ارزیابی تعادل عمومی از شوک های خارجی و داخلی در اسپانیا

After many years of growth, the Spanish economy plunged into the most severe and prolonged recession recorded since reliable national accounts data have been available. The main goal of this paper is to quantify the effects of the external and domestic shocks that hit the Spanish economy in 2008–2009 by employing a disaggregated general equilibrium model calibrated to a 2000 SAM elaborated by the authors. External shocks are simulated by employing the neoclassical closure (private investment is determined by domestic and external savings) and the Keynesian closure (investment is exogenous). External and domestic shocks are also jointly simulated with the Keynesian closure. The results provide a good approximation to observed changes in key macroeconomic variables. Highlights ► Evaluation of reductions in exports, non‐resident consumption and investment. ► Investment is disaggregated by capital goods in the SAM we have elaborated. ► CGE models are very sensitive to the closure rules used to evaluate external shocks. ► The Keynesian closure is the most appropriate to analyze external shocks.

The main goal of this article is to present the results of some simulations performed with a disaggregated computable general equilibrium (CGE) model to assess the impact of external and domestic shocks that struck the Spanish economy in 2008–2009. From 1996 to 2007, Spain had been a model of success of the EU with gross domestic product (GDP) growing at a 3.7% average rate. Almost 8 million new jobs were created and the unemployment rate fell from 23.5% in 1995 to 8.03% in the third quarter of 2007.1 However, the economic situation changed suddenly in 2008. GDP grew just 0.9% that year and fell 3.7% in 2009, by far the largest drop recorded since the Statistics National Institute (INE) started publishing GDP estimates in the 50s. The number of unemployed grew at a disturbingly fast pace, and the unemployment rate hit 11.3% in the third quarter of 2008 and 17.9% a year later. During the boom years, financial institutions, nonfinancial firms and households financed a growing share of domestic and international investments with credits from non‐resident institutions. Although in 1996 there was a small current account surplus (0.2% of GDP), the difference between national savings and domestic investment as shares of GDP grew steadily in the following years (3.3% in 2000, 5.7% in 2004 and 10.0% in 2007). Gross external debt also rose at a fast pace in that period (341.1 thousand million euro in 1996, 784.5 in 2000, 1.29 billion in 2004 and 2.16 billion in 2007) and the net international investment position of Spain deteriorated without pause (− 113.3 thousand million euros in 1996, − 201.5 in 2000, − 505.5 in 2004 and − 822.8 in 2007). Any objective observer could easily foresee (Polo, 2005) that non-resident institutions would not be indefinitely willing to finance Spain's domestic2 and international investments. Put another way: it was clear that any slowing of the inflow of external finances would surely stop Spain's growth. During the second half of 2008, most advanced economies plunged into recession as financial instability spread out throughout the world. An immediate effect of the Great Recession was a sharp reduction in international trade. In the subsequent four-quarters, Spanish exports of traded goods and services fell 9.4% and non-resident consumption dropped 8.75%.3 The closure of international financial markets to domestic financial institutions and nonfinancial firms also put a sudden end to the investment process, which fell 13.07%. In this paper, we evaluate the impact of the decline in exports, non-resident consumption and private investment in the Spanish economy at the very beginning of the economic recession with a CGE model. These models have been widely used to assess the effects of changes in exogenous variables and parameters. Roberts and Zolkiewski (1996) used a static CGE model for Poland to estimate in 4% of GDP the effects of the drop in exports due to the end of the CMEA. Polo and Valle (2008), also with a static CGE model but under alternative closure rules, estimate in 5% of GDP the negative effects of a 10% permanent fall in non-resident consumption in the Balearic Islands. Furthermore, there are other recent papers that use a CGE multi-sector model to assess the effects of very different economic policies such us tax rate cuts on labor and VAT (Boeters et al., 2010 and Böhringer et al., 2005), or the impact of variations in the endowments of labor force (Learmonth et al., 2007), capital stock of multinationals (Latorre et al., 2009), etc. CGE models have been widely used since they can provide valuable quantitative insights into changes in the allocation of resources among sectors and major economic aggregates. However, the results drawn from the simulations are quite sensitive to the closure rule chosen, which determines the endogenous and exogenous variables in macroeconomic balances.4Rattso (1982), Dewatripont and Michel (1989) and Robinson (1991) described the main characteristic of the closure rules most commonly used in the literature and their effects on theoretical models and empirical results. The neoclassical closure rule assumes that the value of aggregate investment is determined by the value of domestic savings and the current account surplus of foreign sectors. In this setting, a negative external shock from the domestic viewpoint, such as a fall in exports, sets up an implausible investment boom in the economy. Under the Keynesian closure rule, aggregate investment is arbitrarily fixed as the sum of the value of private savings, the government deficit and the current account deficit that adjusts to equal the value of investment when there is an external negative shock. In this case, an investment boom is ruled out by hypothesis and the external shock raises the unemployment rate to reduce the value of private savings. Under these closure rules, if the exchange rate is fixed, foreign savings are flexible. The role of the exchange rate is a question that has received special attention in trade CGE models, Devarajan et al. (1993), as well as the effects of closure rules and the capital accumulation process, Francois et al. (1996), in dynamic CGE models.5 In the case of Spain, the empirical effects of closure rules in CGE models have been recently evaluated (Polo and Valle, 2008 and Polo and Viejo, 2009).6 In our article, the simulations are performed using both neoclassical and Keynesian closure rules. The paper is organized as follows. Section 2 depicts the main features of the model, and Section 3 presents the simulation scenarios and the results under the neoclassical closure. The results of the same simulations and two additional scenarios under the Keynesian closure are presented in Section 4. Finally, Section 5 gives some general conclusions.

The effects of external and domestic shocks that hit the Spanish economy in 2008–2009 have been analyzed using a general equilibrium model calibrated to a SAM elaborated by the authors for the year 2000. First, the fall in non-residents demand and exports of goods and services other than tourism were simulated using the neoclassical closure rule where investment is savings determined but the real wage is sensitive to the unemployment rate. Second, the impact of external and domestic (a sharp decline in private investment) shocks were estimated with standard Keynesian closure where private investment is exogenous and the unemployment rate endogenous. The effects of external shocks are extremely sensitive to the closure rule chosen in the model. An improvement in the current account of foreign sectors boosts external savings and domestic investment. Thus a negative external shock that reduces exports is compensated by an implausible investment boom. There is a reallocation of factors from the export sectors to investment-oriented sectors but public revenues and expenditures, unemployment, employment and GDP are almost unaffected by the shock. The simulation results are completely different when investment is fixed and unemployment adjusts to match the sum of the value of private savings, the public deficit and the current account surplus to the value of investment. Still negative external shocks also increase the current account surplus of both the EU and the ROW, although not as much as under the neoclassical closure. The main differences are in the remaining variables: prices and production decline substantially, especially in export-oriented sectors, increasing the public deficit and unemployment and reducing employment and GDP. The fall in employment (4.43%) and in real GDP (2.44%) are more consistent with observed values than the “nothing is happening here” message that derives under neoclassical closure. The data presented in Table 7 indicate that investment in capital goods fell dramatically from the third quarter of 2008 until the second quarter of 2009. The collapse of investment cannot be accounted for in a neoclassical framework but can be simulated under Keynesian closure. The effects of the negative domestic shock also reduce prices and production levels, especially those of investment oriented sectors, increases the public deficit and the unemployment rate and reduces employment and GDP. Those changes are even greater than in the external shock simulation. Of course, the fall in investment lowers the current account surplus of the EU and the ROW. It is worthwhile to compare the results of the joint simulation of negative external and domestic shocks under the Keynesian closure with observed changes in the main macroeconomic variables. Simulation results indicate that the unemployment rate increases 3.81 percentage points and GDP and employment drop by 4.43 and 2.44%. These latter values are very close to the variations reported by National Accounts for the same time span, 4.37 and 2.21%, respectively.16 The simulation of reductions in private investment also predicts a worsening of the EU and ROW surplus of 1.03%age points, which is below the 4.26 value registered in National Accounts. This discrepancy is motivated by the lower drop of imports in simulations and the smaller increase in the unemployment rate. Despite this evolution of the foreign current balances, however, the decline of imports suggests that economic policies aimed to favor exports and tourism services would be best for reducing the effects of the economic recession and for enhancing final demand. In summary, Keynesian closure seems the best for handling external shocks and the only way to account for investment shocks. In the first case, prices and production levels react to changes in final demand but the results are not distorted by an implausible investment boom. Moreover, the investment shock reduces income and savings and unemployment rates go up much more than under the neoclassical closure. As the number of unemployed increases, unemployment benefits and the public deficit increase. These results fit better the performance of the Spanish economy since 2008 and suggest that the selection of closure rule for a particular simulation deserves more attention in CGE modeling.