This paper constitutes a preliminary effort to examine whether, and how, the existing macroeconomic indicator system used in the context of the World Economic Outlook (WEO) might be expanded to include structural policies and performance.2 The paper begins by assessing the potential role structural indicators could play in the current system and by identifying the desirable features of possible new indicators. It then discusses the appropriate scope of structural indicators and reviews their applicability for the domestic economy as a whole, for individual factor markets, and for goods markets and trade protection. The final section contains a brief summary and conclusions of the paper.

This paper constitutes a preliminary effort to examine whether, and how, the existing macroeconomic indicator system used in the context of the World Economic Outlook (WEO) might be expanded to include structural policies and performance.2 The paper begins by assessing the potential role structural indicators could play in the current system and by identifying the desirable features of possible new indicators. It then discusses the appropriate scope of structural indicators and reviews their applicability for the domestic economy as a whole, for individual factor markets, and for goods markets and trade protection. The final section contains a brief summary and conclusions of the paper.

Conceptual Issues

Domestic Effects of Structural Policies

While monetary and fiscal policies operate primarily through the demand side of an economy, structural policies operate primarily through the supply side. Structural policies impinge on the supply side of the economy, first, through influencing the level of potential output, and, second, by affecting the flexibility with which an economy responds to economic shocks. The link between structural policies and potential output is relatively straightforward. By removing economic distortions, structural policies will normally lead to greater economic efficiency, thus raising the level of potential output and permitting a temporary rise in the growth rate. However, the benefits and efficiency gains from removing economic distortions may not be independent of the order in which they are removed.

Structural policies can also affect the way an economy responds to a shock. For example, the relaxation of regulations governing dismissals can make firms more flexible in both hiring and firing. As a result, the national economy may be able to move resources more quickly between sectors in response to changes in comparative advantage. In addition, greater flexibility in hiring and firing could lead to greater sensitivity of wages and prices to current economic conditions.

Both of the major ways in which structural policies operate on the economy suggest that policy making concerned with structural issues must have a longer time frame than traditional macroeconomic policy making. Potential output responds more slowly to structural policies than does actual output to monetary and fiscal policies. The adjustment of the economy to shocks cannot even be measured at a moment in time, but only over time.

While structural policies can have important macro-economic effects, they typically impinge directly on individual markets and most clearly affect participants in the market concerned. For example, eliminating a dis-tortionary agricultural subsidy will, over time, cause resources to move to more highly valued uses and thereby raise potential output. The increase in potential output, however, may not be the most significant aspect of the policy, particularly in the short run. The removal of the subsidy will lower farm income and raise food prices. Farmers will clearly be hurt by the policy, and consumers will pay higher food prices. However, in the long run consumers of nonagricultural goods will benefit, as will taxpayers.

International Spillover Effects

An important aspect of structural policies in an international context is the particular way in which the flexibility of the supply side of a national economy can cause international spillover effects. For example, consider the impact of an equiproportionate worldwide reduction in nominal demand on trade balances and exchange rates. If all labor markets possessed the same degree of flexibility (so that the aggregate supply curves had the same slopes in all countries), the demand shock would cause the same proportional reduction in the price level and real output in all countries. Thus, there would be no pressure on exchange rates or changes in the trade balance. In contrast, however, if countries have differing degrees of labor market flexibility, the countries with the greatest flexibility will experience the largest decline in its price level and the smallest decrease in output. In this case, there will be pressures on exchange rates and trade balances. The differences in the flexibilities of the labor market cause the spillover of external adjustment costs in response to a macroeconomic shock.

The flexibility of labor markets may also influence the effects of more microeconomic disturbances. Taking advantage of changes in comparative advantage requires shifting resources between sectors of the economy. During the transition period, some people are unemployed as a result of the changes in relative demands and prices. More flexible labor markets can allow countries to adjust more quickly and economically to changes in trade flows, and thus give rise to less political pressure for protectionism.

Thus, structural policies that make markets more flexible in response to macroeconomic shocks and disturbances to trade flows have important spillover effects. Indicators of structural policies that affect market flexibility, and intermediate variables of the degree of market flexibility have the potential to be very useful to international policy makers.

It is assumed in this study that structural policies that raise potential output have the most potential for improving national economic welfare over the long run, independently of their international spillover effects. In addition, however, there are at least three types of benefits from raising potential output that relate to international interactions. First, the more affluent a country, the easier it may be for that country to undertake needed macroeconomic adjustment policies. Second, to the extent that structural policies affect output and inflation they will have implications for balance of payments and exchange rate developments. Third, given the growing integration of the world economy, removing distortions in a market in one country may have effects on the international allocation of resources.3

International cooperation may be much more difficult to achieve on structural objectives than on macroeconomic goals. In formulating performance indicators for structural policies care must be taken not to impose uniform views on the proper balance of equity versus efficiency (Is there a unique level of unemployment benefits appropriate for all countries?), or on the proper balance of market goods versus non-market goods (Is there a length to the work week that is appropriate for all countries?). These considerations imply that performance indicators should be focused on the macroeconomic effects of structural policies.

Attributes of Structural Indicators

The foregoing discussion suggests a number of attributes for desirable structural indicators for use in international discussions of the medium-term sustainability and desirability of economic policies.

Performance indicators. The ultimate objectives of structural policies are the same as those of monetary and fiscal policies (output growth, full employment, price stability, current account and trade balance viability). Macroeconomic performance indicators of structural policies should attempt to capture two additional dimensions. First, an attempt should be made to measure effects on the level or the growth rate of potential output; second, indicators should seek to capture the responsiveness of an economy to economic shocks. It is difficult, if not impossible, however, to measure changes in these variables over short periods of time. For this reason, assessments of the macroeconomic impact of structural policies can be made at less frequent intervals than for demand management policies.

Policy indicators. Structural policy indicators should be variables that are closely influenced by the authorities’ actions and that are causally linked with the performance indicators discussed above. In selecting policy indicators emphasis should be placed on policies that have important spillover effects on trade and capital flows. This suggests emphasizing indicators of policies that affect the responsiveness of supply to macroeconomic and microeconomic shocks, and indicators of policies that affect distortions in highly integrated resource markets.

Intermediate variables. Intermediate variables should be available in a timely fashion and should ideally be linked via accepted behavioral relationships to policy and performance indicators. Finally, care must be taken that the interpretation of the indicators is not highly sensitive to the particular economic model linking the indicators.

Selecting Structural Indicators

Macroeconomic Versus microeconomic Focus

This section explores, in a preliminary fashion, how the existing WEO macroeconomic indicator system might be expanded to include structural indicators. While it seems safe to assume that structural, or microeconomic, issues are relevant for policy coordination only to the extent that they have important macroeconomic implications,4 this macroeconomic focus is not easy to achieve. The purpose of this subsection is to identify some of the difficulties involved.

The first such difficulty is analytical. In many cases, empirically estimated models do not exist that reliably link structural policies to macroeconomic objectives. However, microeconomic, partial equilibrium models are often available that describe many important implications of structural measures for individual markets. A second difficulty is that, with the important exception of labor market reforms in the context of excess capacity, past empirical work suggests that the quantitative effects of many structural reforms on the macroeconomic performance indicators are likely to be relatively small in the short run, compared with the effects of substantial changes in macroeconomic policies. Longer-run effects are clearly more significant, but not easy to capture econometrically. And benefits in the form of greater flexibility in the face of shocks, or greater effectiveness of other macroeconomic policies, are most difficult of all to measure.

Choice of Alternative Indicators

No structural indicators are at the same time simple to comprehend, easy to compute, and closely related to the real economic concepts that are relevant for analysis. The choice between indicators seems to come down to simple, though flawed, indicators, or complex, though improved, indicators.

A useful taxonomy is to think of structural policies as influencing the efficiency with which the economy, or in the first instance individual markets, function, and to classify structural measures according to the markets they affect. Considering the likely magnitude of macroeconomic effects that can be associated with microeconomic distortions in particular markets, one would expect factor markets to be particularly important. But substantial distortions also exist in some product markets. The labor and capital markets cut across all goods markets and play a primary role in determining the distribution of income in an economy. In addition, in helping to determine the unemployment rate and the interest rate, the factor markets are directly related to macroeconomic performance.

For example, real wage flexibility is needed for maintaining adequate levels of international competitiveness and employment. Within the factor markets, there is added emphasis on labor markets because these hold the most promise for structural reform. Capital and financial markets are also important, but in the industrial countries these markets have already benefited from widespread deregulation, reform, and integration, so that the remaining distortions are probably small.5

Aggregate indicators

The only way an economy’s output can be lifted on a sustainable basis in the medium term is for its supply capacity, that is, its potential output, to expand. Discretionary aggregate demand policies can help to keep actual output near potential, but they cannot raise the rate of growth of output above this level for long. The best available aggregate indicators that summarize structural performance on an economy-wide basis, therefore, are the rate of growth of potential output and its disaggregation into the growth of factor inputs and total factor productivity (TFP).

Because potential output is not an observable variable it must be estimated econometrically, which gives rise to problems of measurement and interpretation. Specifications differ, for example, because of different views about how the economy works. Measurement is also subject to statistical variations in how well the specified equations are estimated. Finally, different estimates of potential output may give rise to substantially different estimates of output gaps (the difference between actual and potential output), which could give rise to very different policy implications.

Because of the emphasis on the control of inflation in the medium-term strategy of the industrial countries. Fund staff estimates of potential output are based on the level of output that can be sustained without risking an acceleration of inflation.6 Fund staff estimates of the growth of potential output are presented in Table 1. (This and subsequent tables are mainly intended to illustrate the kind of structural indicators available; explanation of developments and interpretation of the data presented are kept to a minimum.) Two features of the estimates in Table 1 stand out: first, the marked slowdown from growth rates achieved during 1966–73 to those of 1974–79, and second, the rise in growth rates from those achieved during 1974–79 to those of 1980–87, but to levels still far below those of the initial period. 7

Table 1.

Geographic Mobility: Persons Who Changed Country or Region of Residence in Certain Years

(In percent of total population)

article image
Sources: OECD (1987).

Regional mobility rates refer to labor force only; ZEAT regions have been defined specifically for the purpose of labor market analyses.

Administrative data, which include multiple and return moves made within a year.

While potential output is the primary indicator of structural performance, different interpretations would be given to two identical growth rates if one reflected modest growth in factor supplies combined with large increases in TFP. While the other reflected very large increases in factor supplies and negative rates of growth of TFP. The former case reflects significant increases in the economy’s overall efficiency, which is generally what is meant by “structural improvements,” while the latter reflects “structural deterioration.” The contribution of TFP is defined residually as the difference between the contributions of factor supplies and the rate of growth of potential output. An important limitation of this approach in the context of structural indicators, therefore, is that one of the primary intermediate variables of importance is defined residually. Since changes in TFP cannot be measured directly, the constituent contributions coming from particular sources cannot be easily or reliably identified.

Another key intermediate variable affecting potential output is the rate of investment. Table 2 provides data on gross investment as a proportion of business output during 1966–73 and 1974–85. Although these data suggest that rates of investment held up fairly well in all countries, it is clear that it is net (not gross) investment that matters for the growth of capacity and output. Yet, net investment is precisely the indicator that is most difficult to estimate in periods of large relative price changes (such as occurred with the oil shocks) that may induce changes in rates of economic depreciation.

Table 2.

Major Industrial Countries: Gross Investment Ratio and Growth of Capital Stock and Capital-Labor Ratio, 1966–85

(Private business sector; excluding residential construction; in percent)

article image
Source: Adams, Fenton, and Larsen (1987).Note: The gross investment ratio is the ratio of gross business investment at current prices to value added; the capital-labor ratio is expressed in relation to total manhours.

Labor Markets

Four major intermediate variables directly related to labor markets have been discussed in the literature concerned with unemployment problems in the industrial countries: aggregate real wage gaps, the non-accelerating inflation rate of unemployment (the NAIRU), the degree of real wage rigidity, and the degree of nominal wage rigidity.

The real wage gap is attractive as an intermediate variable because it attempts to summarize in one number the structural factors that are preventing a return to full employment. The real wage gap is the difference between the observed real wage and an estimate of the real wage needed to achieve full employment, the “warranted real wage.” Warranted real wages are typically estimated starting from a base period when the economy was presumed to be at full employment, and then adjustments are made for productivity growth. Estimatcs of the real wage gap are quite sensitive to the base period and to the adjustment for productivity growth (see Table 3).8

Table 3.

Major Industrial Countries: Alternative Estimated Wage Gaps, 1981

(Percentage deviation of real wages from warranted real wage)

article image

Cyclically adjusted wage gap measures, from Adams, Fenton, and Larsen (1986), Table 15.

Wage gap based on trend productivity, from Gordon (1988).

The NAIRU measures the unemployment rate needed to stabilize the inflation rate. The NAIRU is a desirable indicator because it reflects both the problem of excessive real wage growth, and the more microeconomic labor market distortions, such as the mismatch of workers and jobs, inflexible industrial relations, and policy distortions caused by minimum wage laws, unemployment benefits, and taxes.9

The NAIRU, however, suffers from two major problems, one a measurement problem and the other a problem of interpretation. The measurement problem arises because the NAIRU estimate is derived from estimated wage and price equations; it cannot be measured directly. As a result, it can only be measured for a period of time, and the precision of the estimates is sensitive to how well the underlying wage and price equations have been estimated. In addition, different wage and price equations yield different estimates of the NAIRU. The importance of this latter problem is indicated in Table 4. Which lists several recent estimates of the NAIRU for industrial countries.

Table 4.

Major Industrial Countries: Estimates of NAIRUs

article image
Source: Adams, Fenton, and Larsen (1986).

The problem of interpretation arises because recently many economists have begun to argue that the primary determinant of the NAIRU is past unemployment experience.10 This interpretation of the determinants of NAIRU is called the “hysteresis” hypothesis. Under the hysteresis hypothesis, removing structural distortions will help, but the most important way of permanently lowering unemployment is to lower current unemployment, through demand expansion. There would be a cost to this demand expansion in terms of a permanent increase in inflation,11 Alternatively, under the “structuralist” hypothesis the key to a permanent reduction in unemployment is the elimination of distortions in the labor market. A corollary to this hypothesis is that the rise in European unemployment in the 1970s and 1980s reflected existing structural rigidities in the labor market in the face of successive real external shocks. Although there is more overlap between the two schools of thought than their proponents typically are willing to admit, the different factors they emphasize in the determination of the NAIRU make the link between this intermediate variable and the performance indicators uncertain.

Nominal and real wage rigidity are potentially useful indicators of the responsiveness of the economy to shocks. Nominal wage rigidity measures the responsiveness of wage inflation to an increase in price inflation and real wage rigidity measures the responsiveness of real wages to an excess supply of labor, usually measured by the unemployment rate. In a world experiencing supply shocks, where a rise in prices is usually associated with a fall in the warranted real wage, high nominal rigidity and low real rigidity can reduce the level or unemployment caused by a supply shock. From this perspective, nominal rigidity is desirable and real rigidity is undesirable. In addition, with a tendency to nominal wage rigidity demand management policies are more likely to affect real output in desired directions.

Unfortunately, there are measurement and conceptual problems with the two measures of wage rigidity. As with the NAIRU, the measures of wage rigidity are based on estimated wage equations, and therefore are subject to the same problems discussed above (see Table 5 for different measures of wage rigidity). On a conceptual level, it is very difficult to link precisely policy indicators and nonpolicy factors that determine wage rigidity. For example, the OECD finds that countries with either highly centralized collective bargaining, such as Austria, or with highly decentralized collective bargaining, such as the United States, responded well to the second oil price increase. It was countries with a mix of both systems that fared worst.12

Table 5.

Major Industrial Countries: Estimates of Real and Nominal Wage Rigidity1

article image
Sources: Adams, Fenton, and Larsen (1986).

These measures are based on empirical estimates of wage equations for individual countries and indicate the degree of real and nominal wage rigidity. The estimates by Coe are based on semiannual data for the entire private economy whereas the estimates by Grubb, Jackman, and Layard are based on annual data only for the manufacturing sector only.

In comparing the intercept terms of Coe’s wage equations for the United States and other countries it should be noted that different wage variables are used in these equations. The wage variable for the United States is an hourly earnings measure whereas for other countries a compensation per employee measure is used. This latter measure will be influenced by the strong negative trend in average weekly hours worked which may affect estimates of the intercept term in the wage equation.

Three policy indicators have the potential to be linked to the NAIRU, one of the intermediate variables discussed above: (i) unemployment insurance replacement ratios (the ratio of unemployment benefits to previous earnings); (ii) taxes on labor income, or the employment “tax wedge” (iii) minimum wages. These three factors will jointly influence an individual’s decisions to supply labor.

In theory, an increase in the replacement ratio should lead to an increase in the natural rate of unemployment. Unfortunately, the empirical link is nowhere near as clear as the theoretical link (Burtless (1987)). Moreover, it is not easy to summarize the generosity of unemployment programs in terms of a single ratio (illustrated in Table 6). First, the replacement ratio tends to change over the duration of the unemployment spell; second, health benefits in some countries cease while unemployed (the United States) and are unaffected in others (the United Kingdom), and this effect is not captured in the replacement ratio; third, the total possible duration of benefits seems to be as important in affecting search behavior as the replacement ratio (Table 7).

Table 6.

Major Industrial Countries: Macroeconomic Unemployment Insurance Replacement Ratios, 1960–841

article image
Source: Chan Lee and others (1987).

Unemployment insurance replacement ratio = (standard national accounts unemployment compensation payments/number of unemployed) divided by (compensation in manufacturing/number of production workers in manufacturing). These figures exclude other social transfers and are not adjusted for tax.

Unemployed do not include early retirees.

Does not include supplementary benefits received by the unemployed which are a part of the general U.K. welfare system.

Table 7.

Differences in Jobless Benefits and Unemployment in Five Countries, 1979–80

article image
Source: Burtless (1987).

Average number of recipients of unemployment insurance or unemployment assistance divided by the total number of persons unemployed in 1979–81.

Net replacement rate in first and second years of unemployment for average-wage worker who is married to dependent spouse and has no dependent children.

Potential duration of unemployment insurance and follow-on unemployment assistance for fully insured workers; “indefinite” implies that follow-on benefits can last indefinitely for workers who pass a means test.

Percentage of all 1979 unemployment that lasts longer than six or twelve months, respectively.

Replacement rate and duration depend on age and discretionary extension of allocation base (see text).

Although benefits are in principle related to past earnings, most benefit amounts in fact fall within an extremely narrow range because of a ceiling on benefits.

There is no follow-on program of unemployment insurance or assistance in the second year.

Six months is the usual maximum duration of benefits. Workers in states with exceptionally high insured unemployment are entitled to nine months of benefits.

Taxes on labor income drive a wedge between the wage paid by the firm and the wages the worker can use to buy goods and services. There have been some attempts to calculate tax wedges across countries. In general, non-wage labor costs as a proportion of wages and salaries have been high and rising in most countries (Table 8)). Policy inferences from cross-country comparison of tax wedges, however, must be made with care. Payroll taxes in European countries with national health benefits tend to be higher than in countries such as the United States, where there are few national health benefits. Many U.S. employers, however, pay for workers’ health insurance as a fringe benefit. It would seem to make little difference to the firm whether it pays the government or a private insurer for the workers’ health benefits.

Table 8.

Major Industrial Countries: Non-Wage Labor Costs as a Percentage of Wages and Salaries, Total Economy, 1960–86

article image
Source: Chan Lee and others (1987).


An alternative indicator of the tax wedge on labor income is the estimated total marginal tax rate (Table 9), which includes indirect taxes on consumption of goods and services (which lowers the real after-tax wage just as direct taxes on labor income do). These estimates are very high and even show a tendency to rise. Comparison of the two indicators (Table 8) vs. Table 9) for 1983 shows that the ordinal ranking of the major industrial countries according to the size of the tax wedge changes somewhat, but not by much. In recent years all of the major industrial countries have lowered marginal tax on personal income (Table 10) but in some countries partially offsetting increases in other taxes (such as social security taxes) have reduced the impact on the total tax wedge.

Table 9.

Major Industrial Countries: Total Marginal Tax Rates on Labor Use, 1979–831

article image
Source: Hagemann, Jones, and Montador (1987).

For a single-earner, married couple with two children. Calculated as a percent of total compensation, including payroll taxes, for an average production worker. Recent tax reforms are unlikely to have changed the figures significantly, except perhaps for the United States.

Unweighted average.

Table 10.

Major Industrial Countries: Recent and Proposed Changes in Lowest and Highest Personal Income Tax Rates (National Levels)


article image
Source: Tanzi (1987), updated by Fund staff.

If provincial taxes are taken into account, the top marginal rate averaged about 53 percent before the changes proposed by the tax reform of 1987.

Excluding local income taxes.

Finally, taxes on labor income are important for the allocation of resources as well as for the NAIRU. Economic efficiency requires that the choice between labor and capital should be unaffected by income taxation. This suggests that the proper value of the employment tax wedge must be determined in conjunction with the tax on capital income. While it is difficult to determine exactly the allocatively efficient level of employment taxes, a focus solely on the role of employment taxes in determining the NAIRU, could lead to distortions in capital/labor ratios.

There is a clear theoretical and empirical link between minimum wages and employment.13 An increase in minimum wages lowers employment, particularly of young workers. The size of the employment effect, however, is subject to debate. In addition, a country’s minimum wage laws are difficult to summarize in one number. There are differences in the sectoral coverage of the laws, some countries have separate minimum wages for young and adult workers, and some countries have different minimum wages for different industries (for example, the United Kingdom). Table 11 shows the kind of information it would be possible to provide.

Table 11.

Selected Industrial Countries: Minimum Wages, 1965–86

article image
Sources: Chan Lee and others (1987).

Minimum wage as a proportion of the average wage in manufacturing (percent).

Minimum wage relative to the private consumption deflator (1980 = 1.0).

Weighted average of


Capital and Financial Markets

Profit maximizing firms will invest up to the point where the expected marginal product of capital equals its real cost. The expected marginal product of capital depends on technology and long-term expectations of demand; the real cost of capital depends on real interest rates and on the marginal effective tax rate on new investment (the “tax wedge” on capital). The efficient international allocation of capital requires the equality of the real cost of capital across countries.

Financial market distortions reflect many diverse regulations. The choice of relevant indicators is. However, greatly facilitated by the homogeneous nature of credit. This suggests a focus on differences in the price of credit, that is, differences in interest rates, to judge the performance of financial markets. In particular, two types of interest rate differential that are important: differentials between internal and external interest rates for credit denominated in a particular currency, and differentials between real interest rates denominated in different currencies.

The internal market for credit of a particular country is straightforward: the U.S. commercial paper market is an internal market for short-term credit for the most creditworthy customers. It is internal because borrowers and lenders are predominantly U.S. firms. “External” financial markets allow foreigners to trade in a domestic currency. The differential between the two rates reflects exchange controls, where they exist, as well as other factors (prudential controls, tax provisions, time differences). While the quantitative link between any particular exchange control and the interest differentials may be hard to determine, the existence of such a link is not, as is illustrated in Figure 1. This figure shows the differential for Japan and Germany during specific periods of control on foreign asset transactions.

Figure 1.
Figure 1.

Exchange Controls and Euro-Domestic Interest Rate Differentials for Japan and Federal Republic of Germany

(Euro rate less domestic rate)

Source: Fukao and Hanazaki (1987).

Even in a country with no exchange controls, domestic-financial markets could still be segmented from international markets. Such segmentation in a country’s financial markets could occur if the regulation of financial institutions and the existence and depth of various financial markets differed between the country and the rest of the world. There are limits, however, to the divergence of real interest rates. Arbitrage will tend to limit interest differentials to a level that can be explained on economic grounds (e.g., the cost of regulations or tax provisions).

Figure 2 displays estimates of real short- and long-term interest rates for a number of industrial countries. At least in the top two panels there is evidence of a narrowing of cross-currency real interest rate differentials. The bottom panels show less convergence, but it is interesting to note that three of the countries depicted in this panel, France, Italy, and Switzerland, were also the countries for which there was evidence of ongoing exchange controls.

Figure 2.
Figure 2.
Figure 2.

Behavior of Real Interest Rates, 1969–85

Source: Fukao and Hanazaki (1987).

Three caveats must be entered to the interpretation of real interest rate differentials. First, the “real interest rate” is not an observable variable, but comprises the observed nominal interest rate and the unobserved expectation of inflation. To the extent that the estimate of expected inflation differs from the unobserved market expectation, measured real interest rate differentials may not provide accurate information on structural distortions. The problem of measuring expectations will clearly be more severe for the long-term rates. The second caveat about real interest rate differentials is a conceptual problem. International financial arbitragers are more interested in real rates of return, adjusted for any change in value of the currency in which the asset is denominated. Finally, differences in tax systems may influence calculations of rates of return.

The variety of the types of financial markets, institutions and regulations across countries makes it virtually impossible to list a small group of policy indicators. However, given the importance of these differences to international financial market participants, the private sector has tried to devise qualitative measures of the existence and freedom of markets. An example of such qualitative measures may be found in Kemp (1984).14

In terms of the taxation of capital, the appropriate policy indicator is conceptually clear, that is, the marginal effective tax rate on investment—more simply, the tax wedge on capital. The linkage between this policy indicator and the intermediate variable, the marginal real cost of capital, is also straightforward. But measuring and aggregating the marginal effective tax rate poses major problems. The tax wedge is not a statutory rate. Instead it is a summary measure of the tax liabilities on the income from new investment and on the tax reliefs that arise from making an investment.

Given the importance of the tax wedge, however, a number of authors have attempted estimates. Two sets of these estimates for different countries are presented in Table 12 and 13 (note that the tax wedge can be negative when the value of tax reductions more than offsets gross tax liabilities from new investment). The difference in the levels of the tax wedge between the two sets of estimates presented in the two tables is not important for the present discussion. Of more importance, is the variability of the tax wedge within a set of estimates for a given country. For example, depending on assumptions, the U.K. tax wedge could be 1.3 percent or 42.3 percent in the set of estimates in Table 13, Also the instability of the relative rankings across countries for the two sets of estimates is important. The differences in the table do not imply that the tax wedge is, in principle, unmeasurable. The differences do suggest, however, that in addition to detailed knowledge of each country’s tax laws, cross-country measures of the tax wedge are sensitive to a number of judgmental factors.

Table 12.

Estimated Total Marginal Tax Rates on Capital1

(Percent of pre-tax rate of return)

article image
Source: Hagemann, Jones, and Montador (1987).

1983 data. The total marginal tax rate which would apply to a hypothetical investment in the manufacturing sector which yielded a pre-tax real return of 10 percent. The investment is assumed to be made directly out of personal savings—the effective tax rate would be lower if the investment were channelled via a tax-exempt financial institution. The estimates take account of both the corporate and personal tax systems (assuming that the investor has other income equal to the annual wage of an average production worker).

Table 13.

Summary of Overall Effective Tax Rates, 1980

article image
Source: King and Fullerton (1984).Notes:Cases 1 and 2. The estimated tax rates depend upon the choice of a given value for the pretax rate of return on all projects in the fixed-p case (a fixed pretax annual return of 10 percent on all assets), and upon the choice for the interest rate earned on all projects in the fixed-r case (a fixed real interest rate of 5 percent per annum for all assets). It is the fixed-p results that are more relevant for an analysis of the incentive to save and invest.Cases 3 and 4. Show the effect of inflation on the marginal tax rate (t). Case 4 shows the effect of an increase in the inflation rate from 6 to 7 percent in each country.Cases 5–9. Attempt to decompose tax rate differences into those attributable to differences in rates of inflation, the allocation of investment, and actual depreciation rates, as opposed to, differences in the in-tax parameters themselves. These cases’ results indicate that none of these differences taken by themselves had a major influence on the spread of tax rates, but that in total they do have some influence. Effective tax rates are somewhat more equal among countries when only tax parameters differences remain.

Short-term nominal interest rates

article image

Inflation rates

Price indices

GNP/GDP deflators except for Sweden (CPI). For the recent period, OECD forecasts are used. After converting the quarterly series of price indices into the monthly bases by putting the same monthly figures during the corresponding quarters, all price indices at the time t, P t are smoothed by using the 3-month moving average method.

Long-term nominal interest rates

article image

The formulae used for calculating the short-term and long-term inflation rates are as follows:


For the short-term inflation rate, the formula is equivalent to three-quarter moving average of one-quarter-ahead rate of inflation

Goods Markets and Protection

Goods or product market interventions affect both the level of potential output and the flexibility of the economy in the face of shocks. Subsidies, for example, distort the pattern of resource allocation and thereby lower potential output. They also inhibit the movement of resources following changes in market incentives, thereby forcing the non-subsidized sectors of the economy to undertake proportionately larger adjustments. The conceptual links between policies and the level of potential output are more direct, however. In principle, it is possible to measure the deadweight economic losses to the economy of various government interventions in product markets, although in practice this is a difficult and contentious area. When governments intervene in particular product markets they create a wedge between domestic and international prices of the commodity. The size of this wedge—the degree of protection afforded domestic producers—is the main issue of relevance to the design of an indicator for goods markets distortions.

One choice for a policy indicator of the level of protection would be some average measure of nominal tariffs. This is not straightforward, however. The first problem is how to aggregate individual tariffs into a single summary measure of protection. The most commonly used method is the simple arithmetic average, but the shortcomings of this measure are immediately clear, as it takes no account of the relative importance of various products. Any attempt to take account of differences in importance will involve incorporating some weighting scheme. The major alternative weighting schemes available would seem to be the levels of either imports or domestic production. Since either of these could bias the summary measure by either underweighting the highly protected products (using import weights), or overweighting them (using domestic production weights), a third possibility could be the level imports plus domestic production.

A second complication is based on the fact that many types of nontariff border barriers also affect the overall level of protection, for example, import quotas or voluntary export restraints (VERs). Because all of these types of measures result in a divergence between domestic and international prices, their effects can. in principle, be translated into tariff equivalents that can be incorporated into average tariff calculations. Although this step is straightforward in theory, it is very difficult in practice because many non-tariff barriers are not particularly transparent or identifiable. In addition, the question arises whether a distinction should be made between measures that are directly protectionist and those that are designed to counteract unfair practices on the part of trading partners.

Given the difficulty of calculating tariff equivalents, the ratio of domestic to border prices itself (called the nominal protection coefficient, or NPC) has been used as a shorthand indicator of protection. Table 14 presents calculated NPCs for several major agricultural crops in selected industrial countries. As in all cross-country comparisons, these estimates must be treated with caution. One problem is that domestic prices can be measured at several stages, (for example, in agriculture: at the farm-gate, intervention board, or wholesale levels), and available data may vary across countries. Qualities and varieties of commodities may vary. Retailing markups often reflect custom rather than imposed regulations; such factors are hard to identify and measure. Even within a country NPCs can vary widely through time because world trade prices typically vary much more than do domestic prices. For example, while Table 14 shows NPCs for 1980-82, corresponding figures for 1985 would imply much higher protection because world prices were much lower then. Finally, NPCs do not measure those internal policies that are not supported by border policies; in such cases domestic and world prices are equal. For example, U.S. deficiency payments affect internal and border prices of corn equally.15

Table 14.

Selected Industrial Countries: Nominal Protection Coefficients (NPCs) for Producer and Consumer Prices of Selected Commodities in Industrial Countries, 1980–821

article image
Source: World Bank, World Development Report, 1986.

The NPC is defined as the domestic price divided by the border price.

Excluding Greece, Portugal, and Spain.

Austria, Finland, Norway, Sweden, and Switzerland.

Also including New Zealand.

The third complication relates to the existence of such non-border, industry-specific taxes and subsidies, that restrict trade indirectly and which have similar resource allocation effects to direct trade-restricting measures such as tariffs or quotas. Again, these types of measures can also, in principle, be brought into the analysis, but in practice this is difficult.

A final complication relates to the fact that tariffs apply to commodities, but resources move between economic activities on the basis of differences in value added. Therefore, to reveal the resource allocation effects of a tariff structure, it is necessary to calculate the protective rate for each stage of production, that is, the effective rate of protection (ERP).16 This rate depends not only on the tariff on the commodity, but also on the input coefficients and the tariffs on inputs. ERP calculations therefore usually require detailed input-output tables and a variety of other microeconomic data for a large number of sectors if they are to be accurate and useful.17

A conceptually appealing, though resource intensive, indicator of non-border support for production of particular goods that takes account of value-added problems is the “producer subsidy equivalent” (PSE). The PSE is defined as the payment that would be required to compensate producers for the loss of income resulting from the removal of all policy measures that had been reflected in the higher output price.18 Cross-country comparisons of PSEs in agriculture in the industrial countries show high and rising subsidies almost everywhere (Table 15). Comparison of the real budgetary expenditures with PSEs for various countries reveals the weakness of the former indicator. For example, in Japan real budgetary expenditures fell by 25 percent between 1980 and 1986, but at the same time PSEs as a proportion of the value of agricultural output rose by 50 percent—that is, from about 50 percent to 75 percent. The latter measure gives a more accurate view of the level and trend in agricultural subsidies.

Table 15.

Selected Industrial Countries: Indicators of Agricultural Intervention, 1980–86

article image
Source: OECD (1988), pp. 12 and 28.

Real index, deflated by GDP deflator, 1980 = 100.

The PSE, or producer subsidy equivalent, is defined as “the payment that would be required to compensate farmers for the loss of income resulting from the removal of a given policy measure.” The PSE is expressed as a percentage of the value of agricultural output.

The international counterpart to the PSE is the effective rate of assistance (ERA). ERAs also provide a summary figure that attempts to capture the degree of protection of domestic producers from whatever source, but the ERA is expressed as a rate of value added, not as an amount of money (or as a proportion of sectoral output) as is the PSE. The ERA is defined as the difference between the value added per unit of output in domestic prices and the value added in world prices, expressed as a percentage of the latter. The measure summarizes the entire assistance structure, including the effects of subsidies and other non-border measures in addition to the effects of border measures.19 This indicator requires detailed information on both domestic and world trade prices (including those of all inputs).

Assuming the data requirements could be met, the averaging problem discussed above would still have to be solved. In addition, resource allocation is affected by the dispersion of protection rates, so that a measure of the average ERA of import competing industries should be supplemented by a measure of dispersion around the average. A related issue concerns the fact that many countries, for example, the EEC countries and potentially the United States and Canada, belong to customs unions or free trade areas. Average levels of ERAs in this context could be very misleading. For example, it is conceivable that an ERA calculation for a given industry in France could be a negative rate (i.e., a subsidy) vis-à-vis Germany, but could be a very high positive rate vis-à-vis Japan or Korea. In practice, however, ERA calculations are highly resource intensive and have only been done for a few countries.20

Summary and Conclusions

This paper has reviewed the considerations involved in extending the existing macroeconomic indicator system to include indicators of structural policies and performance. Macroeconomic indicators are used in the context of the World Economic Outlook to assess the medium-term sustainabilily and desirability of economic policies. Structural policies in this macroeconomic context are defined as those that work through the supply side of the economy, primarily in two ways: through affecting the level of potential output; and through affecting the flexibility with which an economy responds to economic shocks. Since expansion of supply is the only way to achieve a sustainable increase in output in the medium term, extension of the formal indicator system to include structural policies is a natural step.

The macroeconomic variables policy makers wish to influence (growth of real output and demand, unemployment, inflation, and the balance of payments) are the same, whether the policies being considered operate through the supply or demand sides of the economy. Expansion of the indicator system therefore essentially involves identifying new policy indicators in the structural area, the intermediate variables (or channels) through which they affect the performance indicators, and the logical links by which the two are connected.

On a general and qualitative level the logical links are fairly straightforward. By removing economic distortions, structural policies generally result in improved economic efficiency which may raise the level of potential output. Alternatively, structural policies can also affect the way an economy responds to a shock. As a result, the economy may be able to move resources more quickly between sectors in response to changes in comparative advantage. Attempting to move beyond such generalizations by defining the precise quantitative impact on macroeconomic targets of specific structural measures is difficult, however. The difficulties are of two main types: first, the long and uncertain time lags involved; and second, the unavailability of much of the necessary data.

First, with the exception of major labor market reforms when there is excess capacity, the time lags with which structural reforms yield greater output are long and highly variable. Potential output responds more slowly to structural policies than does nominal output to monetary and fiscal policies. Moreover, the existing empirical literature contains few estimates of the relevant elasticities involved. These differences imply that while it might be sensible to assess the medium-term implications of demand-side policies with indicators semiannually or annually, such a frequency with structural indicators would not seem to be necessary, or even desirable, on such a frequent basis.

Second, the data required for many structural indicators are not readily available, which considerably raises the difficulty and cost of expanding the exercise to include them. Simple indicators that are derivable from data that are readily available are usually uninformative or misleading. Conceptually adequate indicators must usually be based on a large amount of country- and sector-specific microeconomic data. Use and interpretation of this information always involves a significant amount of additional analysis. Together these two points imply that generating and interpreting many structural indicators represents a substantial and ongoing research project.

Despite these general qualifications, a preliminary effort to identity some useful structural indicators has met with some success, although it must be emphasized that their usefulness in each case is subject to the specific limitations and qualifications discussed in the previous section. For the economy as a whole, the growth of potential output and its disaggregation into the growth of factor supplies and total factor productivity could serve as useful indicators of the performance of the supply side of the economy. In labor markets, potential intermediate variables include real wage gaps, the non-accelerating inflation rate of unemployment (the NAIRU), the degree of real wage rigidity, and the degree of nominal wage rigidity, potential policy indicators include unemployment insurance replacement ratios (the ratio of unemployment benefits to previous earnings), taxes on labor income (the “tax wedge’), and minimum wages. In financial and capital markets, potential intermediate variables include differentials between internal and external interest rates for credit denominated in a given currency, and differentials between real interest rates denominated in different currencies across countries; a potential policy indicator is the marginal effective tax rate on capital. In goods or products markets and foreign trade, possible policy indicators are the “producer subsidy equivalent” (PSE) and the “effective rate of assistance” (ERA).

There is a final point worth emphasizing. While structural indicators may provide potentially helpful signposts, they are certainly not prerequisites for addressing structural problems. Severe distortions can usually be identified without the aid of sophisticated indicators, and measures to address problem areas need not be delayed until refined indicators are constructed. It is in the self-interest of all countries to eliminate structural rigidities and raise potential output, regardless of the actions of others.


  • Adams, Charles, Paul R. Fenton, and Flemming Larsen, “Differences in Employment Behavior among Industrial Countries,” Staff Studies for the World Economic Outlook, World Economic and Financial Surveys (Washington: International Monetary Fund, July 1986).

    • Search Google Scholar
    • Export Citation
  • Adams, Charles, Paul R. Fenton, and Flemming Larsen, “Potential Output in Major Industrial Countries,” Staff Studies for the World Economic Outlook, World Economic and Financial Surveys (Washington: International Monetary Fund, August 1987).

    • Search Google Scholar
    • Export Citation
  • Baumol, William J., “Informal Judgement, Rigorous Theory, and Public Policy,” Southern Economic Journal (October 1965).

  • Bean, Charles, “Comment on Schultze,” Barriers to European Growth: A Transatlantic View, ed. by R. Lawrence & C. Schultze (Washington: The Brookings Institution, 1987) pp.29599.

    • Search Google Scholar
    • Export Citation
  • Blanchard, Olivier, and Lawrence Summers, “Hysteresis and the European Unemployment Problem,” Macroeconomics Annual 1986, ed. by S. Fischer (Cambridge, Massachusetts: National Bureau of Economic Research, 1986).

    • Search Google Scholar
    • Export Citation
  • Brown, Charles, Curtis Gilroy, and Andrew Kohen, “The Effect of the Minimum Wage on Employment and Unemployment,” Journal of Economic Literature, Vol. 20, No. 2, 1982.

    • Search Google Scholar
    • Export Citation
  • Burtless, Gary, “Jobless Pay and High European Unemployment,” Barriers to European Growth: A Transatlantic View, ed. by R. Lawrence & C. Schultze (Washington: The Brookings Institution, 1987), pp. 10561.

    • Search Google Scholar
    • Export Citation
  • Chan-Lee, James, David Coe, and Menahem Prywes, “Microeconomic Changes and Macroeconomic Wage Disinflation in the 1980s,” OECD Economic Studies, No. 8 (Paris: Organization for Economic Cooperation and Development, Spring 1987).

    • Search Google Scholar
    • Export Citation
  • Coe, David, T., “Nominal Wages, the NAIRU and Wage Flexibility,” OECD Economic Studies, No. 5 (Paris: Organization for Economic Cooperation and Development,Autumn 1985).

    • Search Google Scholar
    • Export Citation
  • Corden, Max, Protection, Growth, and Trade: Essays in International Economics (New York: Basil Blackwell, 1985).

  • Crockett, Andrew, “Strengthening International Economic Cooperation: The Role of Indicators in Multilateral Surveillance,” (Washington: International Monetary Fund Working Paper, WP/87/76, November 2, 1987).

    • Search Google Scholar
    • Export Citation
  • Crockett, Andrew, and Morris Goldstein, Strengthening the International Monetary System: Exchange Rates, Surveillance, and Objective Indicators, International Monetary Fund Occasional Paper 50 (Washington: International Monetary Fund, February 1987).

    • Search Google Scholar
    • Export Citation
  • Emerson, Michael, What Model for Europe? (Cambridge, Massachusetts: MIT Press, 1987).

  • Englander, A. Steven, and Axel Mittelstadt, “Total Factor Productivity: Macroeconomic and Structural Aspects of the Slowdown” OECD Economic Studies, No. 10 (Paris: Organization for Economic Cooperation and Development, Spring 1988).

    • Search Google Scholar
    • Export Citation
  • Feldstein, Martin and Lawrence Summers, “Inflation, Tax Rules and the Long-Term Interest Rate,” Brookings Papers on Economic Activity: 1 (Washington: The Brookings Institution, 1978).

    • Search Google Scholar
    • Export Citation
  • Fukao, Mitsuhiro and Masaharu Hanazaki, “Internationalization of Financial Markets and the Allocation of Capital,” OECD Economic Studies, No. 8 (Paris: Organization for Economic Cooperation and Development, Spring 1987).

    • Search Google Scholar
    • Export Citation
  • Gordon, Robert J., “Back to the Future: European Unemployment Today Viewed from America in 1939,” Brookings Papers on Economic Activity, (Washington, forthcoming, 1988).

    • Search Google Scholar
    • Export Citation
  • Hagemann, Robert, B. Jones and R.B. Montador, “Tax Reform in OECD Countries: Economic Rationale and Consequences,” (Paris: OECD Department of Economics Working Paper No. 40, August 1987).

    • Search Google Scholar
    • Export Citation
  • Industries Assistance Commission, Annual Report, 1986-87 (Canberra: Australian Government Publishing Service, 1987).

  • International Iron and Steel Institute (IISI), Steel Statistical Yearbook, 1987.

  • International Monetary Fund, World Economic Outlook (Washington: International Monetary Fund, October, 1987).

  • Jarrett, Peter and Raymond Torres, “A Revised Supply Block for the Major for the Major Seven Countries in Interlink,” (Paris: OECD Economics and Statistics Department Working Paper No. 41, April, 1987).

    • Search Google Scholar
    • Export Citation
  • King, Mervyn, and Donald Fullerton, The Taxation of Income from Capital: A Comparative Study of the United States, the United Kingdom, Sweden and West Germany (Chicago: The University of Chicago Press, 1984).

    • Search Google Scholar
    • Export Citation
  • Kemp, Lynette, The Wardley Guide to World Money and Securities Markets (London: Euromoney Publications Limited, 1984).

  • Layard, P.R.G., D. Grubb, and J. Symons, “Wages, Unemployment and Incomes Policy,” in Europe’s Stagflation, ed. by Michael Emerson (Oxford: Clarendon Press, 1984).

    • Search Google Scholar
    • Export Citation
  • Layard, P.R.G., Giorgio Basevi, Olivier Blanchard, Willem Buiter, and Rudiger Dornbusch, Europe: The Case for Unsustainable Growth (Brussels: Brussels Center for Economic Policy Studies, 1984).

    • Search Google Scholar
    • Export Citation
  • Lipsey, Richard, and Kevin Lancaster, “The General Theory of Second Best,” Review of Economic Studies, December 1956.

  • Meade, James, Trade and Welfare (London: Oxford University Press, 1955).

  • Organization for Economic Cooperation and Development, Structural Adjustment and Economic Performance (Paris: Organization for Economic Cooperation and Development, 1987).

    • Search Google Scholar
    • Export Citation
  • Organization for Economic Cooperation and Development, Monitoring and Outlook of Agricultural Policies, Markets, and Trade, (Paris: Organization for Economic Cooperation and Development, May 1988).

    • Search Google Scholar
    • Export Citation
  • Organization for Economic Cooperation and Development, National Accounts, various years.

  • Schultze, Charles, “Real Wages, Real Wage Aspirations, and Unemployment in Europe,” in Barriers to European Growth: A Transatlantic View, ed. by R. Lawrence & C. Schultze (Washington: The Brookings Institution, 1987), pp. 23091.

    • Search Google Scholar
    • Export Citation
  • Tanzi, Vito, “Tax Reform in the Industrial Countries and the Impact of the U.S. Tax Reform Act of 1986” (Washington: International Monetary Fund Working Paper, WP/87/61, September 1987).

    • Search Google Scholar
    • Export Citation
  • Tanzi, Vito, and Lans Bovenberg, “Economic Interdependence and the International Implications of Supply-Side Policies, (Paper presented for the International Conference on “A Supply-Side Agenda for Germany?”, Cologne, June 29-30, 1988).

    • Search Google Scholar
    • Export Citation
  • Wirtschaftsvereinigung Eisen- und Stahlindustrie, Statistisches Jahrbuch der Eisen und Stahlindustrie (Düsseldorf: Verlag Stahleisen, 1985 and 1987).

    • Search Google Scholar
    • Export Citation
  • Witteler, W. “Tarifäre und Nichttarfäre Handelshemnisse in der Bundesrepublik Deutschland-Ausmass und Ursache,” Die Weltwirtschaft (Kiel), 1987, Heft 2.

    • Search Google Scholar
    • Export Citation
  • World Bank, World Development Report (London: Oxford University Press, 1986).


The authors acknowledge very helpful comments and suggestions by their colleagues throughout the l-und, particularly Max Corden. Andrew Crockett, and Flemming Larsen.


The existing macroeconomic indicator system is described in Crockett and Goldstein (1987). An example of the use of indicators in the context of the WEO is contained in the World Economic Outlook, October 1987, pp. 18–21.


The intense interest of the Fund in structural reforms in adjustment programs in many developing countries is entirely consistent with this approach because many microeconomic rigidities in these countries have important effects on macroeconomic imbalances. For example, interest rate ceilings that result in negative real rates of interest have important effects on saving-investment balances in the economy and therefore on the balance of payments.


This is not the case in the developing countries, which helps to account for the attention financial markets receive in structural reform efforts there.


The inflation constraint is introduced into the estimations by basing potential output on an estimate of the non-accelerating inflation rate of unemployment (NAIRU). The NAIRU is the rate of unemployment below which wage and price inflation would normally be expected to accelerate and is strongly influenced by the structure of labor markets. (See the section below in which the NAIRU is discussed as an intermediate variable of labor market structural performance.)


The reasons for this pattern of developments has been discussed in previous issues of the World Economic Outlook


For discussion of problems of measurement and interpretation of real wage gaps, see Schultze (1987), pp. 239–51, and Bean (1987), pp. 295–302.


Gordon (1988) argues for the importance of hysteresis over structural problems in explaining European unemployment problems.


See OECD (1987), pp. 122–45.


The frequency with which financial market institutions and regulations change and the difficulty of deriving accurate comprehensive measures both suggest that caution is needed in using the results of such attempts at qualitative capital market measures.


World Bank (1986), pp. 112–13.


See Part II of Corden (1985).


Note that effective rales of protection are relevant only for analyzing production effects of protection (our interest here), not consumption effects which should still he analyzed in terms of nominal rates on commodities.


Note that changes in the level of the ERA indicate only trade distortions: a change could be trade-contracting if it raised levels of assistance to import competing activity, but trade-expanding if it applied to exporting activities.


See, for example. Winder (1987) and Industries Assistance Commission (1987) for estimates of ERAs for the Federal Republic of Germany and Australia.