this implies that the share of GDP devoted to investment must be equal to 27.6%. This compares to a current share of 20.9%. Balanced growth will therefore require a substantial increase in the share of investment in GDP.

3.2 The path of consumption and government spending

If gross investment is equal to 27.6% of GDP and the current account is balanced, then the sum of private consumption and government expenditure will have to equal 72.4% of GDP, about 6 points less than the number for 1997. This clearly will be difficult. First, in a context of high growth and lower interest rates it seems more likely that consumption grows by more than GDP rather than by less. Secondly, government expenditure stands at 16.5%, close to the average of EU countries. After several years of fiscal consolidation (with reductions in the proportion of public employment and in the relative pay of public employees), it is difficult to foresee significant further major reductions of government expenditures.

4 External balance and the real exchange rate: A first pass

If Spain grows faster than its neighbors for seven years, and if the export and import functions remain stable, then, at a given real exchange rate, there will be a steady worsening of the current account. Put another way, if the current account is to remain balanced, there will have to be a steady real depreciation. We proceed in two steps. We first derive the evolution of the current account under the assumption of a constant real exchange rate. We then derive the magnitude of the real depreciation required to maintain current account balance.

Think of Spain as producing a domestic good, and buying a foreign good. The trade balance is then given by

\[N X = X (Y ^ {*}, \epsilon) - \epsilon Q (Y, \epsilon)\]

where is the real exchange rate, X is exports, Q is imports, Y is domestic demand, and is foreign demand. Given the growth paths for domestic and foreign demand and the corresponding elasticities, we can solve out for the path of the current account, or of the real exchange rate required to maintain balance.

At a given real exchange rate, the evolution of the trade balance is given by:

\[\frac {d N X}{X} \equiv \eta_ {X Y ^ {*}} \frac {d Y ^ {*}}{Y ^ {*}} - \eta_ {Q Y} \frac {d Y}{Y}\]

We saw that Spain has to grow at 4.5% per year. Let's assume that its trading partners will be growing at 2.5%. What income elasticities should we use to compute the evolution of the current account?

There is a wide variety of papers estimating export and import equations for the Spanish economy (among the most recent ones are Buisán and Gordo, 1994, Bajo and Montero, 1995, Doménech and Taguas, 1997, García and Gordo, 1998, Mauleón and Sastre, 1994; for a survey and evaluation of the estimation performed in some of the previous papers, see Escribano, forthcoming). The papers differ in the choice of the dependent variable (for exports: log exports excluding energy products, log exports of goods and services in real terms, both in levels and in differences; for imports: the rate of growth of imports excluding energy goods, the rate of growth of import of goods and services, the log of import of goods and services, excluding tourism), on the proxies for domestic demand (log domestic demand, log GDP) and foreign demand (log of trade among developed countries weighted and unweighted by trade shares with Spain), and on the proxies for relative prices. They also differ in the set of regressors in both equations (some include the capacity utilization rate and taxes on imports -Doménech and Taguas, 1997, some include foreign direct investment -Bajo and Montero, 1995) and in other restrictions (some impose a unit long-run elasticity of imports with respect to domestic demand, some impose interactions between the export and the import equations). Thus, not surprisingly, they generate a large range of estimated elasticities. The estimated income elasticities of exports are between 0.93 and 2.56, of imports, between 0.68 and 2.1, both with a mid-point of estimated income elasticities of 1.4. The estimated price elasticities of exports are between -0.38 and -1.14, with a midpoint of -1.0; of imports, between -0.39 and -0.87, with a mid-point value of -0.6.

If we assume income elasticities to be both equal to 1.4, then at a given exchange rate, the trade balance will worsen by 1.4 times 2.0% or 2.8% of exports a year, or roughly 0.28 (the ratio of exports of goods and services to GDP for 1997) times 2.8% = 0.78% of GDP per year, adding up to a deficit of 5.5% after 7 years. If, instead, we assume unit income elasticities, the trade balance will worsen by 2% of exports or, equivalently 0.56% of GDP per year, adding up to a deficit of 3.9% after 7 years.

If we require instead that the current account remain balanced, then Spain will have to have steady real depreciation at the rate:

\[\frac {\eta_ {Q Y} \frac {d Y}{Y} - \eta_ {X Y ^ {*}} \frac {d Y ^ {*}}{Y ^ {*}}}{\eta_ {X \epsilon} - \eta_ {Q \epsilon} - 1}\]

where the and are the elasticities of exports and imports with respect to the real exchange rate. Using a value of 1 for the exports elasticity, and of -0.6 for imports, gives a rate of real depreciation of 2.8/.6 = 4.6 (if income elasticities are assumed equal to 1.4) to 2.0/.6 = 3.3 if the income elasticities are assumed equal to 1.0). We shall conservatively assume in what follows that the rate of real depreciation must be equal to 3.5%.

5 From the real exchange rate to inflation

The proportion of Spanish exports of goods to EU countries is above 80%, of services around 87%. The proportion of Spanish imports of goods from EU countries is around 75%, of services around 70%. It is therefore not too bad an assumption to think of Spanish trade as Euroland trade, and thus to assume that Spain now operates under a fixed nominal exchange rate.

Suppose that inflation in the Euroland is 2%, the current EU target for inflation. The implication of our calculations on the real depreciation required to keep the current account balanced is straightforward: The GDP deflator of Spain must decrease by 3.5%-2% = 1.5% per annum. With labor productivity growing at 1.0 to 1.5%, this means roughly constant nominal wages (real wage growth in terms of the consumption basket, assuming a weight of .25 for foreign goods, .75 for domestic goods, of 0.0 - .75*(-1.5) - .25*(2.0) = .625% a year).

The question is then whether Spain can achieve zero wage inflation in the face of rapidly declining unemployment. Surely, the answer is that it cannot do so without drastic reforms in the labor market. Absent reforms, decreasing unemployment is sure to lead to higher wage inflation, not constant nominal wages.

To explore this issue, we must look at the implications of the relation between the equilibrium (or “natural”) unemployment rate, the actual unemployment rate and the inflation rate in Spain. Previous empirical work on Spain suggests that this relation is well characterized by an extended Phillips curve relation of the form:

\[\pi = \pi_ {- 1} - \beta (u - u ^ {*}) - \gamma (u - u _ {- 1})\]

where is inflation, u is unemployment, and is equilibrium unemployment.

This equation differs from the basic Phillips curve in two important ways. First, it is very clear from the data that equilibrium unemployment is not constant, but has instead increased enormously since the early 1970s. Second, the data suggest a strong effect of the change in unemployment in addition to the deviation of unemployment from its equilibrium value; the effect is present in most countries, but seems stronger in Spain (see Dolado and Jimeno 1997). Estimating this relation requires specifying . We use three (all black box) approaches here: (i) a smooth trend (from a 4th-degree polynomial in time), (ii) the series provided by the OECD on the NAWRU, and (iii) the result of applying the Hodrick-Prescott filter (with a high bandwidth). Figures 4a and 4b plot the trend and the cyclical component of unemployment obtained with each of the three methods.

The estimation results are given in Table 4. We use quarterly data for the period 1978:3 to 1997:4. The dependent variable is the quarterly rate of change of the GDP deflator (excluding indirect taxes, and seasonally unadjusted). The sum of coefficients of past inflation is restricted to be one. The sum of the coefficients of the unemployment gap (current and three lags), is below -0.1 in all three specifications, and close to be significantly different from zero only at the 10% level. The coefficient on the change in the unemployment rate (with respect to the same quarter of the previous year) is around -0.2 and very significant.

Based on this estimation we take and We can then answer two questions:

- What would happen to inflation if unemployment decreased along the target path, but equilibrium unemployment, remained at its current level, around 18%? The estimated equation tells us that inflation would soon start to increase from its current level of 2.2%, stand at 5% in 2001, and 11% in 2005.

- What must be the path of equilibrium unemployment if unemployment is to decrease along its target path, and inflation goes from its current value of 2% (a conservative forecast for the change in the GDP deflator in 1998) to a negative -1.5% and remains there until 2005?

The mechanical answer is given in Figure 5, that also plots the path for inflation and unemployment rate (inflation falls to -1.5% in 1999 and remains at this level until 2005 to converge to the EU level afterwards, and unemployment steadily falls to 5% in 2005 to stay at that level). Going from inflation to deflation at the start requires a large gap between actual unemployment and equilibrium unemployment. Given the requirement that actual unemployment declines, this requires a very large decline in equilibrium unemployment for one year at the start. After the initial adjustment, equilibrium unemployment must remain five points below actual unemployment to avoid any inflationary pressure from the change of unemployment. It must obviously be equal to the target rate in 2006, namely 5%, after unemployment stabilises at 5%. The details of the simulation—and in particular the large initial decline and final increase—should be taken with a grain of salt; they would go away along smoother paths for inflation. But the basic message, namely that that a steady fall in unemployment without accelerating inflation can only be achieved with structural reforms and thus a sustained and large fall in equilibrium unemployment, should be taken seriously.

These values are consistent with previous results on the Spanish inflation-uncemployment trade-off. For instance, Dolado and López-Salido (1998) estimate a bivariate VAR of inflation and unemployment to find that the effects of unemployment on inflation are between 0 and -.3 (within the year) and between 0 and -.6 (in the long-run) depending on the identification scheme.

5.1 How to interpret the recent evidence on inflation and unemployment?

Can such a sustained decrease in equilibrium unemployment rate be achieved? The recent evidence on unemployment and inflation seems to suggest that the answer is yes. From 1994 to 1998, the unemployment rate has decreased by about 5%, or roughly 1% a year. And inflation, far from increasing, has also decreased by about 2%, or roughly 0.4% a year. This suggests that the equilibrium unemployment rate has already decreased substantially in the last five years.

Using the Phillips curve relation we estimated above, we can actually derive the evolution of equilibrium unemployment consistent with such an evolution for unemployment and inflation. Rewrite the equation as:

\[u - u ^ {*} = - \frac {1}{\beta} (\Delta \pi + \gamma \Delta u)\]

Using and , gives . With actual unemployment around in 1998, this implies that the equilibrium unemployment rate stands at roughly . Taken at face value, this computation implies that unemployment can be decreased down to before inflation starts increasing again.

This back-of-the-envelope computation comes with more than the usual warnings. In particular, the actual value of depends very much on the small value of , which is not estimated with great precision. Nevertheless, it suggests that equilibrium unemployment has declined, perhaps substantially, over the last five years.

Where does this decline come from? Some argue that the labor market reforms in 1994 and 1997 have substantially changed employment protection legislation, the wage setting mechanisms and the welfare system in Spain, and that these are the factors that lie behind the decline in equilibrium unemployment. If they are right, then one can be hopeful that more decline is on the way, making it easier to have high growth and low inflation or even deflation. We are somewhat less optimistic. We think that the reforms of 1994 and 1997 fall short of fundamental changes in the labor market, and we favor an explanation that puts more weight on two one-time events. Our interpretation of events goes like this: In order to avoid the decrease in employment protection threatened by the reform of 1994, unions traded lower wage increases for maintaining employment protection. As a result, there was a transitory reduction in wage inflation. At roughly the same time, there was a change in the monetary policy regime (the Bank of Spain became independent in 1994), the probability that Spain would join the Euro started increasing, and inflation expectations fell, leading again to a shift in the Phillips curve relation. Thus, we are worried that the favorable inflation-unemployment trade-off observed in the last four years may not last. We believe that further progress still requires substantial labor market reforms.

Indeed, in our mechanical exercise, equilibrium unemployment becomes negative for a year, an impossibility which points to the difficulty in practice of achieving both a decrease in unemployment and a shift from inflation and deflation right away. Allowing for the current account to go into deficit early on would require a more gradual decrease in inflation, and thus less of a decrease in equilibrium unemployment at the start—and more of a decrease later on. More on this in the next section.

What labor market reforms may be needed to achieve such an outcome have been discussed elsewhere. It is clear that they involve at the least a further reduction in employment protection and a tightening of unemployment benefits as unemployment comes down. Past experience with the introduction and development of fixed duration contracts indicates that progress is possible, but may be not at the speed required in the simulation above. Any slowdown in growth would almost surely lead to a smaller scope for reforms, and thus make it even harder to achieve the decrease in unemployment without inflation.

5.2 Supply side reforms and productivity growth

We have ignored any potential effect of reform measures on productivity, any “supply side revolution”. Is this a reasonable assumption? One can think of two types of reforms: i) labor market reforms which decrease the bargaining power of workers and therefore allow for a lower equilibrium unemployment rate, but do not have dramatic effects on productivity (for example, a reform of the unemployment insurance and assistance system), and ii) structural reforms in the labor and goods markets which may lead to an increase in productivity (for example, in the labor market, a decrease in firing costs, allowing firms to more efficiently adjust labor). Finding productivity effects of goods market reforms (for example, privatization, deregulation of airlines and so on) has proven sufficiently difficult that we have decided to ignore them. But it is worth asking how our conclusions would change if reforms lead to an increase in productivity growth. The answer is: not much. There is no question that higher productivity growth would be good for the Spanish economy, but it may not not make it easier to achieve the required decline in unemployment. This is because there are two effects at work.

On the one hand, higher productivity growth requires higher GDP growth to achieve the same employment creation. Higher GDP growth means a higher current account deficit at a given real exchange rate. On the other hand, higher productivity growth means lower price inflation given wage inflation, and thus a higher rate of real depreciation for given wage inflation. Under the benchmark elasticities we used earlier, the net effect of higher productivity leads to a deterioration of the current account: Additional growth of 1% leads to a deterioration of the current account of 1.4% of exports; a 1% real depreciation (due to the effect of higher productivity growth on price inflation given wage inflation) leads to an improvement in the current account of only 0.6%. This computation implies that, if reforms led to an increase in productivity growth, output growth would clearly be higher, but an even larger rate of wage deflation would be needed to achieve the target decline in unemployment. One can again question the specific numbers. But the robust conclusion here is that, while higher productivity growth would be good for Spain, it would not make it easier to achieve a return to low unemployment.

6 Further discussion

The scenario we have developed is based on many of assumptions and back of the enveloppe computations. All these assumptions can be challenged. In this section, we discuss two of them, the assumption of stable export and import equations, and the requirement of external balance along the path.

6.1 Exports and imports: Time trends and the relevance of foreign direct investment

An important assumption underlying our derivation is that of stability of the export and import equations. It is this assumption which implies that high growth leads to high import growth, to a trade deficit, and to the need for real depreciation along the path.

It is a fact that faster cyclical growth typically leads to a deterioration of the trade balance: Imports increase and exports do not change, leading to a trade deficit. But it is also true that countries that grow faster than others do not systematically suffer steady real depreciation; indeed, things often go the other way. This is because, in that case, growth is associated with a change in the number of of goods produced in the country. As output increases, so does the set of goods produced, and so do exports. Growth need not in this case be associated with a larger trade deficit—or with a real depreciation if one imposes a balanced current account.

The question is therefore whether transition growth in Span (the growth needed to return unemployment to a lower level) should be thought as a cyclical adjustment, or instead as faster growth with faster underlying structural change. We do not know the answer. A look at the recent numbers for Spain (Table 3) suggests that things may be more favorable than we have assumed so far. Growth in Spain since 1995 has been slightly higher than for the EU (a cumulative 1.6%). Yet the current account has turned from a deficit of 1.4% of GDP in 1994 to a surplus of .5% of GDP in 1997; the trade balance has improved from a deficit of 3.0% to a deficit of 2.5% (not shown in the table). The numbers for exports and imports separately show high growth rates for both, suggesting indeed structural changes in the menu of goods and trade.

Rather than extrapolating from the past few years, it may be more useful to look at the experience of other countries. One clearly relevant example here is that of Ireland, which has experienced, for now more than ten years, both very high growth (an average annual GDP rate of 6.4% for the 1988-97 period) together with a sustained trade surplus (7.4% of GDP in 1988, 13.2% in 1997). There are many elements that explain the Irish performance and cannot be replicated by Spain. But one appears directly relevant, the importance of foreign direct investment (which has increased from 0.3% of GDP in 1988 to 3.7% of GDP in 1997). There are at least three ways in which foreign direct investment has been relevant for Ireland and is relevant to our scenario. First, higher foreign direct investment may lead to higher productivity growth. Second, the higher foreign direct investment, the more potential for growth through new products, and thus the more scope for shifts in the export and import functions. Third, the higher foreign direct investment, the easier it is to finance a given current account deficit. We have already discussed the implications of higher productivity growth in the previous section. We now take up the second effect in this section. We shall discuss the third when we return to the external balance constraint below.

6.1.1 Exports and imports revisited

To see whether one can find an effect of foreign direct investment on exports and imports, we reestimate export and import equations over the period 1970-1996, allowing both for time trends—to reflect ongoing structural changes—and for the stock of foreign direct investment as regressors.

Our specification builds on the work of Doménech and Taguas (1997). The dependent variables are the (log) of exports and imports of goods and services, respectively. The regressors for exports are the relative price of exports, the level of foreign income, a time trend, and the accumulated stock of foreign direct investment; for imports, the relative price of imports, Spanish GDP, an index of taxes on imports, a time trend, and the stock of foreign direct investment.

The estimation results are presented in Tables 5a (exports) and 5b (imports). Specifications which do not include the stock of FDI as a regressor and leave the coefficient on the income variable unconstrained yield estimated trends of 2.7% for exports and 2.2% for imports; when imposing the restriction that the coefficient on the income variable is equal to one, the estimated trend for exports does not change. When adding the stock of foreign direct investment as a regressor, its coefficient and statistical significance are sensitive to whether one imposes the constraint that the coefficient on the income variable be equal to one. Absent the constraint, the stock of FDI increases both exports and imports. When the constraint is imposed however, the stock of FDI decreases both exports and imports.

For definitions of the variables, see Doménech and Taguas (1997). We are grateful to David Taguas for making his data available to us. We also thank Esther Gordo for providing us with the series on foreign direct investment (and also her data on exports, imports, and relative prices which we used to check for robustness of the results in the text). The stock of foreign direct investment is obtained from the Balance of Payment statistics, net of disinvestments, and deflated by the GDP deflator.

These inconclusive macro results appear consistent with the microeconometric evidence on the effects of foreign direct investment on exports and imports. In a cross section of firms for 1991, Merino and Salas (1995) find, after controlling for industrial sectors, that foreign manufacturing firms (meaning firms with a participation of non-residents in property higher than 30 per cent) have a larger probability to import and export than Spanish firms, but this difference is only significant for firms below 200 employees. They also find that the propensity to export (measured as the proportion of exports to sales) is not significantly different between Spanish and foreign firms, but the propensity to import (measured in a similar fashion) is significantly larger for large foreign firms. Overall, they conclude that foreign firms contribute to worsening the trade balance. Moreno and Rodriguez (1998), using a sample of firms for the period 1990-94, find that this negative contribution of foreign firm on the trade balance is only significant for large firms.

Our focus here has been on FDI, but the estimated time trends are worth a short discussion. They imply that, for equal growth rates in Spain and abroad and a constant real exchange rate, the current account improves over time, by 2.7%-2.2% = 0.5% of exports per year. Put another way, if these trends continue, they imply that the required real depreciation needed to maintain current account balance under our scenario is about 0.5%/0.6 = 0.9% lower than we computed earlier. This in turn implies that, along our target path, wage inflation can be nearly 1% higher than we computed earlier.

6.1.2 The determinants of foreign direct investment

Leaving aside the question of whether foreign direct investment affects export and import relations, there is still the issue of how much foreign direct investment one should expect in Spain in the future.

Figure 6 plots foreign direct investment (to Spain and from Spain) as a proportion of Spanish GDP for the 1970-97 period. FDI to Spain, which had increased significantly in the 1980s (largely reflecting a general increase in FDI flows in the world during that period), peaked in 1990, and has decreased since. FDI from Spain has increased since the mid 1980s. As a result of these two evolutions, net foreign investment has turned negative (but small) since 1997.

Can one hope for a turnaround? Is Spain, as some have claimed, on the verge of an impending FDI boom? Net FDI flows have been typically found to be inversely related to relative labor costs. To the extent that this is true, there is clearly still substantial scope for large flows to Spain. Table 6 shows that the differences in labor costs across Euroland are still large. In fact, the coefficient of variation of hourly labor costs was larger in 1995 than in 1988 (and has been increasing except for the 1988-92 period). Since Spain is still a “low wage country”, one might guess that it will attract a larger share of the EU international investment flows once the common currency eliminates currency risks and makes costs comparisons more transparent. But there are also a number of reasons to be less optimistic. First the recent evolution is not encouraging. Second, the comparison with Ireland is revealing: Despite having roughly the same hourly labor costs as Ireland, Spain is receiving much less FDI than Ireland. Differences in the educational level of the labor force (lower in Spain ) and the language (Spanish versus English) both appear to both be important in explaining the difference. Neither difference will go away any time soon.

Our thanks to David Taguas for providing us the data.

See for example OECD 1994, chapter 3. In that study, net FDI to Spain during the 1985-92 period is above the corresponding regression line. Of more interest for our purposes would be an explanation of gross rather than net flows. Unfortunately, not much is known about the

6.2 Relaxing the current account constraint

Another important assumption underlying our scenario is that of current account balance. This is the assumption which leads to the need for a real depreciation and thus for very low inflation or deflation along the path. One might well question this assumption. Spain can surely afford to run a current account deficit, at least for some time: Its external debt is only around 20% of GDP. Being part of Euroland will also make it much easier to finance current account deficits. This raises the question: How different would our scenario look if we did not impose current account balance, but allowed Spain to run a current account deficit for some time before stabilizing its debt? Suppose for example that Spain followed the same path of output as under our scenario, but kept its inflation rate equal to that of its trading partners, namely equal to 2%. This would have two implications.

First, higher inflation (relative to our original scenario) would require less of a decrease in equilibrium unemployment at the beginning of the adjustment. From the Phillips curve above, equilibrium unemployment must follow:

\[u ^ {*} = u + \frac {\gamma}{\beta} \Delta u = u - 2. 5 (2 \%) = u - 5 \%\]

Equilibrium unemployment would therefore have to remain 5 points below actual unemployment throughout (until unemployment has been reduced to its target level), instead of having to drop significantly at the beginning to achieve disinflation as under the original scenario.

determinants of the patterns of gross international investment flows. For studies specifically addressed at Spain see Bajo and Sosvilla (1994), and Bajo and López (1997), for foreign direct investment into Spain, and Campa and Guillén (1996) for Spanish direct investment abroad.

In 1990 the proportion of adult population -above 25- with a university degree was 14.6% in Ireland, 8.5% in Spain; the proportion of adult population with a secondary degree was 43.7% in Ireland, 25.6% in Spain. For international comparisons of educational attainments, see Boscá, de la Fuente and Doménech (1996).

Second, constancy of the real exchange rate in the face of high growth would lead to a steadily increasing trade deficit. From our computations above, the trade deficit would increase each year by 0.56% of GDP (this is under the assumption of unit income elasticities; 0.78% of GDP if we used estimated income elasticities of 1.4). Thus, under this scenario, the trade deficit would reach 3.9% of GDP in 2005.

What would happen to the ratio of foreign debt to GDP is given by the accumulation equation:

\[b _ {t} - b _ {t - 1} \equiv (r - g) b _ {t - 1} - \left(x _ {t} - q _ {t}\right)\]

where is the ratio of debt to GDP in year t, r and g are the real interest rate and the rate of growth of real GDP, and are respectively the ratios of exports and imports to GDP. Under our target path, g is equal to 4.5%, and if we take the nominal interest rate for Euroland to equal 4.5%, r = 2.5%, and thus . Using this equation, the increase in the ratio of debt to GDP due to the accumulation of trade deficits would reach 12% of GDP in 2005.

Suppose that in 2005, Spain decided to stabilize its debt to GDP ratio as the then prevailing level. With Spain now on its normal growth path, g would be back at 2.5%. If we assume for simplicity that r remains equal to 2.5%, then r - g is equal to zero, and stabilizing the debt just implies eliminating the trade deficit. Using the estimated price elasticities we discussed earlier, this would require a depreciation equal to 3.9% divided by .28 (the share of exports in GDP) and divided by 0.6 (the effect of the real exchange rate on the trade balance divided by exports), or 23%. If Euro inflation was running at 2% a year, this would require about 10 years of zero inflation, starting in 2005.

Thus, (partly because r - g is negative under this scenario, and is likely to remain low in the future), debt dynamics are favorable, and the costs of waiting to adjust the real exchange rate are not very large. A substantial real depreciation—and thus a rate of inflation lower than the EU for some time—is eventually required. But it need not be achieved right away. Keeping inflation initially at the same level as the EU (already a substantial achievement, if unemployment decreases by 2 points a year), and decreasing it later on is clearly feasible.

r - g

If we assumed that r - g was positive—as we think it should be in steady state—waiting to eliminate the trade deficit would lead to further accumulation of debt and thus to a further increase of the eventual trade surplus needed to pay interest on the accumulated debt, and, thus, requires a larger real depreciation.

The steady appreciation along the path under this alternative scenario raises however another issue, namely that it may choke off the demand for domestic goods and make it more difficult to sustain growth. An offsetting effect is that, given the EU nominal interest rate, this scenario, with higher inflation than our benchmark, means a lower real interest rate. Working out whether this combination of steady appreciation and lower interest rates than under the benchmark can generate a level of demand consistent with steady growth is something which goes beyond our back of the enveloppe computations, and requires a general equilibrium model.

7 Concluding remarks

We started the paper by asking what it would take to reduce Spanish unemployment to 5% in 2005. The answer is that it will take a lot. It will take a growth rate of output of 4.5% per year. If internal balance and current account equilibrium are to be maintained, it will take a major shift in the composition of demand, towards investment and away from consumption and government spending. And if current account equilibrium is to be maintained, it will require inflation 3.5% below that of Spain's partners, thus price deflation and wage inflation close to zero.

Choosing a less ambitious path, say a reduction of the unemployment rate to 5% in 2008 both reduces required growth to about 4%, and the inflation differential to 2.5%. The message remains the same however: high growth and very low inflation, perhaps deflation, are needed to achieve the required unemployment reduction.

Our exercise has been an exercise in reverse engineering, starting from a target path, and working out macro implications. Our purpose was to highlight a number of features that have to hold along such a path. If they do not hold, then Spain will not achieve the growth it needs to achieve low unemployment. The problems may come in many forms. If demand is not there, growth will obviously fall short of what is needed to reduce unemployment. If growth falters, reforms will be politically more difficult to implement. If reforms are not there, growth will lead to higher, not lower inflation; this in turn will lead, through appreciation and the need to limit the current account deficit, to a slowdown in growth. Thus, both demand and reforms are of the essence. At the same time, the usual macroeconomic tools will not be there to help. Luck will have to be an essential ingredient of success.

8 References

1. Bajo, O. and C. López (1997): “An Industry Analysis of Foreign Direct Investment in Spanish Manufacturing, 1986-92”, mimeo.
2. Bajo, O. and M. Montero (1995): “Un modelo econométrico ampliado para el comercio exterior español, 1977-92”, Moneda y Crédito, 201, 153-182.
3. Bajo, O. and S. Sosvilla (1994): “Un análisis empírico de los determinantes macroeconómicos de la inversión extranjera directa en España, 1961-89”, Moneda y Crédito, 194, 107-48.
4. Boscá, J., A. de la Fuente, and R. Doménech (1996): “Human capital and growth: Theory ahead of measurement”, Universidad de Valencia, mimeo.
5. Blanchard, O. (1998): “Revisiting European Unemployment: Unemployment, Capital Accumulation, and Factor Prices”, NBER, working paper no. 6566.
6. Buisán, A. and E. Gordo (1994): “Funciones de importación y exportación de la economía española”, Investigaciones Económicas, 18(1), 165-92.
7. Campa, J.M. and M.F. Guillén (1996): “Spain: A boom from economic integration” in J.H. Dunning and R. Narula (eds.), Foreign Direct Investment and Governments: Catalysts for Economic Restructuring, New York: Routledge.
8. CEPR (1995), Spanish Unemployment: Is There a Solution?, London: CEPR.
9. Dabán, T., A. Díaz, F.J. Escribá, and M.J. Murgui (1998): “La base de datos BD.MORES”, Ministerio de Economía y Hacienda, working paper D-98001.
10. Dolado, J.J. and J.F. Jimeno (1997): “The Causes of Spanish Unemployment: A Structural VAR Approach”, European Economic Review, 41, 1281-1307.
11. Dolado, J.J. and J.D. López-Salido (1998): “Unemployment and Inflation Persistence in Spain: Are There Phillips Trade-offs”, mimeo.
12. Doménech, R. and D. Taguas (1997): “Exportaciones e Importaciones de Bienes y Servicios”, Moneda y Crédito, 205, 13-44.
13. Escribá J. and J.R. Ruiz-Tamarit (1992): “Una contribución a la medición económica del capital (España 1964-1990)”, Departamento de Análisis Económico, Universidad de Valencia, QT. 233.
14. Escribano, A. (forthcoming): “Predicción y análisis de funciones de exportación e importación en España”, Investigaciones Económicas.
15. Fernández Cordón, J.A. (1998): “Proyecciones de la población española”, FEDEA, working paper 98-11.
16. García, C. and E. Gordo (1998): “Funciones trimestales de exportación e importación para la economía española”, Bank of Spain, mimeo.
17. Hunt, J. (1998): “Hours Reductions as Work-Sharing”, Brookings Papers on Economic Activity, (1) 339-81.
18. Mas, M., F. Pérez and E. Uriel (1996), El “stock” de capital en España y sus Comunidades Autónomas, Madrid: Fundación BBV.
19. Mauleón, I. and L. Sastre (1994): “El saldo comercial en 1993: Un análisis econométrico”, Información Comercial Española, 735, 161-72.
20. Merino, F. and V. Salas (1995): “Empresa extranjera y manufactura española”, Revista de Economía Aplicada, 9, 105-31.
21. Moreno, L. and D. Rodriguez (1998): “Efectos de la inversión directa extranjera en los flujos comerciales de las empresas”, Investigaciones Económicas, 22(2), 179-99.
22. OECD (1994), Employment Outlook, Paris: OECD.

Table 1a

Table 1b

Notes: d1: dummy 1980-85, d2: dummy 1986-91, dummy 1992-97. : 0.029 for 1978-85, .028 for 1986:91,.022 for 1992:97. Standard errors (robust to serial correlation) in brackets.

Table 2

Notes: Population and Labor Force in thousands. Source: Labor Force Survey, INE, Fernández-Cordón (1998) and authors' calculations.

Table 3

ance of Payments, Bank of Spain, for the second panel. Note: Net exports, rents and current transfers do not add up to the balance of the current account due to some minor methodological differences between the National Accounts and the Balance of Payment Statistics in the calculation of net exports. *Annual average. Source: National Accounts, INE, for the first panel, Bal-

Table 4

Dependent variable: Quarterly Changes in GDP Deflator (excluding indirect taxes) (Sample period: 1978:3-1997:4)

Notes: Regressions also include segmented trends for 1978:3-1985:4, 1986:1-1991:4, and 1992:1-1997:4. Standard errors in brackets. Unemployment gap1: Deviations of the unemployment rate from a 4th-degree polynomial in t. Unemployment gap2: Deviations of the unemployment rate from quarterly structural unemployment (obtained by interpolation of the annual OECD series).

Unemployment gap3: Cyclical component of the unemployment rate (obtained by Hodrick-Prescott filtering with )

Table 5a

Dependent variable: (log) Exports of goods and services (Sample period: 1970-96)

Table 5b

Dependent variable: (log) Imports of goods and services (Sample period: 1970-96)

Notes: Restricted. p-value for the restriction. Standard errors (robust to serial correlation in brackets).

Table 6 Hourly labor costs (ecus)

Figure 1. Rate of return to capital, capital share and capital-output ratio

Figure 2. Tobin's q and user's cost of capital.

Figure 3. The rate of growth of technological progress

Figure 4a. Unemployment and proxies for structural unemployment

Figure 4b. Proxies for the unemployment gap.

Figure 5. Unemployment, inflation and equilibrium unemployment

Figure 6. Foreign direct investment as proportion of GDP.

COLECCION RESUMENES

98-01: “Negociación colectiva, rentabilidad bursátil y estructura de capital en España”, Alejandro Inurrieta.

TEXTOS EXPRESS

98-02: “Sector turístico y crecimiento del empleo en la Comunidad Autónoma de Canarias: Un ejercicio de prospección al horizonte 2011”, José A. Herce y Simón Sosvilla.

98-01: “El gasto sanitario en España: Evolución reciente y perspectiva”, Javier Alonso y José A. Herce.

DOCUMENTOS DE TRABAJO

99-02: “Reducing Spanish unemployment under the EMU”, Olivier J. Blanchard y Juan F. Jimeno.

99-01: “Further evidence on technical analysis and profitability of foreign exchange intervention”, Simón Sosvilla-Rivero, Julián Andrada-Félix y Fernando Fernández-Rodríguez.

98-21: “Foreign direct investment and industrial development in host countries”, Salvador Barrios.

98-20: “Numerical solution by iterative methods of a class of vintage capital models”, Raouf Boucekkine, Marc Germain, Omar Licandro y Alphonse Magnus.

98-19: “Endogenous vs exogeneously driven fluctuations in vintage capital models”, Raouf Boucekkine, Fernando del Río y Omar Licandro.

98-18: “Assesing the economic value of nearest-neighbour exchange-rate forecasts”, F. Fernández-Rodríguez, S. Sosvilla-Rivero y J. Andrada-Félix.

98-17: “Exchange-rate forecasts with simultaneous nearest-neighbour methods: Evidence from the EMS”, F. Fernández-Rodríguez, S. Sosvilla-Rivero y J. Andrada-Félix.

98-16: “Los efectos económicos de la Ley de Consolidación de la Seguridad Social. Perspectivas financieras del sistema tras su entrada en vigor”, José A. Herce y Javier Alonso.

98-15: “Economía, mercado de trabajo y sistema universitario español: Conversaciones con destacados macroeconomistas”, Carlos Usabiaga Ibáñez.

98-14: “Regional integration and growth: The Spanish case”, Ana Goicolea, José A. Herce y Juan J. de Lucio.

98-13: “Tax burden convergence in Europe”, Simón Sosvilla-Rivero, Miguel Angel Galindo y Javier Alonso.

98-12: “Growth and the Welfare State in the EU: A cusality analysis”, José A. Herce, Simón Sosvilla-Rivero y Juan J. de Lucio.

98-11: “Proyección de la población española 1991-2026. Revisión 1997”, Juan Antonio Fernández Cordón.

98-10: “A time-series examination of convergence in social protection across EU countries”, José A. Herce, Simón Sosvilla-Rivero y Juan J. de Lucio.

98-09: “Estructura Demográfica y Sistemas de Pensiones. Un análisis de equilibrio general aplicado a la economía española”, María Montero Muñoz.

98-08: “Earnings inequality in Portugal and Spain: Contrasts and similarities”, Olga Cantó, Ana R. Cardoso y Juan F. Jimeno.

98-07: “Labour reallocation and labour market institutions: Evidence from Spain”, Carlos García-Serrano y Juan F. Jimeno.

98-06: "Benchmark priors for Bayesian model averaging", Carmen Fernández, Eduardo Ley, Mark F. J. Steel.

98-05: “The effects of externalities on value added and productivity growth in Spanish industry”, Juan J. de Lucio, José A. Herce y Ana Goicolea.

98-04: "Employment segmentation, labour mobility and Mismatch: Spain, 1987-1993", Sonsoles Castillo, Juan F. Jimeno y Omar Licandro.

98-03: “Un análisis global, regional y sectorial de los efectos externos de conocimiento”, Juan J. de Lucio.