II Structural Reform and Macroeconomic Adjustment in Industrial Countries

Abstract

The record of economic performance in many industrial countries during the 1970s and early 1980s did not measure up to expectations. Part of the performance shortfall can, with hindsight, be attributed to deepseated and irreversible factors that made it unlikely that growth rates in many countries could return to the levels of the 1950s and 1960s: part of the shortfall can also be attributed to inappropriate demand management policies. But some of the difficulties in adjusting to the global oil shocks clearly emanated from dysfunction on the supply side of economies, the operation of which was not well understood. To improve its understanding of supply side factors, the Research Department has carried out a number of studies as background analysis for the World Economic Outlook. This paper, which is the fourth in the series, seeks to extend earlier staff work on structural issues by analyzing structural reform in a macroeconomic framework.2 It also reviews structural reforms in key areas and evaluates the evidence on the quantitative impact of structural measures. An annex presents the microeconomic foundations of structural reform.

The record of economic performance in many industrial countries during the 1970s and early 1980s did not measure up to expectations. Part of the performance shortfall can, with hindsight, be attributed to deepseated and irreversible factors that made it unlikely that growth rates in many countries could return to the levels of the 1950s and 1960s: part of the shortfall can also be attributed to inappropriate demand management policies. But some of the difficulties in adjusting to the global oil shocks clearly emanated from dysfunction on the supply side of economies, the operation of which was not well understood. To improve its understanding of supply side factors, the Research Department has carried out a number of studies as background analysis for the World Economic Outlook. This paper, which is the fourth in the series, seeks to extend earlier staff work on structural issues by analyzing structural reform in a macroeconomic framework.2 It also reviews structural reforms in key areas and evaluates the evidence on the quantitative impact of structural measures. An annex presents the microeconomic foundations of structural reform.

Structural Reform and Macroeconomic Theory

Definition of Structural Reform

The beginning of the 1980s marked a major turning point in the thinking of economic policymakers in the industrial countries. There was widespread disappointment with the ability of fiscal, monetary, and incomes policies to foster noninflationary growth with full employment during the previous decade. Indeed, since the early 1970s, many countries had been experiencing low growth, high and rising unemployment, accelerating inflation, and large budget deficits. The challenge facing policymakers in the early 1980s was therefore to restore, through a sustained commitment to fiscal and monetary restraint, an environment conducive to stronger economic performance over the medium term. This medium-term orientation of demand management policies implied that attempts to moderate short-run cycles in economic activity had to be resisted, even at the cost of some temporary loss in output and employment. At the same time, there was widespread recognition that improvements were needed in the way the industrial economies functioned—primarily through the adoption of “structural policies” to promote flexibility in factor and goods markets and more efficient economic adjustment.

The medium-term strategy recognized that many of the structural rigidities that were thought to impede growth were associated with the effects of earlier policy actions. Some of these actions sought to redistribute income with little attention to efficiency and budgetary implications. Others sought to reduce or delay the costs of adjustment for specific industries, regions, or labor groups in the face of successive external shocks throughout the 1970s. There also was a growing perception that the costs of these earlier actions were much greater than earlier recognized. In part this was because some of the external shocks turned out to be of a lasting nature, thus requiring adjustment rather than financing. Policymakers gradually realized that the adjustment process would require modifications in economic structure that would allow changes in market forces to be better understood by economic agents. The strong interest in structural reform therefore reflected, to a considerable extent, an erosion of confidence in the efficacy of government intervention. Instead, reliance on market forces was seen as the central principle in the economic organization of the industrial countries.

The new medium-term orientation of policies has highlighted the fact that, beyond the short run, economic performance depends primarily on the economy’s productive capacity. The emphasis has therefore been placed on the introduction of measures that increase the “flexibility” of the economy or raise its potential output.3 Such measures are called “structural” (or sometimes, “supply-side”) measures, and they include many traditional fiscal policy changes. Conceptually, structural measures can be classified into two broad categories: those that eliminate existing inefficiencies of resource use (so that output can be maximized, given existing resources) and facilitate more rapid adjustment to shocks and those that raise potential output, either by adding to existing productive resources—capital and labor—of the economy or by raising total factor productivity (TFP) (Hernández-Catá, 1988).4 In practice, of course, most structural measures are likely to involve effects of both types.

Structural measures of the first type include removal of rigidities that hinder the mobility of resources—such as institutional or regulatory barriers to the intersectoral (or international) mobility of labor—as well as elimination of price distortions, such as those created by tax wedges in various markets, including those that affect financial markets. Such measures may improve resource allocation and thereby raise potential output directly (while altering its sectoral composition), but they also facilitate adjustment following exogenous disturbances (e.g., technological innovations, changes in relative prices, or trade liberalization) by lowering transitional costs in terms of unemployed resources.

Structural measures of the second type include those that raise the level (or rate of growth) of potential output by increasing TFP or by raising savings and capital formation. Examples of such measures include efforts to promote research and development activity or to apply existing technical knowledge more effectively in production processes, and the elimination of tax measures that distort private saving and investment decisions.

Because structural measures often change the way individual markets work, they frequently are called “microeconomic” policies.5 It is important to remember, however, that they relate to fundamental macroeconomic concerns about the growth of potential output and the speed with which the economy weathers external shocks.

“Structural reform” emphasizes the removal of market rigidities and distortions. This is in stark contrast to earlier usage of the term, which was synonymous with “industrial” or “regional” policies. Such policies usually involved efforts to introduce incentives (i.e., “strategic” market distortions), in order to channel resources into specific sectors or regions with a view to correcting some perceived market failure. In contrast, the new view on structural reform implies a reduced role for government in the economy and an unfettering of private markets.

Domestic and International Macroeconomic Effects of Structural Reform

Effects of an Increase in Potential Output

This section focuses on the domestic and international macroeconomic effects of an increase in potential output resulting from a structural measure.6 The approach adopted here is based on the widely used two-country version of the Mundell-Fleming model. This framework is an analytically convenient vehicle for studying the international spillover effects of structural policies, and it also forms the theoretical basis for the Fund’s MULTIMOD model, which will allow the theoretical results to be illustrated with model simulations.

The analysis will proceed from a simple version of the Mundell-Fleming model, to more complicated versions.7 It is initially assumed that each of the two countries specializes in the production of one good and trades for the foreign good. The price level in each country is sticky, so that changes in aggregate demand lead to changes in both the level of output and the price level. Capital mobility is assumed to be perfect, implying that the expected nominal rates of return on all assets are equal. The price deflator for real money balances and real wages is taken to be the deflator for domestic output (or GDP deflator). Finally, expectations are assumed to be backward looking or adaptive. The issue to be examined is how an increase in potential output in one country (say the “home” country) affects the levels of output and prices and the external balances of both countries in the short and the long runs.8

In studying this problem it is natural to compare the effects of an increase in potential output in the home country, with those of more traditional demand policies. Those effects are clearly not identical although there are some important similarities. In the former case the productive capacity of the economy has increased; in the latter case it has not. This difference has very important microeconomic implications (and benefits) for the home country. From the perspective of the foreign country, however, disturbances in the home country are transmitted via changes in trade volumes, interest rates and the real exchange rate, i.e., the same variables that transmit the effects of changes in demand management policies. Given any pattern of these variables facing the foreign country, that country will be indifferent in the short run to whether this pattern resulted from structural reforms that raised potential output or from altered demand policies. In the long run, however, there would generally be a preference for the potential output channel since foreign potential output would also rise.9

(1) Demand-side Effects of an Increase in Potential Output

The effects of changes in potential output can be described in terms of the IS-LM and aggregate supply/demand framework, illustrated in Figure 1. This figure shows for a closed economy the short- and long-run effects of an increase in potential output with a given stance of demand policies. The initial equilibrium in Figure 1 occurs at point A where the aggregate supply curve in the top panel intersects the aggregate demand curve at potential output. The initial price level (P0) thus corresponds to a long-run, full-employment equilibrium where all nominal contracts have adjusted to the correct expectation of the price level. The increase in potential output resulting from a structural measure is shown by the shift from YP to YP. If the price level stayed at P0, the long-run, full employment equilibrium could be maintained. This fact is depicted in the figure by the new short-run aggregate supply curve Y1s intersecting the new level of potential output, YP at the initial price level, P0, At that price level, however, there is insufficient aggregate demand, so that the new short-run equilibrium (point B) occurs at a lower price level, P1, and at a level of output Y1 which is higher than the original level, but lower than the new level of potential output.

Figure 1.
Figure 1.

Demand-Side Effects of an Increase in Potential Output

Note: Subscripts denote time periods; O denotes original equilibrium (point A): I denotes short-run equilibrium (point B); z denotes the long-run equilibrium (point C). YS and YD are aggregate supply and demand curves. YP and YP are the original and new levels of potential output. P is the aggregate price level. Y is real output, and r is the interest rate (changes in which measure changes in both the nominal and the real interest rate since expectations of inflation are backward looking). IS is the locus of goods market equilibria: and LM, money market equilibria.It is assumed that money growth and potential output growth are zero, that there is no foreign sector and are no wealth effects, and that expectations are formed adaptively.Explanation: The diagram illustrates that the demand-side effects of an increase in potential output are equivalent to a monetary expansion in the short run (viz., LM1 is to the right of LM0 and IS1 = IS0), and equivalent to a balanced monetary and fiscal expansion in the long run (namely, both ISz and LMz are to the right of IS0 and LM0, respectively.In period 1 (i.e., the new short-run equilibrium) the level of potential output increases to YP’. Which shifts the aggregate supply curve to Y1s The new equilibrium occurs at a higher level of output and a lower aggregate price level. The IS-LM panel shows the effects of the lower price level on the demand side of the economy. With constant nominal money balances, the lower price level implies higher real money balances, which explains the shift in the LM curve to LM1. In the first period the IS curve doesn’t shift, so that the interest rate falls to r1. The increase in aggregate demand is explained by the fall in the real interest rate, which increases interest-sensitive spending which then has feedback effects on total spending.In the long run, the aggregate demand curve must intersect the aggregate supply curve at potential output (i.e., at YP). Thus, the IS-LM must also intersect at YP. The IS curve, however, will shift from its original position. Because of the permanent increase in real output, consumption will respond more in the long run than it does in the short run In addition, the increase in output also increases the productivity of capital, and thereby increases the optimal capital stock and investment. Finally, one might also expect that the permanent increase in output would lead to an increase in public consumption (i.e., government spending) as well as private consumption. AH of these factors would lead the IS curve to be further to the right in the long run. The diagram is drawn in the special case where the ratios of C/Y, I/Y, and G/Y are constant in the long run. In this case the aggregate demand expands to the new higher level of output at a constant interest rate in the long run. More generally, the IS curve shifts to the right and the interest rate rises above r1 in the long run.

The increase in aggregate demand results from the drop in the price level from P0 to P1 as shown in the bottom panel. The decline in the price level with no change in the nominal money supply increases the real money supply. The excess supply of money in turn causes the interest rate to fall. The interest-sensitive components of spending rise and a new demand-side equilibrium (point B) is achieved at a lower real interest rate and higher level of output. The new short-run, IS-LM equilibrium in Figure 1 is similar to the equilibrium that would result from an increase in the nominal money supply, even though the shifts in the aggregate supply and demand curves and the change in the price level would be quite different.10

In the long run, expectations of the price level and nominal contracts must adjust until a new full-employment equilibrium (point C) is restored at Pz and Yz (see note to Figure 1).11 The increase in output eventually will be reflected in the expectations of a rise in permanent output. As a result, demand is likely to rise more in the long run than in the short run. In order to determine how much long-term demand will rise, it is necessary to introduce additional assumptions about the determinants of spending in the long run. In the current context it is reasonable to assume that the long-run changes in demand will be proportional to the increase in potential output and permanent income. To locate the IS curve in the long run, it is necessary to specify the determinants of all components of spending, including government spending. The ISz curve in Figure 1 is drawn on the assumption that government spending in the long run is proportional to permanent income.12 Under these assumptions, the demand-side effects of the increase in potential output in the long run are basically similar to the short-run effects of a simultaneous monetary and fiscal expansion. (In both cases the IS and LM curves have shifted to the right; the effects on economic capacity differ, however.)

The spillover effects from the home country will affect the foreign country through trade volumes, the interest rate, and the real exchange rate. The impact on trade volumes is determined by real aggregate demand in the two countries and by the real exchange rate. From the foreign country’s perspective it is therefore the real exchange rate, the interest rate, and real demand in the home country that matter. These last two variables are determined by the intersection of the IS and LM curves in the home country. The final determinant of the spillover effects, the real exchange rate, must be analyzed in an explicit open economy context.

(2) Short-run International Effects: Adaptive Expectations

So far the analysis has shown that the demand-side spillover effects to the foreign country of an increase in potential output are similar in nature to the effects of a monetary expansion in the short run and of a balanced monetary-fiscal expansion in the long run. Now the analysis must be extended to include the determinants of the real exchange rate. As mentioned above, the simple version of the Mundell-Fleming model assumes perfect capital mobility, which implies (because of arbitrage), that the expected nominal rates of return on financial assets in the two countries must be equal. The real exchange rate must adjust to maintain macroeconomic equilibrium in both countries, while fulfilling the arbitrage condition.

To analyze the impact of structural measures on the real exchange rate, it is convenient to focus on how the initial effects of the increase in potential output in the home country affect the foreign country. Given perfect capital mobility and adaptive expectations, the fall in the home country’s interest rate leads to a fall in the foreign country’s interest rate. In addition, the rise in output in the home country will increase net exports in the foreign country. Both effects tend to raise aggregate demand and output in the foreign country.

The simple case of “negative transmission.” The impact on the exchange rate can now be inferred from the two initial effects on the foreign country’s money market. The fall in the interest rate and the initial rise in output tend to increase the real demand for money. In turn, excess demand for money in the foreign country puts upward pressure on the foreign interest rate. However, the foreign interest rate must satisfy the arbitrage condition, and therefore it cannot adjust to eliminate the excess demand arising solely in the foreign country’s money market. Thus, the upward pressure on the foreign interest rate leads to a capital inflow and a real appreciation of the foreign currency.

The effects of the real appreciation on the money market of the foreign country depend upon the particular analytical model used. In the original Mundell-Fleming model, it is assumed that the exchange rate has no direct effect on the aggregate price level.13 This assumption implies that the price level used to deflate nominal money balances is the domestic output (GDP) deflator, which does not depend directly on the exchange rale. Therefore, changes in the exchange rate will only influence the money market through changes in net exports which in turn reflect changes in aggregate demand and output in the foreign country. However, with the interest rate below its original equilibrium level, the excess demand for money can only be eliminated if foreign output falls below its original equilibrium level. The real exchange rate must adjust to lower foreign output and restore money market equilibrium in the foreign country. Specifically, the foreign currency must appreciate by enough to more than offset the initial rise in foreign country exports. The foreign currency appreciates until the fall in its net exports is larger than the rise in interest-sensitive spending, so that equilibrium in the money market is restored at a lower interest rate and a lower level of output. These effects are summarized in Table 1. The nominal depreciation of the home country’s currency follows from the necessity of a real depreciation in the home country and the falling price level (associated with the expansion of potential output).

Table 1.

Impact Effects of an Increase in Home Country Potential Output: The Simple Mundell-Fleming Model

article image

The effect on potential output in the foreign country is discussed in the subsection below dealing with the transmission of supply effects.

As noted at the outset, in many respects the short-run spillover effects of an increase in potential output are similar to those of an expansionary monetary policy.14 There is, however, one notable difference. A monetary expansion by the home country leads to an increase in that country’s price level, whereas an increase in potential output leads to a decrease. This difference has no effect on the international transmission of the effects of the two policies: residents of the foreign country are not concerned with the home country prices of home country goods, but rather with the foreign country prices of home country goods. In other words, it is the real exchange rate that is important. Trade volumes and real output, however, depend on real magnitudes. As a result, the difference in the behavior of the nominal exchange rate and the home country’s price level under the two types of policies does not lead to any differences in real effects.15

“Positive transmission.” The similarities of the results of structural measures in Table 1 to the results of a monetary expansion include one of the key results of the simple Mundell-Fleming model, that is. That a monetary expansion in the home country may lead to a fall in output abroad (“negative transmission”). This strong and somewhat counter-intuitive result has attracted a great deal of attention and analysis as many economists have suspected that this was a reflection of the particular assumptions of the simple Mundell-Fleming model, a suspicion that was confirmed by Argy and Salop (1979).

In Argy and Salop’s analysis, households are concerned about the consumption value—as opposed to domestic output value—of their money balances. With this assumption, the correct deflator for money balances is the consumption deflator as opposed to the GDP deflator. As a result, changes in the exchange rate can directly affect the foreign country deflator for money balances.

Real appreciation of the foreign exchange rate has two effects on the excess demand for money in Argy and Salop’s analysis. As in the simple Mundell-Fleming case, the real appreciation lowers net exports and therefore aggregate demand and output in the foreign country. However, the real appreciation also lowers the consumption deflator, thereby increasing real money balances. The net effect on output in the foreign country-becomes ambiguous. If the effect of the real appreciation on real money balances in the foreign country is relatively small, the new equilibrium will involve lower output in the foreign country as shown in Table 1. If the impact of the real appreciation on real money balances is large enough, however, money market equilibrium in the foreign country will involve a higher level of output (hence, a “positive transmission” effect). The result of positive transmission of the potential output increase in the home country to foreign output is more likely to occur the greater the degree of openness of the foreign economy. The qualitative results for the modified Mundell-Fleming model are summarized in Table 2.

Table 2.

Impact Effects of an Increase in Home Country Potential Output: The Extended Mundell-Fleming Model1

article image

The extended Mundell-Fleming model assumes that exchange rate changes have a large effect on the consumption deflator, which is used as the deflator for money balances.

The sign of the change in net exports in Table 2 is ambiguous. In the extended model, there are two offsetting effects working on net exports. The expansion of demand in the home country will increase net exports in the foreign country, whereas the real appreciation of the foreign exchange rate will tend to lower net exports. In addition, interest-sensitive spending generally tends to increase in the foreign country. Thus, when foreign output falls (as in Table 1), the fall in foreign net exports more than offsets the rise in interest-sensitive spending. When foreign output rises, however, the change in the sum of interest-sensitive spending and net exports rises (Table 2). Foreign output could rise because both interest-sensitive spending and net exports are rising, or because interest-sensitive spending is rising by more than the decline in net exports in the foreign country.

The ambiguity of the change in net exports has important implications for external and internal balance. A similar ambiguity occurs with expansionary monetary policy that is positively transmitted. Boughton (1988) and Genberg and Swoboda (1987) have argued that because fiscal policy has a more predictable and significant effect on the current account, fiscal policy should be aimed at achieving external balance while monetary policy should be aimed at internal balance. These same arguments would imply that policies designed to stimulate potential output should be more concerned with internal than with external issues.

Transmission of Supply Effects. The previous section examined several aggregate demand channels through which changes in potential output are transmitted internationally. This section examines transmission effects through the supply side—that is, how an increase in potential output in the home country may lead to an increase in potential output in the foreign country.16

International transmission effects of an increase in potential output may also arise because of the effect of exchange rates on labor supply decisions. Wage earners care about the value of their nominal wages in terms of the goods they buy, which include imports; they therefore deflate nominal wages by the consumption deflator. Firms, however, care about the real product wage, which is the nominal wage deflated by the price of output or GDP deflator.

In this situation an appreciation of the foreign country’s exchange rate causes a decrease in consumer prices, which under conditions of real wage flexibility would lead to lower nominal wages. From the firms’ perspective, a real appreciation will thus increase the supply of labor. As a result, the real product wage at full employment falls and foreign potential output rises. This result, which does not affect the qualitative results in Table 2, implies that some of the benefits to the home country stemming from the increase in its productive capacity eventually are shared by the foreign country.

The increase in potential output in the foreign country was caused by the real appreciation of its currency, the counterpart of which is a real depreciation in the home country. But as the home country wage earners face higher import prices, they will demand higher wages and reduce their labor supply. This channel of international transmission therefore implies that part of the initial increase in potential output in the home country is lost to the foreign country. Of course, if the degree of real wage flexibility differs across countries, the ultimate impact on world potential output may be either greater or smaller than the initial expansion of capacity in the home country.

An additional channel through which supply effects could be transmitted abroad is directly through transfers of technology. For example, technical progress that raises total factor productivity and potential output in the home country could be embodied in capital goods (e.g., computers) that are exported, thus raising productivity and potential output in the foreign country.

(3) Long-Run Effects17

The similarity between the effects of an increase in potential output and of a monetary expansion breaks down in the long run. In virtually all macroeconomic models, a one-time change in the money supply is neutral (i.e., has no effect on any real variable) in the long run. Thus, an increase in the nominal money supply in the home country will ultimately lead to an equal percentage increase in the home country’s GDP deflator and no change in its real exchange rate or any other real variable. In contrast, a one-time increase in potential output in the home country will have real effects as actual output converges toward potential output in the long run.

The more interesting long-run results of an increase in potential output on the demand-side concern the impact on the real exchange rate and the trade balance. In order to determine these long-run results, the determinants of spending in the long run must be specified. As noted previously, the simplest specification is to assume that the ratio of any component of spending to real output is constant in the long run. This specification of long-run demand implies that absorption and import demand in each country would be a constant fraction of output if the interest rate and real exchange rate are unchanged.

Such a scenario, however, cannot describe the new long-run equilibrium following an increase in potential output, because output and demand have risen in the home country, and therefore imports have also risen. This would imply an excess demand for goods in the foreign country (with a corresponding excess supply in the home country), which would cause the exchange rate (and perhaps the interest rate) to change. To eliminate the home country’s excess supply of goods its currency would have to depreciate. In the long run, an increase in home country potential output is therefore likely to result in a real depreciation of its currency. In contrast, with an expansionary monetary policy there are no real exchange rate effects in the long run.

(4) Rational Expectations

The similarity between the demand-side effects of a monetary expansion and an increase in potential output breaks down even in the short run when forward-looking, rational expectations are introduced. Both developments increase demand in the short run by raising real money balances. In the rational expectations case, however, market participants anticipate that a long-run real depreciation of the home country’s currency will result from an increase in potential output, while in the case of a monetary expansion the real exchange rate will return to its original value. Under rational expectations, this difference in long-run impact shows up immediately.

A rise in potential output and a monetary expansion also have different long-run effects on permanent income, the optimal capital stock, and the debt-GNP ratio among other variables. Under rational expectations these differences in long-run effects also have immediate and significant effects. These effects can be illustrated on the basis of simulations derived from an empirical rational expectations model, such as MULTIMOD.18

(5) Simulation Results

The major findings of the two simulations that are reviewed here can be summarized as follows: (i) the MULTIMOD simulations are qualitatively consistent with the simple theoretical analysis above (Table 1); (ii) the quantitative impact on current accounts of structural policy measures that raise potential output is small; (iii) the simulations confirm the similarity between the short- and medium-run) international transmission of the effects of an increase in potential output and the transmission of the effects of more traditional demand policies. As discussed earlier, however, the difficulty of linking (both analytically and empirically) specific structural reform measures to the subsequent rise in aggregate potential output underlines the stylized nature of the simulation exercise and suggests caution in using the results.

Table 3 lists the simulation results for an increase in potential output of 1.5 percent in all industrialized countries other than the United States. Thus in terms of the theoretical discussion, the United States is the foreign country and the rest of the industrial world is the home country.19 The results are given in terms of the change in the value of each variable relative to the baseline levels. For the current account, the effect of the increase in potential output on the aggregate of the rest of the world can be inferred by changing the sign of the change for the United States. The effects on the home country price level, real output and interest rate, can be inferred from any of the countries that comprise the home economy. (The behavior of these variables is not completely uniform among the home countries, which indicates there are significant distributional effects working among the countries experiencing the increase in potential output.) Finally, the results for the first four years will be taken as conforming roughly to the “short run” of the theoretical discussion, and the later years indicate the adjustment toward the long run.

Table 3.

Simulated Effects of an Increase in Potential Output in All Industrial Countries Except the United States, 1988–99

(Percentage deviation from baseline unless otherwise noted)

article image
article image

A positive value indicates an appreciation.

The short-run simulation results for the United States in Table 3 are qualitatively similar to the theoretical predictions for the foreign country in Table 1. In the short run there is negative transmission of the increase in potential output in the home country. Interest rates fall in the United States, but the effect of this fall on spending is more than offset by the real appreciation of the dollar.20 The short-run effects on current accounts are small in relation to the effects on output. Over the medium and longer term, the qualitative results gradually change as demand in the United States increases and eventually reflect positive transmission and an improvement in the U.S. current account. These longer-term results are due to the outward shifting of the home countries’ IS curve as expectations of permanent income in the home countries increase. Indeed, by the mid 1990s the fall in prices and interest rates in the home country diminishes and eventually reverses itself. For example, in Germany during the first four years of the simulation, the GNP deflator falls by over 3.5 percentage points and the long-term interest rate by almost 70 basis points. In the four-year period from 1993 to 1996, the GDP deflator falls by only 1 percentage point and the long-term interest rate rises. This suggests that some of the increase in demand is coming from IS shifts as opposed to the initial LM shifts. Although MULTIMOD has a relatively complicated economic structure, the simulation results of Table 3 are broadly consistent with the simple theoretical results of Table 1 in the short run and with those of the above discussion of the long-run effects.

Table 4 provides simulation results of a monetary expansion.21 On the basis of the impact on real GNP in the United States, the immediate impression is that the effects of a monetary expansion are transmitted much more quickly than a rise in potential output. By 1990 and 1991, however, the magnitudes of the effects are about the same. The major difference in the two sets of simulations is in terms of the impact on nominal exchange rates. In the monetary expansion the U.S. currency appreciates by over 5 percent, whereas in the case of the increase in potential output the nominal exchange rate appreciates by less than one half of 1 percent. This difference, however, solely reflects the fact that in the monetary expansion the home countries’ price levels are rising, whereas in the case of the increase of potential output the home countries’ price levels are falling. As noted in the previous section, this difference between the effects on the home countries’ price level is irrelevant for international effects. Instead, what matters is the real effective exchange rate. By 1990–91 the change in the real effective exchange rate of the U.S. dollar is in the 1–2 percent range for both simulations.

Table 4.

Simulated Effects of an Increase in Money Supplies of All Industrial Countries Except the United States, 1988–99

(Percentage deviation from baseline unless otherwise noted)

article image
article image

A positive value indicates an appreciation.

Some of the difference in the timing of the effects of the two policy changes is due in part to overshooting. For the monetary expansion the nominal and real appreciations are largest in the first year, whereas in the case of the increase in potential output the change in the real exchange rate is modest in the first year and slowly increases in the short run. The greater overshooting in the case of the monetary expansion therefore explains in part the more immediate transmission to foreign output and the current account.

Structural Policies and Macroeconomic Flexibility

In addition to affecting the level of potential output, structural policies can affect the way an economy reacts to shocks by increasing “flexibility.” Structural policies that increase flexibility include efforts to reduce wage rigidities or increase openness.22 Not all types of flexibility are desirable from a macroeconomic perspective, however. For example, while improvements in real wage flexibility are always desirable, it is possible for nominal wage flexibility to be too high in the short run.

(1) Short-Run Flexibility and Macroeconomic Adjustment

Short-run nominal and real wage flexibility are measures of the responsiveness of wages to short-run disturbances.23 Short-run real wage flexibility is measured by the responsiveness of the real wage to unemployment.24 Perfect real wage flexibility would correspond to the case where wages adjust to their “warranted” level, where the unemployment rate is kept at its long-run equilibrium rate. Nominal wage flexibility is measured by the elasticity of the nominal wage with respect to the aggregate price level. Perfect, short-run nominal flexibility would correspond to an elasticity of 1, so that any increase in the price level would give rise to an equiproportionate increase in the wage rate, with no change in the real wage.

Many analysts associate short-run rigidity of wages with the periodic nature of wage bargaining. With contracts of a fixed duration, contracts that are currently scheduled for renegotiation can be adjusted in response to current economic disturbances. Contracts are negotiated with fixed durations because of bargaining costs. One way of minimizing bargaining costs while still increasing wage flexibility is the use of contingent wage contracts. In contingent contracts, the formula for determining the wage is negotiated, rather than the level of the wage. A well-known example of a contingent wage contract is the union wage contract that includes a cost-of-living (COLA) clause. In such a contract, the actual level of the wage depends on the level of consumer prices during the life of the contract. COLA clauses increase nominal flexibility, at the expense of lower real wage flexibility in the short run.25

An economy with perfect short-run nominal wage flexibility is protected from the effects of demand shocks on unemployment, but not from the effects of supply shocks. Perfect short-run nominal flexibility makes the aggregate supply curve vertical, but not necessarily at potential output. For example, consider an economy with perfect, short-run nominal flexibility, which is subject to an adverse supply shock that lowers the warranted real wage. In order to preserve full employment, the actual real wage must fall. Perfect nominal flexibility, however, makes the real wage constant for any price level. As a result, the aggregate supply curve is vertical at a level of output below potential,26 and unemployment increases until nominal wages are renegotiated to reduce the real wage.

In contrast, perfect nominal flexibility will tend to protect an economy from the real effects of a demand shock on unemployment. In addition to purely cyclical influences on the demand for labor, demand shocks may increase unemployment by changing the price level, thereby bringing the real wage out of line with the warranted real wage. With perfect short-run nominal flexibility, however, changes in the price level no longer affect the real wage. Put another way, in the absence of a supply shock, perfect, short-run nominal flexibility makes the aggregate supply curve vertical at potential output.

With perfect, short-run real wage flexibility, the degree of nominal flexibility is irrelevant. With less than perfect real wage flexibility, the desirable degree of short-run nominal flexibility depends on the relative importance of demand and supply shocks. Gray (1976) has shown in the context of a closed economy that the optimal degree of nominal wage flexibility leads to a positive wage elasticity which is closer to unity the more prevalent are demand shocks relative to supply shocks. Aizenman and Frenkel (1985) have extended Gray’s analysis to a model of a single open economy and analyzed how the optimal degree of nominal wage flexibility depends on the demand policies being pursued.

(2) Medium-Run Flexibility and Macroeconomic Adjustment

So far the discussion of flexibility has been concerned with the short-run response of the economy to shocks. In the absence of perfect short-run real wage flexibility, however, the initial response will not preclude adverse effects on unemployment. It is also important, therefore, to analyze how an economy adjusts to eliminate unemployment that results from past shocks. It is convenient to call this latter type of flexibility medium-run, real wage flexibility. The higher the degree of medium-run, real wage flexibility, the more quickly will unemployment be eliminated.27

The causes of medium-run real wage rigidities are not well known, but in light of recent persistence of European unemployment, a number of theories have been proposed. Taylor (1980) argues that medium-run rigidities may result from the wage bargainers’ concern with relative as well as real wages. In an economy with high unemployment and staggered wage negotiations, bargainers who focus on relative wages will not negotiate their real wage to the warranted real wage because such a negotiation would lower the relative wage. Thus, high unemployment may persist even after all contracts have been renegotiated.

Another set of theories relies on the relative power of “insiders” vs. “outsiders” (Blanchard and Summers (1986) and Lindbeck and Snower (1988)).28 These types of theories explain the persistence of deviations of the real wage from the warranted real wage, for example on the basis of senior union members pushing for wage increases even though this would lead to unemployment among junior members. A related argument based on empirical analysis (Bruno and Sachs (1985)) argues that unions often fail to lower their real wage goals after a decrease in trend productivity growth. Finally. Bean and Layard (1988) argue that the real wage may be sticky because of “outsider ineffectiveness.” In their view, the long-term unemployed, perhaps through the loss of job skills and motivation, fail to exert any pressure to lower wages and restore high employment.

Regardless of which of these theories proves to be correct, the international consequences of differences in medium-run. Real wage flexibility are important. Consider a shock that causes an equal rise in unemployment in all countries. Further assume that the real wage temporarily has risen over the warranted real wage in all countries (so that there is a rise in structural unemployment) and that one group of countries has a high degree of medium-run, real wage flexibility and another group has very low real wage flexibility. In the former group of countries, real wages will fall rapidly in response to the rise in unemployment, and these countries will return to high employment and high levels of demand quickly. In the low flexibility countries, a return to full employment with price stability requires a reduction in real wages, which takes time. Thus, differences in medium-run flexibility are likely to lead to quite different cyclical patterns of adjustment.

The delay in the fall of real wages puts the obvious costs stemming from global disturbances on the low flexibility countries. However, it may also place an external adjustment burden on both groups of countries. If the shock takes the form of an increase in demand in the high flexibility countries, caused by a fiscal expansion, then these countries may experience an appreciation of the exchange rate and a widening trade deficit. Alternatively, if the increase in demand is brought about by an increase in the real money supply, then the impact on the trade balance is likely to be small. In the latter case the demand effects on the trade balance are offset by a real depreciation of the currency. However, there may still be external costs in this situation to the extent that individual countries may be affected by the changing patterns of demand and relative prices. In either case, the difference in the degree of wage flexibility adds an external adjustment burden to the internal adjustment costs caused by differences in the stance of demand policies.

Flexibility and Macroeconomic Adjustment: Simulation Results

The theoretical discussion has drawn a distinction between the short-run effects of increased flexibility and the longer-term effects as contracts are renegotiated. Because MULTIMOD is a model of annual observations, and some contracts will be renegotiated during the year in which a shock occurs, the results of the simulations cannot be broken down into these two categories very easily. Even the shortest possible simulation period will thus contain some of the effects of what was described as “medium-term” flexibility in the theoretical discussion. However, the longer the period of the simulation, the more the results correspond to the effects of medium-term flexibility. To illustrate the effects on international macroeconomic adjustment, two MULTIMOD simulations of a change in the degree of real flexibility were implemented (Table 5). In both cases, the macroeconomic disturbance is a reduction in government spending in all industrial countries.29 The two simulations show the response of the world economy in the cases of high and low real flexibility.30

Table 5.

Macroeconomic Adjustment and Real Flexibility

(Percentage deviations from baseline levels)

article image

Changes in percent of baseline GNP.

One result in Table 5 is particularly surprising: the change in output is smaller in the United States in the less flexible case over both time periods, and the same relationship holds true in Germ an the over the shorter time period. In terms of changes in real output, greater flexibility does not appear to be stabilizing. Notice, however, that none of the differences in the cumulative changes in output between the high and low flexibility cases is ever very large. The surprising result of output stability is apparently due to the forward-looking expectations in the MULTIMOD model. With forward-looking expectations, the greater price flexibility that results from increased real wage flexibility exacerbates disturbances in aggregate demand. Greater real flexibility is always stabilizing in terms of aggregate supply. Table 5 suggests that the effects of price flexibility on aggregate supply and demand in the MULTIMOD model are roughly offsetting in the short run.

In the simulations reported in Table 5, the aggregate demand effect works through forward-looking expectations of inflation. Because the price level falls by more in the high flexibility case, forward-looking expectations of inflation decline more in the short run. As a result, real interest rates rise by more, and thus the “crowding in” effect following the fiscal contraction is attenuated. The possibility that increased price flexibility could destabilize aggregate demand has been long recognized in the theoretical macroeconomic literature (although not in the literature on the optimal degree of wage flexibility, such as Gray (1976)). Recently it has been argued that this theoretical possibility has empirical relevance for the U.S. economy.31 While MULTIMOD seems to be supporting this last contention, it should be noted that the destabilizing effect is confined to the short run, and that the net change in real output over any time span is small.

While the gains from increased real flexibility in terms of real output stability seem to be small, there are significant gains in terms of inflation control. By comparing the output and inflation effects, it is possible to infer the output costs that may be associated with policies of demand restraint aimed at lowering inflation. By comparing the high flexibility to the low flexibility columns in Table 5, there is clearly evidence that increased flexibility can substantially lower the “sacrifice ratio.” In other words, the simulations show that for about the same cumulative reduction in output, the price level would fall by much more in the case of high real flexibility. The results in Table 5 also have important implications for the external costs of adjusting to a common macroeconomic shock. These external costs result from differences in the degree of flexibility among countries that leads to different cyclical responses to a common shock.

After four years all three countries have experienced a real appreciation of their currencies (the currencies of the other industrial countries have depreciated, which is not shown in Table 5) with the real appreciation in Japan and the United States the largest. This result holds true for both the low flexibility and the high flexibility column. In addition to the quantitative differences, there are qualitatively different changes in current account positions. The United States experiences a reduction in its current account deficit, while Germany’s surplus falls; there is very little change in the Japanese current account. At least part of the differences in the behavior of the current accounts reflect different speeds of adjustment to the demand shock among countries.

In the longer run (after eight years) the change in real GNP is at or near its peak in all three countries, but the value at this peak is much different. In particular, Germany and Japan experience much more of a change in the price level than in real output. In the case of Japan, the real appreciation broadly offsets the effects of the decline in output and demand on the current account. In the United States, however, where output and demand are declining the most, there is a substantial improvement in the current account. In Germany, the real appreciation is more than enough to offset the demand effects, so that there is a marked decline in the current account surplus. Overall, these simulations confirm the theoretical result that differences in medium-term flexibility can add significant external costs to the internal costs of adjusting to common macroeconomic shocks.

Structural Reform in Practice

This section builds on the theoretical framework for analysis of structural policies developed above and examines how structural reforms might affect macro-economic performance and international adjustment in practice. The section reviews examples of some of the important initiatives taken over the last several years in the area of structural reform, with the emphasis on the types of reforms common to most industrial countries. Specific measures in industrial countries are mentioned as examples within each category; however, these examples do not aim at providing an exhaustive inventory of structural measures taken so far.32 The presentation will distinguish between aggregate demand effects and aggregate supply effects. Table 6 presents an overview of the analysis.

Table 6.

Structural Policies and Macroeconomic Relationships

article image
article image