Spain: evaluating the effects of macro policy using an econometric model.

Anderton, Bob

Introduction(1)

In recent years Spain has experienced rapid economic development during a period of increasing European integration. Spanish policymakers have enhanced their credibility by becoming a member of both the EC and the ERM. A strong commitment to Spanish participation in a future European Economic and Monetary Union confirms Spain's determination to 'catch-up' and converge towards the model of the more advanced economies of Europe. The National Institute's Global Econometric Model (NIGEM) now contains a detailed econometric model of Spain. This note begins with an overview of economic events and policy in Spain over the last twenty years. This is followed by sections which describe in more detail the reforms and economic outcomes regarding the labour market, price and wage determination, the balance of payments and the finances of the public sector. The econometric model is then described and the effects of the structural reforms upon the individual equations evaluated. The penultimate section uses the model to simulate the effects of macroeconomic policies which the Spanish authorities have either implemented or may be considering. This is followed by a conclusion.

Overview

Table 1 shows that over the period of the two oil price shocks, from the mid-1970s to the mid-1980s, Spain experienced sluggish growth. During the 10 years between 1975 and 1984, real investment declined by almost 2 1/2 per cent per year, employment fell by 15 per cent and real consumption and GDP only grew by around 1 per cent per annum. As a result, the unemployment rate increased from 3 per cent in 1974 to 22 per cent in 1985. At the end of 1982 the new government implemented stronger monetary policy combined with rapid restructuring of the industrial and financial sectors supplemented by policies aimed at wage moderation. These policies improved profitability and restrained wages with the result that interest rates fell slightly. These acts were augmented by further supply-side legislation between 1984-86 which began the process of dismantling financial and exchange controls combined with labour market measures aimed at improving flexibility.(2) In particular, temporary employment contracts were introduced along with rules allowing easier termination of these contracts. EC membership in 1986 started the process of tariff reduction and facilitated the introduction of VAT which, along with a deceleration of public spending, helped to reduce the government deficit. These policies attracted substantial foreign investment as foreign multinational corporations began to perceive Spain as both a high return investment opportunity and a gateway to the larger EC market.

After EC membership Spain experienced rapid real GDP growth of above 4 per cent for the four years 1985-88. Although consumers' expenditure expanded at an annual rate of more than 4 per cent over this period, investment was also a major source of growth rising by almost 11 per cent in each year. Consumption was fuelled by strong growth in both real incomes and wealth. Real incomes grew partly as a response to robust employment while wealth benefitted from inflated asset prices resulting from strong capital inflows. However, a significant proportion of domestic demand was satisfied by imports which grew by around 14 per cent over the same four years. This acceleration in imports was due to both poor price competitiveness, arising from high relative inflation, and strong demand. By 1990 Spain's current account deficit was around 3 1/2 per cent of GDP. In terms of counterparts, this primarily reflected a significant decline in the savings rate of households. This was not offset by increased savings of the public sector which itself was financially adversely affected by an unemployment rate above 15 per cent throughout the second half of the mid-1980s. This caused a savings/investment imbalance resulting in a current account deficit.

Spain's consumer price inflation has declined substantially from an average of 16 per cent between 1975-84 to around 6 1/2 per cent in 1990. This reduction was achieved in a period of strong activity with unemployment falling from a peak of around 21 per cent in 1986 to approximately 16 per cent in 1990. The decline in price inflation has largely followed the deceleration in wage inflation which, in turn, has benefitted from the TABULAR DATA OMITTED vast expansion in temporary employment in low paid service sector jobs, which has resulted in a decline in average wages. Trade liberalisation has successfully encouraged convergence of industrial price growth but price inflation in services is not responding so well. Furthermore, a poor productivity performance, in terms of growth in output per head, contributed to robust growth in unit labour costs in the second half of the 1980s. This is partly a consequence of employment growth being concentrated in the service sector which is characterised by a relatively small rate of labour augmenting technical progress.

Although Spain has made progress in implementing several economic reforms, many structural deficiencies remain. Further reforms are required to reduce imperfections in the labour market, to make the service sector more competitive and to bring the public sector deficit and debt ratios down to sustainable levels. If Spain is to become a member of a future monetary union, or even maintain membership of the ERM, reforms are necessary to enable further progress towards nominal convergence.

The labour market and price and wage determination

Spain is characterised by a relatively low rate of labour force participation, particularly for females.(3) Furthermore, over the last 20 years most G7 countries have had rising participation rates but Spain has remained fairly static. However, this was associated with a period of economic stagnation in Spain (mid-1970s to mid-1980s) which possibly resulted in 'discouraged worker' effects. The period of stronger growth from 1986 has increased total participation slightly but this masks strong differences between ages and genders. From 1986 adult female participation rose strongly whereas adult male participation remained static and male youth participation continued on a strong downward trend. The latter is probably related to the particularly high unemployment rates for this category, the type of employment growth over this period and barriers to entry into permanent employment.

Spain is characterised by a labour market with a fairly strict segmentation into permanent and temporary labour. Table 2 shows that a high proportion of the rapid employment growth from 1986 onwards was concentrated in the temporary sector. At the end of 1987 temporary employment accounted for 17 per cent of total employment. By 1991 it had risen to 32 per cent with most of the new temporary employment concentrated in low paid jobs in the service sector. The Employment Promotion Programmes (EPP) and other legislation easing the restrictions on the adoption of temporary contracts began around 1984 and explains the growth in temporary work. One component of the EPPs encouraged employment of temporary workers via relatively lower employer social security contributions and much lower termination payments for fixed contract workers relative to permanent labour. In the case of permanent worker dismissals the law provides for compensation, TABULAR DATA OMITTED depending upon the size of firm, of between 20 and 45 days pay for each year worked. This, in turn, encourages low turnover of permanent employees with the result that redundancy payments average around two years of salary. The EPPs allowed firms to avoid the higher firing costs and rigid rules regarding permanent workers by introducing fixed-term employment contracts of six months duration that could be renewed for up to a period of three years.(4) The EPP therefore satisfied a demand for flexible low-cost labour. Furthermore, the rise in adult female labour participation showed that, combined with changing attitudes towards working women, there was a sufficient supply of this type of labour.

The Spanish economy seems to be characterised by both high wage inflation and high unemployment. Table 3 shows unemployment by age, gender and duration. It is clear that over 50 per cent of the unemployed have been without jobs for more than a year which indicates a high degree of unemployment persistence. Furthermore, prime age males, who dominate employment in the permanent sector and would probably put the most downward pressure on wage inflation, have a lower unemployment rate compared to the other categories. In addition, strong growth regions dominate wage bargaining and the resulting wage settlements are emulated by low activity regions. As a consequence, wages do not seem to respond strongly to local unemployment or productivity conditions hence the dispersion of regional wage relativities is limited. Many of these factors encourage a high degree of real wage rigidity and the economy gets locked into a high inflation-high unemployment trajectory.(5)

The high unemployment rate may also be the result of a high sustainable rate of unemployment. Labour market mismatch and a high replacement ratio are often cited as the major structural impediments to low unemployment. One reason for a skills mismatch is that temporary workers eventually become unemployed as employers want to stop these workers becoming permanent employees and avoid paying the higher fixed costs. Consequently, training and skills acquired during temporary employment are often wasted particularly given the high incidence of long-term unemployment. Furthermore, difficulties in matching workers and jobs are compounded by the official public employment offices (INEM). INEM do not offer an efficient job placement service as their work is dominated by the distribution of unemployment benefit to the large numbers of unemployed. In addition, Bentolila and Blanchard (1990) argue that improvements in unemployment benefit in Spain explain a substantial fall in job search.(6) Unemployment benefit is approximately 80 per cent of the last wage for the first six months of unemployment declining to 70 per cent for the next six months.

Table 3. Unemployment by age, gender and duration

 1975 1980 1985 1991

Total unemployment 4.3 11.1 21.2 16.3

Males:

Adults 3.4 7.7 15.4 9.8
Youth 9.8 23.0 39.3 25.7

Females:

Adults 1.8 5.9 16.0 21.0
Youth 7.8 29.3 51.0 37.9

Long-term unemployment(a) - - 56.7 51.1

Notes:

Adults refers to those between the ages of 25-54.

Youths are 16-24 year olds.

(a) Those unemployed for 12 months and over as a percentage of
total unemployment.

Major attempts to decelerate wage inflation began in 1978 as both labour organisations and the government agreed a formula for wage indexation incorporating ex ante price expectations based on declining inflation targets. But these agreements came to an end in 1987 as trade unions felt that government inflation targets were far too ambitious. However, Table 4 shows that wage TABULAR DATA OMITTED inflation has actually declined most rapidly since 1985. High unemployment explains some of this fall but the major reason may be compositional. The rapid movement towards temporary work, which has been primarily female and in the low-paid service sector, has resulted in a downward movement in nominal compensation per employee during the period of rapid growth between 1986 and 1989. More recently earnings growth has remained fairly buoyant even though unemployment is currently high and GDP growth is declining. This may be partly explained by the fact that temporary workers are usually the first to lose their jobs in a downturn thereby changing the composition of employment back towards more highly paid permanent workers.(7)

Unit labour costs throughout the whole economy did not show much deceleration during the upturn in the second half of the 1980s as productivity growth was not very impressive given the strength of output growth. But Table 4 shows that manufacturing unit labour costs decelerated more rapidly. In addition, price inflation followed wage inflation in a downward direction. The GDP deflator declined from a growth rate of almost 16 per cent between 1975-84 to around 7 per cent in the second half of the decade. However, the exchange rate played a considerable part in dampening price inflation as strong TABULAR DATA OMITTED capital inflows put upward pressure on the exchange rate from 1985 onwards.

Table 5 shows Spain's comparative consumer price inflation performance vis-a-vis fellow EMS members. It is apparent that in the second half of the 1980s the total inflation differential has substantially narrowed. However, manufactured goods inflation has virtually converged to the EMS average. It is inflation in the service sector which is preventing convergence of total inflation, but some progress has been made in reducing distortions. Spain, in compliance with EC directives, introduced a new competition law in 1989 which forbids agreements which jeopardise competition. The resulting court cases primarily concern the service sector (e.g. against unreasonable commissions charged by banks, or repair tariffs and profit margins in second hand vehicle sales and so on). The deregulation of the insurance market in 1992 was partly in response to insurance premiums rising so rapidly. Distortions in the property market have restricted the supply of urban land which has also forced prices upwards. The inevitable rental and home price inflation feeds into wage claims. Public sector wage inflation has also frequently outstripped private sector pay therefore encouraging higher wage claims in the latter sector.

However, the Spanish authorities have taken action on a macro scale in an attempt to achieve lower inflation. ERM membership should help to reduce inflationary TABULAR DATA OMITTED expectations and Spain's commitment to EMU should encourage bargaining groups to realise that nominal convergence is a necessity. Furthermore, Spain's plans to implement legislation to create a Bundesbank-style independent central bank should give greater priority to inflation reduction. EC membership has exposed Spanish domestic producers to a greater degree of foreign competition. The resulting downward pressure on both profit margins and costs has encouraged inflation convergence, particularly in manufacturing.

Balance of payments

Table 6 shows that Spain's current account deficit averaged around 2.5 per cent of GDP between 1980-82 but then moved into a similarly sized surplus over the next 5 years to 1987. This improvement was largely attributable to the trade account which benefitted from improvements in competitiveness resulting from a prolonged period of exchange-rate depreciation and weak domestic demand. Chart 3 shows that an 8 per cent depreciation of the peseta in December 1982 was followed by further decline in the next year amounting to a 23 per cent depreciation against the dollar and a nominal effective depreciation of over 15 per cent. Total exports responded dramatically, rising in volume terms by 17 per cent in 1984 whereas total import volumes remained flat in the same year. Imports remained generally subdued from 1982-85 not only due to domestic competitiveness gains but also because of subdued domestic activity. Exports also seemed to benefit from weak internal demand probably as a result of redirection of production from domestic markets to overseas. The invisibles account, particularly tourism receipts, also contributed to the improvement in the current account up to 1987. In 1986 tourism brought in 12 billion dollars of foreign exchange and comprised the largest component of the invisibles surplus.

However, from 1987 onwards the current balance again began to deteriorate, moving from a surplus of over 2 per cent of GDP in 1986 to a 3 per cent deficit in 1992. This reflected several factors; a loss in competitiveness resulting from a rise in the real exchange rate, strong domestic demand and a dramatic reduction in import tariffs after joining the EC in January 1986.(8) This rise in the real exchange rate primarily reflected a nominal peseta appreciation resulting from the relaxation of constraints governing Foreign Direct Investment (FDI) inflows. These constraints were lifted in mid-1985 and TABULAR DATA OMITTED FDI inflows were further encouraged by Spain's entry into the EC. However, a large proportion of FDI were short-term in nature. Net long-term capital inflows only began to grow strongly towards the end of the 1980s. This has continued into the early-1990s perhaps encouraged by the prospect of a more stable exchange rate after Spain joined the ERM in 1990.

The rapid growth of real GDP, averaging just under 5 per cent between 1985 and 1990, combined with the process of tariff reduction, provoked a dramatic rise in imports. In particular, the strength of investment over this period encouraged imports of capital goods to grow by an average of more than 25 per cent per year in the years between 1987 and 1990. Between 1986 and 1989 the volume of total imports grew by an annual average of 19.5 per cent in comparison to just below 4 per cent for total exports volume. In more recent years, as domestic demand has decelerated, total export and import volumes have registered similar growth rates of approximately 12 per cent and 10 per cent respectively in 1991. Furthermore, exports of goods are now benefitting from previous FDI inflows as 40 per cent of Spain's manufactured exports are now produced by foreign MNCs. However, this may also help explain Spain's deteriorating invisibles surplus as net factor income payments to foreigners have increased which partly reflects repatriation of MNCs profits. But the upward trend in the net surplus on tourism has also flattened out reflecting stability of receipts from foreign tourists combined with rapid growth of tourism from Spain. This seems to be a consequence of both the substantial appreciation of the peseta in recent years and the relatively strong growth in Spain's service sector inflation.

Public sector debt and deficit

Table 7 shows that the general government deficit on a national accounts basis (excluding net lending) gradually deteriorated after 1975 but grew worse between 1980-85. The deficit rose from an average of just over 1.8 per cent of GDP between 1977-79 to an average of almost 5 per cent between 1980-85. Although this reflected a period of economic stagnation, and hence slow revenue growth, the deterioration was also due to rapid expansion of the public sector, improvements and extension of coverage in social security benefits and subsidies to enterprises. By 1986 the deficit had reached almost 6 per cent of GDP. The accumulation of deficits caused gross public sector debt to spiral from around 15 per cent of GDP in 1977-79 to almost 48 per cent of GDP in 1986.

In the second half of the 1980s the government deficit began to decline somewhat. Strong GDP growth increased tax revenues and unemployment benefit payments decreased as employment growth started to push unemployment downwards. Furthermore, the introduction of VAT in 1986, although adding about 2 percentage points to the consumer price index, caused a sharp increase in government revenues. In addition several measures were implemented to make tax evasion more difficult. By 1989 the deficit had declined to about 2.8 per cent of GDP and debt had stabilised at roughly 43 per cent of GDP. Given the stage and strength of the cycle, combined with the above improvements to the structural deficit, one would have expected a more substantial recovery in the government's finances. However, the opportunity to increase spending on the infrastructure afforded by strong growth was given greater priority and the policy of deficit reduction was diluted from 1988. The deceleration in GDP growth in 1990 and 1991 has put renewed upward pressure on the deficit which is now close to 5 per cent of GDP and the debt ratio is rapidly approaching 50 per cent.(9)

The options available for reduction and stabilisation of the deficit are limited. Public expenditure may be difficult to reduce as it is already small as a proportion of GDP by international standards. However, budgetary decentralisation has allowed a large borrowing requirement to develop among the regional authorities amounting to about 1 per cent of GDP. More effective control of regional public spending may be able to decrease the deficit. Employers' social security contributions are already high by European standards and it would not be wise to increase them further particularly when unemployment is so high. But further advances concerning the reduction of tax evasion are possible. Tax evasion not only decreases tax revenue but reduces the effectiveness of fiscal policy. Given that ERM membership requires greater use of fiscal policy for demand management, because monetary policy is primarily devoted to maintaining the exchange rate parity, the elimination of tax fraud takes on a greater urgency.

An econometric model of Spain

This section begins by describing the key equations and long-run parameters of our estimated econometric model. The equations are then used to evaluate whether the structural changes and reforms discussed above have had any discernible influence upon the Spanish economy.(10)

Econometric model

Wages. The compensation equation is based upon the bargaining framework popularised by Layard, Nickell and Jackman (1991).(11) The equation assumes that wages grow in line with labour productivity and producer prices in the long run. Wages are also negatively related to unemployment as the latter is used as a proxy for bargaining power. Tax wedge effects are only present in the short-run dynamics. The equation is estimated by instrumental variables as it is forward-looking in prices. The actual out-turn of inflation one period ahead is used as the proxy for expected inflation. The dynamic forward term has a coefficient fairly close to unity which is consistent with the move to ex ante price expectations indexation from 1978 onwards.

Prices. The wholesale and consumer price equations are a weighted sum of unit labour costs and import prices plus pressure of demand terms proxied by capacity utilisation. The latter have positive signs which indicate a procyclical mark-up. The greater exposure to foreign competition of the wholesale sector relative to the service sector is indicated by the relative sizes of the capacity utilisation parameters. The larger parameter in the consumer price equation (which includes services) indicates that the excess rent generated in the service sector allows a substantial variation in the mark-up over the cycle. Another difference between the sectors is that wholesale prices have a relatively higher weight for import prices. This reflects the high proportion of imports of semi-manufactured goods, the large volume of capital goods imports and the increased 'openness' of the tradeable goods sector in recent years.

Trade. Exports of goods are assumed to grow in line with demand (which is proxied by imports of goods of other countries weighted by their importance in Spanish export markets). The exports equation also includes a relative price elasticity of around unity and a time trend which indicates that, other things being equal, Spain's exports grow by 2 per cent per annum.(12) The import volumes equation has a large demand elasticity of just over 1.5. But this is not unusual when demand is represented by total final expenditure as NIGEM has similar marginal import propensites for the US and UK. The price competitiveness elasticity for imports is again relatively large at 0.7.

The export price of manufactures equation indicates that Spain's trade prices are equally influenced by domestic and competitors' prices. The weight allocated to foreign prices is large compared to some countries on NIGEM (e.g. USA, Germany and France) but this is to be expected given that Spain is a smaller country than most of the G7 members.

Consumption and investment. Real consumers' expenditure is assumed to grow in line with real personal disposable income and is negatively related to short-term interest rates (proxying consumer credit effects). However, Spanish consumption seems to be relatively more interest sensitive than other countries on NIGEM. Total real investment is assumed to grow in line with real GDP and also has a negative elasticity with respect to long-term interest rates. An intercept shift dummy variable with a value of 1 from 1986 onwards is required to capture the strong investment growth after EC entry. The semi-elasticity of long rates seems small compared to other investment equations on NIGEM and seems to reflect two factors. First, interest rate increases in the second half of the 1980s did not subdue investment as much as one would expect due to the greater opportunities for investors from EC entry. Second, housing investment is less responsive to interest rates in Spain because regulations severely restrict urban land use and housing construction.

The model and structural reforms

There are several aspects of the above model which reflect the structural changes described earlier:

First, the compensation specification displays instability over the post-1984 period. This seems to be associated with the introduction of flexible labour contracts. We find that the dynamic forecast of a wage equation estimated using data until the end of 1984 overpredicts the growth of earnings in the second half of the 1980s. This is consistent with the fact that the growth in temporary labour was associated with lower employers' social security contributions and low-paid employment in the service sector. Therefore, the instability is probably capturing the change in the composition of employment over this period. Consequently, we performed stability tests by re-estimating the wage equation over the whole sample period with the addition of intercept, time-trend and slope dummy variables for the period after the supposed change in structure. We found that the intercept and time-trend dummies were significant with positive and negative parameters respectively. We interpret this result in a similar fashion to Dolado and Bentolila (1993). The positive dummy indicates that the legislation which eased firing restrictions for temporary workers increased the relative job security of permanent workers. Hence the legislation reduced the threat of unemployment for this category of workers and put upward pressure on wages. The negative time trend may be explained by the growth in low-paid temporary employment which gradually decreased the average wage as the composition of employment shifted away from permanent workers. Therefore, in our simulations we use the wage equation incorporating the dummies.(13)

Second, the investment equation contains a positive intercept dummy variable to capture the rapid rise in capital expenditure in the second half of the last decade. This represents several factors which cannot be captured by the equation. Ceteris paribus, flexible labour contracts allowed labour demand to become more cyclically responsive and improved both current and future profitability. In addition, the relaxation of capital controls, EC membership and the prospect of a Single European Market encouraged foreign and domestic investment in Spain. Furthermore, financial deregulation allowed easier financing of investment by allowing firms greater access to the stock market.

Third, the removal of trade protection, combined with membership of the EMS, created pressure for Spain's tradeable goods sector prices to converge to those of her competitors. Chart 5 shows a 'rolling window' regression of the export price equation and it is clear that the weight given to competitor's prices, relative to domestic prices, is greater after the trade liberalisation associated with EC membership. However, the process seems to have stabilised in recent years and our equation should be adequate for simulation purposes.(14)

Model simulations

We conducted three simulations using our model of Spain. First, we reduce short-term interest rates temporarily by 1 percentage point to get some idea of the inflationary consequences of a loosening of monetary policy. The latter may occur because continued Spanish membership of the ERM implies a tendency towards lower interest rates. Conversely, reversing the signs of the simulation results gives us some idea of the effects of tighter monetary policy in the 1980s. Second, we simulate the effects of a permanent reduction in government expenditure equivalent to 1 per cent of GDP. This allows us to compare the effects of monetary policy and fiscal policy. This is an important issue as fiscal policy may have to play a vital role in dampening demand if ERM membership restricts monetary policy to the exchange rate target.(15) In addition, Spain's public sector deficit needs to be reduced to satisfy the Maastricht criteria. If this is to be achieved by spending cuts then we can gain a rough idea of the effects upon GDP. Third, we can examine the effects of a 10 per cent devaluation of the peseta against the dollar (which also represents a 10 per cent depreciation against the D-Mark). This will give us a valuable insight into the effects of the recent devaluations of the peseta within the ERM.

In the following simulations nominal exchange rates and interest rates are exogenous. However, economic interactions between Spain and the rest of the world are fully captured as the whole of NIGEM is simulated. It is assumed that shocks are unexpected and therefore do not influence forward-looking variables, such as wages, before the event occurs. The final simulation is followed by a brief comparison of the results with other countries.

A temporary 1 percentage point reduction in short-term interest rates

Short-term interest rates are reduced temporarily by 1 percentage point for two years and then gradually return to base over the following year. The temporary nature of the reduction means that long-term rates, which represent the key borrowing rate for investment, decline temporarily by a much smaller amount. Chart 6 shows that the increase in the level of GDP peaks at roughly 0.4 per cent after about three years and then declines back to base when the monetary loosening is reversed. Chart 7 shows that an increase in total investment accounts for a large part of the rise in GDP. The increase in investment peaks at about 2 per cent after three years and then returns to base. However, the increase in consumers' expenditure is only slightly above 1/2 per cent before returning to base. Chart 6 indicates that the extra activity causes the consumer price level to rise by about 1 per cent after five years. Consequently the percentage effects upon the price level of an interest-rate reduction are greater than those for real GDP. Furthermore, the degree of nominal inertia is evident from Chart 6 as prices continue to rise for two years after the rise in output has peaked. The extra activity from the temporary reduction in interest rates also permanently reduces the government debt/GDP ratio by about half a percentage point.

A permanent reduction in government expenditure equivalent to 1 per cent of GDP

Chart 8 shows that the reduction in government expenditure initially decreases real GDP by 1 per cent in the first year. But this effect rapidly declines to around 0.6 per cent by the start of the second year and almost disappears after about seven years. Chart 9 shows that the decline in the price level peaks at about 2.5 per cent after about six years. These price effects reduce the short-term government expenditure multiplier by improving competitiveness and reducing the fall in real personal disposable income. Hence consumption does not fall as much as real GDP and exports and imports begin to make a positive contribution to GDP. If we compare the results to the simulation above we can see that a substantial reduction in government expenditure is required to offset the inflationary effects of a temporary cut in interest rates. However, the reduction in the government debt/GDP ratio is of the order of 5 percentage points when spending is reduced. Therefore, the combination of falling interest rates offset by reduced fiscal expenditure would help Spain satisfy the Maastricht criteria concerning government debt and deficits.

A 10 per cent devaluation of the peseta

Chart 10 shows that the devaluation eventually feeds through to the price level resulting in a return to the original real exchange rate. However, it takes some time for this to happen--about 4 years for about two thirds of the increase in the price level to occur--and in the short term there are gains to output resulting from temporary gains in competitiveness. Chart 12 shows the effects upon total export volumes. As mentioned above, Spanish export volumes are quite price sensitive with a competitiveness elasticity of around unity. However, export volumes only register a maximum increase of 6 per cent in response to a 10 per cent devaluation as domestic prices rapidly respond to the rise in import prices. Eventually exports returns to base as the depreciation of the real exchange rate is eroded by domestic inflation.

Comparing simulation results with other countries

The output effects of a simulation of a 10 per cent devaluation are reported for the European economies in Anderton et al. (1992). Even though the simulations in the latter paper also include an interest rate cut the effects on output for Spain are somewhat larger relative to the UK, Italy and France. This is primarily due to the large price competitiveness elasticities embodied in the Spanish trade equations, particularly for exports. As a consequence of the higher activity, it seems that the price level for Spain remains higher, and for a longer period of time, relative to other European economies after a devaluation.(16)

The effects upon real GDP of a temporary 1 per cent cut in short-term interest rates is similar to that reported for the USA in the November 1992 Review. However, the US reaction to a monetary loosening tends to favour the supply-side (investment) more than the demand-side (consumption) relative to the Spanish response.

Table 6 of Barrell et al. (1993) shows that the fiscal multipliers for the major European economies are very similar to those reported for Spain in this paper. It seems that a fiscal shock equivalent to 1 per cent of GDP changes output by around a half to three quarters of a per cent for the first few years before declining to zero in the long run.

Conclusion

We have described recent economic events and policy in Spain and highlighted some of the problems facing the Spanish authorities. Our econometric model seems to capture several of the changes experienced by the Spanish economy over the last decade and consequently may be a useful tool for policy analysis. Simulations using the model were conducted with specific macroeconomic problems in mind. We found that monetary policy seems to have strong effects on output relative to fiscal policy. Spain's membership of the ERM and commitment to EMU implies interest rate convergence with the other participants. This suggests a substantial loosening of Spanish monetary policy which may have inflationary effects. The extent to which fiscal policy would have to be tightened to offset these effects would be rather severe. However, Spain's positive inflation diffential vis-a-vis the other ERM members will depress demand in a semi-fixed exchange rate regime given the magnitude of the competitiveness elasticities. Hence any required fiscal tightening will not be so harsh. Furthermore, the competitiveness elasticities of our model suggest that the recent devaluations of the peseta will mitigate the severity of the current Spanish recession.

NOTES

(1) This note is an abbreviated version of Anderton (1993). The latter contains full details of the econometric model used here.

(2) Larre and Torres (1992) provide a summary of structural problems and reforms relating to Spanish labour markets and wage and price determination. For a summary of financial liberalisation and the effects upon the Spanish economy see Galy et al (1993).

(3) In 1990 the male and female participation rates for Spain were 79 per cent and 42 per cent respectively. Whereas in the UK, for example, the figures are 89 per cent and 68 per cent.

(4) For a more thorough summary see OECD Economic Survey of Spain (1992) and Emerson (1988) which compare redundancy regulations between Spain and other countries. Bentolila and Saint-Paul (1992) compare the regulations governing temporary contracts in Spain with other countries.

(5) However, Bentolila and Saint-Paul (1992) argue that the EPP's introduced much greater labour flexibility by decreasing firing costs. Hence the magnitude of the response of employment to cyclical fluctuations may have increased since 1986.

(6) See Chan-Lee et al (1987) for an international comparison of replacement ratios.

(7) Although compositional changes in employment may have decreased average wages the effects upon wage setting of flexible labour contracts may be far more complex. For example, Dolado and Bentolila (1993) claim that the threat of unemployment to permanent workers is lower if they are protected by a buffer of temporary workers with lower firing costs. Hence this will raise the wage demands of permanent workers. They calculate that a percentage point increase in the proportion of temporary employment raises the growth rate of permanent workers wages by 1/3 of a percentage point.

(8) Larre and Torres (1991) point out that half the difference between Spanish tariffs and the common external tariff was wiped out in three years and virtually all quotas abolished.

(9) Cointegration tests reported in Caporale (1993) show that Spain's government debt to GDP ratio is not sustainable in the long run.

(10) This section is an abbreviated description of the model and only highlights some of the major equations. The full model contains many other equations and identities including complete models of the public sector, balance of payments and national income determination. See Anderton (1993) for full details or the NIGEM Model Manual (NIESR (1993)).

(11) A detailed explanation of this specification and application to Spain and other European countries is contained in Anderton and Barrell (1993).

(12) The export price elasticity is fairly high relative to other countries in NIGEM. However, other studies have estimated high elasticities for Spain. For example, Martin and Moreno (1991) estimate equations for Spain's industrial exports to the EEC disaggregated by commodity. They find export price elasticities as high as 3 for sectors such as textiles, leather and clothing.

(13) It is possible that this latter effect may be cyclical as temporary workers will probably account for most of the job losses in a downturn given that firing costs are much lower for fixed contract workers relative to those on permanent contracts. However, our wage equation stability tests did not indicate cyclical effects as the dummy variable for unemployment was not significant. In contrast, Bentolila and St. Paul (1992) estimate a model which contains significant cyclical effects.

(14) For a thorough analysis of the effects of EC membership and the European Single Market Act upon Spanish trade see Vinals et al (1990).

(15) See European Economy--One Market, One Money for a description of the possible requirements of fiscal policy within a monetary union.

(16) The effects of a devaluation for the major European economies are also analysed in Box 2 of the World Economy Chapter in the May 1992 National Institute Economic Review.

REFERENCES

Anderton, R., Barrell, R., in't Veld, J.W. and Pittas, N. (1992), 'Forward-looking wages and nominal inertia in the ERM', National Institute Economic Review, no. 141, August.

Anderton, R. (1993), 'Spain: evaluating the effects of policy using an Econometric Model', National Institute Discussion Paper (forthcoming).

Anderton, R. and Barrell, R. (1993), 'The ERM and structural change in European labour markets: a study of 10 countries', National Institute Discussion Paper no. 40.

Barrell, R., Sefton, J. and in't Veld, J.W. (1993), 'Interest rates, exchange rates and fiscal policy in Europe: the implications of Maastricht', National Institute Discussion Paper no. 44.

Bentolila, S. and Blanchard, O. (1990), 'Spanish unemployment', Economic Policy, 10, pp. 233-281.

Bentolila, S. and Saint-Paul, G. (1992), 'The macroeconomic impact of flexible labour contracts: an application to Spain', Centre for Economic Policy Research Discussion Paper no. 596.

Caporale, G. (1993), 'Some stochastic implications of the government budget constraint: an empirical analysis', paper presented at the Applied Economics Association Conference on Budgetary Policy Modelling, April 1993.

Chan-Lee, J., Coe, D. and Prywes, M. (1987), 'Microeconomic changes and macroeconomic wage disinflation in the 1980s', OECD Economic Studies, no. 8, Spring, pp. 121-57.

Commission of the European Communities (1990), 'One market, one money: an evaluation of the potential benefits and costs of forming an economic and monetary union', European Economy, no. 44, October.

Dolado, J. and Bentolila, S. (1993), 'Who are the insiders? Wage setting in Spanish manufacturing firms', CEPR Discussion Paper no. 754.

Emerson, M. (1988), 'Regulation or deregulation of the labour market: policy regimes for the recruitment and dismissal of employees in the industrialised', European Economic Review, no. 32, pp. 775-817.

Galy, M., Pastor, G. and Pujol, T. (1993), 'Spain: Converging with the European Community', IMF Occasional Paper no. 101.

Larre, B. and Torres, R. (1991), 'Is convergence a spontaneous process? The experience of Spain, Portugal and Greece', OECD Economic Studies, no. 16, Spring.

Larre, B. and Torres, R. (1992), 'Real and nominal convergence in the EMS: the case of Spain, Portugal and Greece', in Economic Convergence and Monetary Union in Europe, edited by R. Barrell, Sage publications.

Layard, R., Nickell, S. and Jackman, R. (1991), Unemployment: Macroeconomic Performance and the Labour Market, Oxford University Press.

Martin, C. and Moreno, L. (1991), 'Spain's industrial exports to the EEC: a panel data approach', CEPR Discussion Paper no. 521.

NIESR (1993), NIGEM Model Manual, November.

OECD (1989), 'Economies in transition; structural adjustment in OECD countries'.

OECD (1992), Economic Survey of Spain 1992.

OECD (1993), Economic Survey of Spain 1993.

Vinals, J. et al (1990), 'Spain and the 'EEC cum 1992' shock', CEPR Discussion Paper no. 388.