The Macroeconomic Implications of Wage Retaliation Against Higher Taxation

In standard keynesian theory an increase in direct personal taxes reduces consumption expenditure and, by means of the multiplier process, reduces national income. Profit maximizing behavior on the part of individual producers then implies readjustments which lead to lower equilibrium levels of output, employment, and prices.


In standard keynesian theory an increase in direct personal taxes reduces consumption expenditure and, by means of the multiplier process, reduces national income. Profit maximizing behavior on the part of individual producers then implies readjustments which lead to lower equilibrium levels of output, employment, and prices.

I. Introduction

In standard keynesian theory an increase in direct personal taxes reduces consumption expenditure and, by means of the multiplier process, reduces national income. Profit maximizing behavior on the part of individual producers then implies readjustments which lead to lower equilibrium levels of output, employment, and prices.

An essential element in this result is that the money wage rate remains constant in the short run. Classical theory assumes that there will be competitive labor markets, homogeneity in labor supply, and continuation of adjustment until all markets are cleared. Under these conditions, a rise in personal taxes lowers the level of prices and causes the money wage rate to fall in the same proportion. Since the real wage does not change, the equilibrium levels of output and employment remain unaffected. The fall in real consumption expenditure caused by the tax is then exactly offset by a rise in planned investment, which occurs in response to the fall in the price level, since lower prices increase the real value of the money supply and lower the rate of interest.

While classical theory implies that a rise in personal taxes causes money wage rates to fall, there have been recent suggestions that the opposite happens because workers attempt to offset the effects of higher personal taxation by seeking wage increases. According to the Economic Council of Canada,1

… increases in direct taxes … may also tend to contribute to price increases—in the case of personal income taxes, perhaps largely through the route of increased wage demands.

If it is difficult to lower disposable income by means of personal taxation because the increased taxation evokes a wage push, it becomes tempting to conclude that it is futile to raise such taxes as a means of slowing inflation. It is also tempting to conclude that the government may not be able to increase its claim over real resources, even if they are financed by increased personal taxes, without setting off inflation. The present paper attempts an analysis of these concerns by examining the implications of wage adjustment—or what might loosely be described as disposable income bargaining—in response to increased direct personal taxation. This first section of the paper provides some descriptive background material and concludes with a brief discussion of labor supply.

Small unions that bargain at the firm or industry level cannot create a credible threat against higher personal taxation, so that wage retaliation is apt to be more acute in countries where most of the labor force is organized into a few large unions and where wage bargaining is nationwide rather than by firm or industry. Surprisingly, perhaps, evidence suggests that positive wage adjustment occurs in response to higher personal taxes even in industrial countries where the institutional conditions do not appear favorable to effective bargaining for disposable income. In the United States, for example, there is some evidence that wages and payroll taxes are positively correlated.2 Indeed, as shown in Section III, the model developed in this paper gives a reasonable interpretation of some of the successes and failures of U. S. tax policy during the 1960s. It is, therefore, fair to say that the inquiry may have significant implications for economic policy in a wide range of countries.

The threat of wage retaliation against higher taxes is taken very seriously by many governments. In Austria, the tax reduction of 1972 was the product of a direct negotiation between the unions and the Austrian Government. In Ireland, there is a “social compact,” which concedes that government expenditure and tax restraint are important elements in securing noninflationary wage behavior. In Sweden, the Swedish Confederation of Professional Associations (SACO) set its wage demands in terms of real take-home pay in 1966, and the Swedish Government eliminated the 5 per cent social security withholding tax in 1973, openly recognizing the need for tax relief as a means of reducing wage demands.3

An empirical study of tax trends in the United Kingdom by Jackson, Turner, and Wilkinson 4 holds heavy wage taxation responsible for inflation as well as for increased labor unrest. The authors speak of “wage explosion,” and they blame upward wage pressure, even when there is substantial unemployment, on the pressure placed on workers by increasingly heavy taxation. In their view there is what amounts to a taxpayers’ revolt by the back door—if one cannot legally escape the clutches of the tax collector, perhaps one can accomplish much the same end by making sure that pretax income is higher.

The authors suggest that the central problem is the very rapid and extensive increase in the taxation of persons who were not previously subject to heavy personal taxation in the United Kingdom. They state

… that the typical manual worker was hardly liable for income tax until the late 1950’s and that even in 1960 his tax and social insurance contributions together took less than 8 per cent of his earnings. In 1970/71, however, this proportion had risen to nearly 20 per cent…. Over the whole period from 1948 to 1970, gross money earnings rose at an average annual rate of about per cent, increasing prices reduced the real value of this to 2½ per cent, and taxes cut it again to 1½ per cent. But whereas up to the mid-1950’s the tax effect was negligible, from 1964 to 1970 these deductions took back the lion’s share of the workers’ annual increase in real income.5

…whereas up to the mid-1950’s it was increases in living costs that were responsible for reducing the value of workers’ monetary gains, from then on direct taxes represent a very significant offset to increases in workers’ real income.6

A circumstance shared with other industrial countries is that

The clearest example of deliberately increased direct taxation is that of social insurance contributions, which … represented about 3 per cent of men manual workers’ average income throughout the 1950’s, but then grew to some 6 per cent in 1970/71.7

The present paper assumes, but does not attempt to verify empirically, that increased taxation of persons may evoke an upward push of money wage rates. Section II develops a macroeconomic model, described as the wage adjustment model, which provides a means of analyzing the comparative static implications of a positive association between changes in personal taxes and changes in money wages.8 Section III translates the implications of the comparative static model into the disequilibrium states of inflation and recession—defined, respectively, as states of excess or deficient demand in both commodity and labor markets. The wage adjustment model will show that either the product market or the labor market can be in a state of excess demand, while the other market is in a state of excess supply. The model may, therefore, help to explain the source of “stagflation”—defined as a situation in which commodity prices are rising at the same time as which there is excessive unemployment—and it suggests methods of overcoming the problem.

Section IV returns to comparative statics. The concern is to analyze the short-run macroeconomic effects of transferring resources to the public sector when it is known that increased personal taxation may evoke a wage increase. Section V summarizes the findings and concludes with suggestions for further research.

labor supply and personal taxes

It is fitting to conclude the present section with some notes on the wage adjustment phenomenon. Although it is not the intention to delve at length into the theory of labor supply, or to attempt to make a theoretical or empirical case for the phenomenon, it should at least be noted that the possibility of upward wage adjustment in response to increased personal taxation is quite consistent with labor supply theory.

Keynes spoke of worker money illusion and postulated an infinitely elastic labor supply curve with respect to the money wage rate. The earlier economists believed in rational behavior and regarded labor supply as a function of the real wage rate. Subsequent evidence showing that workers are very conscious of the cost of living supports the classical view, but the classicists wrote before heavy wage taxation became an important factor.9 Recent years, as documented by Jackson, Turner, and Wilkinson, have seen an increasing fraction of the work force placed under heavy wage taxation. In the words of these authors,

… because the steady increase in wages in the market industrial countries has pushed a rising proportion of workers into the net of generally progressive tax systems, trends in net real wages—i.e. after allowing for direct tax and related deductions as well as for price changes—may now be the more relevant parameter of reference.10

In a similar vein Robert J. Gordon notes that

… a change in tax rates may have a substantial short-run impact on the rate of inflation. If the supply of labor offered is greater at higher anticipated real wages—reckoned as the after-tax wage rate divided by the price level that workers anticipate during the contract period—an increase in the rate of direct taxation on labor income raises the before-tax wage that will be necessary to induce workers to supply a given quantity of labor.11

Formally, these arguments suggest that the classical supply function of labor, which may be written N = S(w/p) where N is the supply of labor and w/p is the real wage, should be amended to read:


where t is the tax rate applied to labor income. If the substitution effect of a real wage change dominates over the income effect, ∂N/∂(w/p) is positive, and ∂N/∂t must then be negative. In other words, a rise in the tax rate will cause labor supply to decline in relation to any given gross real wage. The intersection of the aggregate labor demand and supply functions would then be at a higher gross real wage rate and at a lower level of employment.

This argument is based on the proposition that the substitution effect of a rise in real wages predominates over the income effect. This proposition has been the subject of much debate, and considerable empirical evidence appears to refute it. However, it has been suggested in a recent analysis that the aggregate short-term labor supply curve is positively sloped. According to Barzel and McDonald,

… in the aggregate, at a point in time, wealth effects tend to wash out due to the constraint imposed by the resource endowment of the economy. And in the absence of wealth effects, no Giffen goods are possible, and the aggregate demand curve for leisure, as for any other good, has to slope down throughout.12

On the practical level, much of the discussion of wage adjustment in response to increased taxation has been within the context of the behavior of labor unions. Here, too, a systematic attempt to raise wages in response to increased taxation is consistent with the theory of union preferences. Fellner, for example, assumed that union leadership has in mind a utility function which is characterized by the fact that employment and wages are imperfect substitutes.13 In the absence of an all or none contract, the wage rate and level of employment are set where the highest union indifference curve is tangent to the marginal revenue product curve of the firm. An increase in wage taxes implies utility maximization at a higher pretax wage rate and lower level of employment. A fall in such taxes provides disposable income relief and prompts the substitution of a lower wage for higher employment.

The analysis of this paper is greatly facilitated by adopting several simplifying assumptions. First, it is assumed that direct taxes on persons are exogenous. Second, for the comparative static portion of the paper, it is assumed that the money wage rate is fixed in the short run. It is then further assumed that a change in personal taxes provokes adjustment of the money wage rate in the same direction. The hypothesis may be expressed by the wage adjustment relation:


where N0 is the pre-existing level of employment, dT is the nominal aggregate value of the tax charge, and the parameter є is designated as the wage adjustment coefficient.

Of course, these simplifying assumptions are unrealistic, but they are useful because they help to isolate the consequences of the wage adjustment phenomenon. The essentially Keynesian treatment of wage behavior implies money illusion and eliminates from the model the fact that a change in the price level will, in turn, tend to affect money wages. However, this phenomenon is so well known that it does not seem necessary to make explicit allowance for it in the model. Similarly, it would be a great deal more realistic to make personal taxes a direct function of the wage bill. However, like the wage-price interaction, this would produce a tax-wage interaction that unduly complicates matters. It seems better, for present purposes, to focus on the once-for-all effect of a discretionary tax change. As a consequence, taxes enter the model as a lump sum that intervenes between personal income and disposable income. The taxes of the model are therefore not wage or income taxes, but rather per capita taxes, and they will be referred to simply as personal taxes. It should be understood that tax policy, for the purposes of this paper, refers only to the taxes explicitly recognized in the model and not to the whole gamut of direct and indirect taxes.

Some additional assumptions in the paper should be set forth at the outset. First, it is assumed that wage adjustment takes place in response to changes in personal taxes and government transfer payments to persons, but not in response to other policy action. Consequently, monetary policy actions are assumed to go unnoticed by workers even though a rise in interest rates could well raise cost of living indices. Similarly, a change in government purchases is assumed to have no effect on wage behavior. However, if these purchases are financed by increased taxation, the perception of the value of the services they provide may implicitly be assumed to affect the magnitude of the wage adjustment coefficient.

Most references to induced wage retaliation give the impression that the adjustment occurs in response to tax increases but not to reductions. One could attempt to defend the proposition that there is a ratchet in preference functions, but this is not necessary. The possibility that tax reduction will increase labor supply just as a tax increase reduces it is quite consistent with common observation. Money wage rates tend to trend upward because of productivity growth, cost of living adjustment, and other factors; therefore, in practice a tax reduction would temporarily slow down the rate of money wage increase, while a tax increase would accelerate it.

The magnitude of the wage adjustment coefficient will depend upon a wide variety of factors. Although it is not important to explore these in detail, the following considerations should be noted.

(1) A tax increase should not evoke a wage adjustment at all when the tax is perceived as having been imposed for the purpose of financing some specific desired social service. This perception may be weak or missing when the worker does not regard expansion of the government sector as of substantial benefit to himself, when the tax increase is imposed purely for stabilization purposes, and when the taxpayer and the beneficiary of public services are not the same person. The latter applies particularly to social security taxes. Although these are imposed with the intention of yielding direct benefits to the worker, these benefits are received by older persons while the taxpayers are younger persons who may heavily discount future benefits. Nevertheless, it could well be that a tax increase unaccompanied by a rise in government expenditure will evoke a more sizable wage adjustment than would be the case if the tax increase is perceived as financing an increase in public services.

(2) The degree of wage adjustment will depend in part on the tightness or ease of the labor market. In general, effects of labor market tightness will work in a direction opposite to the effects cited above. Thus, a tax imposed for stabilization purposes may be viewed as a greater threat to employment than a tax combined with government expenditure increase, and it may therefore evoke a smaller wage adjustment. By the same token, it seems likely that the wage adjustment coefficient will be higher as the government’s commitment to full employment becomes greater. This reduces anticipated incremental unemployment in response to wage adjustment and therefore probably tends to magnify the upward push of wages.

(3) Wage adjustment may be more pronounced if the tax increase is imposed exclusively on persons. If the increase in taxes also covers company income, the measure may be viewed as less inequitable and, by reducing after-tax profit, it may stiffen employer resistance to wage demands. On the other hand, heavy taxation of profits, especially at the margin, causes the government to share in the cost of a wage increase, and this will work in the opposite direction.

(4) Wage adjustment is likely to become more pronounced as traditional methods of avoidance and evasion become less accessible due to payroll withholding and other improvements in tax administration, especially if these improvements bear more heavily on labor income. It should also be noted that wage adjustment may take invisible forms, such as the granting by employers of greater nonmonetary benefits in order to ease the burden of taxation of their employees.

II. The Comparative Statics of Wage Adjustment in Response to Personal Taxation

This section is devoted to a comparative static analysis of the effects of an increase in personal taxes on employment, output, the price level, and other macroeconomic variables. Although such an analysis cannot speak to the problem of inflation control, where inflation is defined as a disequilibrium situation in which excess demand causes continuous price rises, it does lay the groundwork. The discussion commences with a list of symbols and then proceeds to the specification and development of the model. As it will be useful, on occasion, to give numerical examples, these symbols are assumed to have certain initial numerical values. The most important variables and their values are:

  • N0 = 50, the level of employment

  • p0 = 1.0, the level of prices

  • i0 = 5.0 per cent, the real rate of interest

  • X0 = 100, the level of real output of goods and services.

All other variables are expressed in nominal terms and are:

  • w0 = 1.5, the money wage rate

  • Y0 = 100, national income

  • C0 = 75, consumption expenditure

  • I0= 25, planned net investment

  • G0, government purchases of goods and services

  • T0, direct wage taxes net of government transfer payments

  • DI0 = 81.25, personal disposable income

  • π0 = 25, profits

  • M0 = 40, the stock of money.

The specification of the model is as follows:


Equation (1) defines national money income as the sum of the wage bill (wN) and the level of profit. Equation (2) defines disposable income as equal to the wage bill plus the portion of profit paid out as dividends (απ) minus personal taxes net of government transfer payments (T).

Equation (3) states that nominal income is the sum of consumption, net investment, and government purchases of goods and services, all expressed in nominal terms. The economy is assumed to be closed, so that exports and imports are ignored. Equation (4) is the consumption function. Real consumption is assumed to be a linear function of real disposable income. Equation (5) is the investment function. It states the conventional hypothesis that planned real net investment is a decreasing function of the rate of interest.

Equation (6) defines monetary equilibrium. The real value of the money supply (M/p) is equated with the demand for money in real terms. The latter is assumed to be an increasing function of the level of real income and a decreasing function of the rate of interest.

Equation (7) is a definition, stating that nominal national income is the product of real output and the price level.

Equation (8) is the production function. The analysis is short run in character, and no explicit recognition of capital stock or technical change is provided. The exponent λ is a positive fraction which measures the elasticity of output with respect to the quantity of labor input. Diminishing marginal productivity is built into this production function, since the marginal product of labor,


is positive but diminishes with respect to incremental labor input.

Equation (9) is the profit maximizing relation, stating that the real wage must equal the marginal product of labor under conditions of competitive equilibrium.

Equation (10), finally, is the wage adjustment equation. As discussed previously, the money wage rate is exogenous in the short run but subject to change in response to an exogenous change in the level of personal taxation.

The exogenous variables of the model are the policy variables G, T, and M, and the money wage rate w. It should be noted that all policy changes are designated in nominal terms. It is further assumed that the fiscal and monetary variables are independent. Changes in government purchases and/or taxes are therefore assumed not to affect the money supply. Furthermore, it is assumed that government deficits are financed by the sale of securities to the private sector, whereas surpluses are used to retire debt held by the private sector, and that the changes in private wealth and portfolio composition that are implicit in government finance have negligible short-run spending effects.

The analytical procedure is as follows: Equations (1) through (9) are reduced, by successive substitution, to a set of three equations expressed in the three endogenous variables that are most central to the analysis—the level of employment, the level of prices, and the rate of interest. These remaining three equations are then differentiated totally with respect to all variables. The result, expressed as a matrix equation, is:


where the term K = [1 - β(1 – α)λαβ] and is positive.

This system can be written in short form as AE = X, where A is the coefficient matrix, E is the vector of endogenous changes, and X is the vector of exogenous changes. The wage adjustment equation


implies that the exogenous vector may also be written as


To calculate the effects on the equilibrium values of the endogenous variables of exogenous policy changes, it is necessary to invert the coefficient matrix. The resultant matrix of multipliers is:


The term Δ is the determinant of the coefficient matrix and is given by the expression


It is evident, given the assumed signs of the parameters, that the first three terms must be negative. The fourth term will also be negative as long as the initial weighted surplus (βT0G0) is negative. Thus, Δ will be presumed throughout to be definitely negative.

Equation (12) is first used to calculate the effects of a tax change without wage adjustment on employment and the price level. The results, respectively, are:


They confirm the standard proposition that both employment and the price level are reduced by an increase in personal taxes.

The marginal additional changes in these variables that are caused by positive money wage adjustment are given by the expressions


These results show that an exogenous rise in the money wage rate lowers the level of employment and raises the level of prices. When equations (14) and (16) are summed, the result shows that a tax increase combined with positive wage adjustment must lower the level of employment, that wage adjustment magnifies the loss in employment, and that this magnification increases as the wage adjustment coefficient increases.

As shown by summing equations (15) and (17), the direction of change in the price level, which is brought about by the tax increase plus wage adjustment, is unpredictable in the absence of quantitative information about the values of the parameters. The fall in aggregate demand due to the increase in the tax is observed as a fall in demand in individual markets. Profits are therefore maximized at a lower level of prices. However, upward wage adjustment simultaneously raises marginal costs, and this implies profit maximization at a higher price level. Thus, the depressing effect on the price level of the reduction in aggregate demand is counteracted by the cost effect of the wage adjustment, and if the wage adjustment coefficient is sufficiently large, the equilibrium price level rises.

As can be seen from the term in the numerator of equation (16)—namely, [(βT0G0)δ2-γM0]є—wage adjustment affects output and employment through its effects on investment (as reflected by the term γM0) and through possible money illusion effects that will occur when nominal taxes and government purchases are held constant as the price level varies.

If nominal government purchases are held constant, a rise in the price level lowers the real value of government purchases, thus tending to depress output. On the other hand, the rise in the price level lowers the real value of tax yield, raises real disposable income, and therefore shifts the consumption function upward. Thus, conceivably, wage adjustment could raise employment and output if the positive consumption effect overcomes the negative government purchase effect. However, this is most unlikely, since it could occur only in the event that the weighted budget surplus is positive—and even then it would also have to overcome the negative real investment effect. In the numerical examples to be presented below, these money illusion effects are eliminated by assuming that the initial values of government purchases and taxes are zero.

Without budgetary money illusion, the effect of wage adjustment on the real variables of the system operates through its effect on investment. If, for example γ = 0, there would be no effect at all on the equilibrium levels of employment, output, and the other real variables. The reason is as follows. A rise in the money wage rate causes an initial reduction in output and real income. Since the marginal propensity to consume is less than unity, real consumption declines by less than the fall in real income; therefore, excess demand in the commodity market develops, and the price level rises. With fixed real investment and government purchases, real aggregate demand can equal real output only at the original level of real consumption and real output. It follows that the price level must rise in exact proportion to the change in the money wage rate.14

If, however, real investment is not exogenous, the equilibrium values of the real variables will change. The increase in the price level lowers the real value of the money supply. This raises the rate of interest and lowers real investment which, in turn, induces a fall in real consumption. The depressing effect of the tax increase on the real variables of the system is therefore accentuated.

Taken by itself, the tax increase unambiguously raises the level of real investment. The wage adjustment, taken by itself, however, lowers real investment. In combination, the effect on investment is ambiguous. This can be seen by combining the solution for the change in the equilibrium rate of interest implied by equation (12) with the investment function. The expression is considerably simplified by assuming that To = G0 = 0 and becomes


Without wage adjustment, this is unambiguously positive. However, with positive adjustment, the term KM0є will be negative, and it could be dominant. Consequently, a tax increase combined with wage adjustment may actually lower real investment expenditure. The monetary system may, under these conditions, be said to accentuate rather than cushion the fall in employment and real output.

To examine the effect of a tax increase on the complete system, it is useful to turn to specific numerical examples. For this purpose the following parameter and intercept values are assumed:


These values, when combined with the assumed initial values listed above, imply the numerical results shown in Table 1. Columns A and B show the changes in the equilibrium values of all the variables of the system that results from a tax increase of one nominal unit. Column A presents the standard case with fixed money wage (є = 0), whereas column B shows the effects under the assumption of complete wage adjustment (є = 1).

Without wage adjustment the tax has the effect of lowering output (item 7), employment (item 4), and the equilibrium price level (item 5). On the expenditure side, real consumption (item 9) drops by more than the fall in total output, but this is offset by an increase in real investment (item 10). The latter occurs because of the reduction in the rate of interest (item 6). This, in turn, occurs because the fall in the price level increases the real value of the money supply (item 12) and because the fall in real output reduces the quantity of money demanded at the original rate of interest. Because of the fall in the price level, all of the nominal magnitudes Y, C, I, DI (items 17–20) fall by less than their real counterparts.

Table 1.

Effects of a Rise in Personal Taxes1

article image

See the text for explanation of symbols.

Turning to the wage adjustment case, the magnitude of the assumed wage adjustment coefficient is such as to offset completely the rise in taxes. In this extreme example, the equilibrium price level increases. Moreover, as proved earlier, the employment and output effects are magnified. Despite the initial tendency of wage adjustment to compensate for the loss in disposable income, real disposable income (item 8) and consumption ultimately fall by more than they would have fallen without wage adjustment. Because of the tendency of the price level to rise (fall) by less under conditions of wage adjustment, the fall in the rate of interest is moderated, preventing real investment from rising as much as it does in the conventional case. Consequently, the tendency for a drop in consumption to be offset by a rise in investment is moderated, and a tax change therefore has a much greater impact on output and employment than it would have if there were no wage adjustment.

To what extent is labor benefited by wage retaliation in response to increased personal taxation? Comparison of the entries in item 13 shows that the real wage rate rises by more under wage adjustment, but the benefit of this is offset by the greater loss in employment. Although wage adjustment prevents the nominal wage bill (item 15) from falling by as much, the loss in real disposable income and real consumption is nevertheless greater. Thus, wage adjustment in response to a tax increase may imply rational behavior from a ceteris paribus point of view, but it could quite easily redound to the disadvantage of labor once the system adjusts to the changes and settles down to its new equilibrium. However, as suggested in the next section, the fault may ultimately lie with inappropriate monetary and fiscal policies and not with labor.

Columns C and D of Table 1 show the effect of combining the income tax increase with just that reduction in the nominal quantity of money needed to maintain a constant rate of interest and therefore to fix the real value of investment. Inasmuch as government purchases are initially assumed to equal zero, there can be no change in real government purchases or investment, so that any change in output must exactly equal the change in real consumption. The model, so restricted, is equivalent to assuming that real investment is exogenous.

In this case, the loss in employment, output, real disposable income, real consumption, and the rise in the real wage rate are identical with and without wage adjustment, and these changes are therefore entirely attributable to the tax increase. As shown in item 3, the required change in the nominal money supply needed to prevent the interest rate from changing differs with the magnitude of the wage adjustment coefficient. However, the magnitude of the change in the real value of the money supply (M/p) is the same in the two cases.

The discussion thus far has dealt with the effect of policy on the equilibrium price level. As such it has little to say about the control of inflation where the term inflation is meant to be descriptive of a dynamic phenomenon characterized by persistent disequilibrium. Nevertheless, the wage adjustment model does provide an approach to the analysis of the disequilibrium problem. As discussed in the next section, the wage adjustment model shows that policies may be mixed in such a way as to achieve general macroeconomic equilibrium.

III. Disequilibrium, Stagflation, and the Appropriate Mix of Policy

The preceding section has dealt with comparative static analysis. In the present section attention shifts to disequilibrium states. Inflation implies excess demand in markets for goods and services (product markets) and labor markets, respectively. Recession implies excess supply in these markets. “Stagflation” occurs when excess product demand coexists with excess supply in the labor market.

the wage adjustment model and labor and product market balance

The wage adjustment model suggests that as the wage adjustment coefficient increases, personal taxation becomes a progressively less efficient way of lowering the equilibrium price level, while it simultaneously becomes a relatively more efficient way of changing the levels of employment and output. This fact itself cannot be used to infer that increased personal taxation is an inappropriate way to control inflation, because it is quite possible that a policy which tends to raise the equilibrium price level may, nevertheless, be quite effective in eliminating excess demand and therefore in eliminating a persistent tendency for the price level to rise.

The first step in adapting the wage adjustment model to the analysis of disequilibrium situations is to utilize the model to construct a pair of iso-employment and iso-price level loci.15 Let a tax change together with wage adjustment be accompanied by a simultaneous change in the money supply. Equation (12) shows that the respective changes in employment and the price level are


To calculate the change in money supply that exactly offsets the employment effect of a change in tax, dN is set equal to zero, and equation (18) then implies:


This expression traces the slope of an iso-employment locus. This slope is positive, since a rise in the money supply raises employment and a tax increase would therefore be needed to provide offset. This locus might appear as in Chart 1. If the particular level of employment implied by the locus is consistent with equilibrium in the labor market, points above the locus will imply excess demand for labor and wage inflation, whereas points below it imply excess supply of labor with consequent involuntary unemployment and downward wage pressure. Since points on the NN′ function imply that the labor market is cleared, the function may be described as the labor market balance function.

Similarly, equation (19) can be utilized to derive an iso-price level locus. When dp is zero, the equation implies:


which may be positive or negative, depending upon the magnitude of the wage adjustment coefficient. The iso-price level locus defines the combinations of money supply and personal tax that will prevent the price level from changing. Above the curve there is excess demand and a tendency for the price level to rise, whereas below the curve there is deficient demand and a consequent tendency for prices to fall. The pp’ relation may therefore be described as the commodity market balance function.

Chart 1.
Chart 1.

Iso-Employment and Iso-Price Level Loci Without Wage Adjustment

Citation: IMF Staff Papers 1974, 003; 10.5089/9781451969344.024.A008

When the wage adjustment coefficient is zero, equations (20) and (21) reduce to


which shows that, in the conventional case, the labor market balance and commodity market balance functions have equal slopes.

Chart 1 shows the superimposition of the two functions. Points above them imply both wage and price inflation of the excess demand type, whereas points below the functions imply recession and downward wage and price pressure. The superimposition implies that if the employment effect of a personal tax change is offset by a change in the money supply, the price level effect will also be exactly neutralized. The respective policies therefore enjoy no comparative advantage with respect to the two targets.

This represents the standard result of macroeconomic theory. It implies that changes in the money supply and changes in personal taxes are perfect substitutes for each other with respect to their impact on labor market equilibrium and product market equilibrium, respectively. Either policy or any one of an infinite number of combinations of the two policies can be used to eliminate recession or inflation. Indeed, one policy can even be used to compensate for a movement in the wrong direction by the other policy. Moreover, and as shown in Chart 1, price inflation and involuntary unemployment cannot coexist. Classic inflation or recession are the only possibilities.

The wage adjustment model enriches the possibilities. From equation (20), it can be seen that the slope of the NN’ function increases as the wage adjustment coefficient increases.16 The reason is that wage adjustment magnifies the employment loss caused by a rise in personal taxes, so that a relatively larger increase in the money supply is required to maintain labor market balance.

Similarly, the slope of the iso-price level locus diminishes as the wage adjustment coefficient increases, and if the wage adjustment is sufficiently large, the slope could become negative.17 Wage adjustment reduces the fall in the price level that results from an increase in taxes, so that a smaller increase in the money supply is required to provide price level offset.18

Chart 2 shows the situation that might emerge when positive wage adjustment exists but is not sufficiently powerful to produce a negatively sloped product market balance function. There now emerge not two regions, but rather four. Above and below the two functions lie standard inflation and recession, respectively. However, between the two functions lie a set of zones that were not anticipated by the theory of employment. The zone to the right of the intersection is consistent with what has been called inflationary recession, or stagflation. At a point such as A, for example, there is excess supply in the labor market, and excess demand in the goods market. Thus, unemployment and inflation may coexist.

Chart 2.
Chart 2.

Iso-Employment and Iso-Price Level Loci With Wage Adjustment

Citation: IMF Staff Papers 1974, 003; 10.5089/9781451969344.024.A008

The interpretation of “stagflation” provided by the wage adjustment model is that it may be the product of a situation in which both the money supply and the level of personal taxes are too high. Such a situation may arise as the result of an attempt to combat inflation by raising taxes, combined with a well-intentioned effort by the monetary authority to avert fiscal “overkill.” It may arise as the consequence of an attempt to raise the economy’s growth rate by reallocating expenditure from consumption to investment by changing the policy mix in favor of easier money and higher taxes. And finally, an economy in the inflation zone could be automatically dragged from there into stagflation because of the tendency of the real value of the personal income tax burden to rise in a progressive tax system when nominal income increases as the result of inflation.

The trouble with such a change in the policy mix is that if the tax increase provokes wage adjustment, it will also be more effective, relative to restrictive monetary policy, in creating excess labor supply than in eliminating excess product demand. The result is stagflation.

Stagflation may be eliminated in time by the operation of natural market forces. The rising price level lowers the real value of the money supply, raises the rate of interest, and thus tends to eliminate excess product demand. Meanwhile, excess supply in the labor market tends to be eliminated by downward money wage pressure. The operation of these automatic forces implies that persistent stagflation will require continuous increases in the nominal stock of money and in the nominal values of personal taxes in a manner that keeps their real values persistently excessive. As noted above, however, the progressive personal income tax may assure that the real value of the tax will automatically be excessive.

The policy response indicated by the wage adjustment model is a reduction in the money supply and in personal taxes. The expected effect is a lull in wage demands, lower unit labor costs, and a bonus in the form of increased output and employment. When combined with a reduction in the money supply, it would also stabilize the price level. Conversely, and at the risk of repetition, a rise in taxes might provoke upward wage adjustment and therefore not yield the effects that such a policy change is normally expected to have.

the appropriate assignment of policies

The wage adjustment model focuses attention upon the importance of the appropriate mix of policies. Referring again to Chart 2, the avoidance of inflation, deflation, or stagflation is a matter of finding the levels of money supply and personal tax that coincide with the intersection of the pre-existing labor market balance and product market balance functions. Thus, the model suggests a Tinbergian approach to policy that emphasizes the appropriate mix of policy and the need to coordinate the actions of the policy-making agencies.

If the equilibrium values for money supply and taxes are known, the choice of policy becomes simple, since the instruments can immediately be set at these values. In practice, however, exact information on which to base policy may not be available. The authorities may know what zone they are in, but they probably do not know the exact shape or location of the respective balance functions; therefore, they will not know the magnitude of the required levels of the instrument variables. Furthermore, a particular policy-making authority may not be willing or able to coordinate its activities with the authority that controls other policy instruments, or to be fully aware of the fact that its own policy actions may prove either inadequate or excessive because of the unanticipated response of another instrument.

One solution to the problems of faulty policy coordination and missing information is Mundell’s well-known principle of effective market classification, according to which each policy instrument ignores the fact that it may have some influence over all targets and concentrates its attention exclusively on that target over which it has the greatest relative influence or comparative advantage.19

The wage adjustment model suggests that the comparative advantage of monetary policy lies in influencing the product market, whereas the comparative advantage of personal taxes lies in influencing the labor market. Thus, if monetary policy aims for the pp’ function, while personal taxes are adjusted in such a way as to reach NN’ the policies will eventually approach their equilibrium values despite the absence of coordination. If the assignment is reversed, there will be a progressive divergence of the instrument values from their equilibrium values.

This proposition can be illustrated with reference to Chart 2. Starting at point A, monetary policy may attempt to secure full employment by expanding the money supply to G. This raises the rate of inflation and may prompt a personal tax increase which places the system at H. However, this creates an even worse employment problem than the initial one. The assignment of policy here is incorrect and destabilizing.

Alternatively, beginning again at point A and supposing that taxes are reduced so as to place the system at point I, the resultant easing of wage pressure secures a large employment gain at the cost of relatively little additional inflation. Monetary policy then moves to point J to eliminate the inflation, and the resultant small increase in unemployment is then combatted by a smaller tax reduction than the previous one. This sequence, if carried forward continuously, eventually attains equilibrium and shows that the instruments have been appropriately assigned to their respective targets.

The possibility that the pp’ curve might be negatively sloped is, perhaps, remote. Nevertheless, the consequences of such a possibility should be noted. In this event, wage adjustment would be so pronounced that a rise in personal taxes would raise the equilibrium price level and a reduction in the money supply would be needed to produce offset. This situation is illustrated in Chart 3, where it can be seen that the policy assignment remains the same as before, provided that the slope of NN’ remains steeper than the slope of pp’. Chart 3 also confirms that personal taxes ought not to be raised under conditions of stagflation and that they might profitably be reduced instead.

It may be useful to attempt to provide concreteness to the wage adjustment model by utilizing it to interpret what took place during certain episodes. Such episodes are the product of numerous complex factors and what follows should be interpreted as only one of several plausible suggestive hypotheses.

Chart 3.
Chart 3.

Iso-Employment and Iso-Price Level Loci When Wage Adjustment Raises the Price Level

Citation: IMF Staff Papers 1974, 003; 10.5089/9781451969344.024.A008

Consider, for example, the effort of the United States to raise employment by reduction of direct taxes in early 1964. The economic expansion that occurred during this period was accommodated by monetary growth, but this was not excessive as indicated by the upward drift of interest rates during the period. This combination of policies appears to have been highly successful in attaining the goals of higher output and employment without a rise in the rate of inflation.20

The wage adjustment model would provide the following interpretation. Policy in 1964 was properly assigned, taxes being used to raise output and employment, with monetary restraint helping to hold down growth in the price level. The tax cut provided disposable income relief. Upward wage pressure was therefore reduced, unit labor costs fell, and the result was a large gain in real output and employment combined with excellent price performance.

It is important to emphasize again that equally plausible alternative interpretations undoubtedly exist. However, in support of the wage adjustment model, it should be noted that the unit labor cost function developed by Robert J. Gordon shows that unit labor costs in the last three quarters of 1964 were substantially below those predicted by the equation.21 On the other hand, in 1968 following the imposition of the surcharge at mid-year, unit labor costs rose sharply and exceeded the values predicted by Gordon’s equation. This latter circumstance perhaps explains the disappointing economic performance of that period.

Gordon’s empirical findings lead him to conclusions similar to those of this paper. In interpreting more recent events in the United States, he writes

A [tax] shifting parameter greater than zero implies that a tax increase designed to raise the unemployment rate in order to reduce the rate of inflation will … raise the price level associated with any given path of unemployment. In this case a movement from a low unemployment rate to a higher unemployment rate engineered by tight monetary policy … would be associated with less inflation than the same unemployment path engineered by tight tax policy. A corollary in the 1971 economic situation would be that a low-tax, tight-money mix would lower the rate of inflation associated with any given unemployment rate, as compared with an easy-money, high-tax policy.22

Unfortunately, the policy of lowering taxes and making money tighter may not be an ideal or a readily available option. This policy mix implies higher interest rates, and it attacks excess commodity demand by restricting investment spending. Thus, the change in mix may be harmful to the economy’s growth prospects. Also, the change in policy mix may attract an inflow of foreign capital. If fixed exchange rates were maintained, the inflow would require the stabilization authority to purchase foreign exchange and this would raise the domestic money supply.

Another option is to combine personal tax reduction with reduction in government purchases. This, too, would have the effect of raising employment and eliminating excess product demand, and it would not increase the rate of interest significantly. However, it carries with it the obvious problem that it may reduce the relative size of the public sector in amounts considered socially undesirable. The next section explores this problem of obtaining adequate resources for the public sector.

IV. Transfer of Resources to the Public Sector Under Conditions of Wage Adjustment

The attempt by government to increase the share of resources that is devoted to the public sector has often been cited as a source of inflation. Reluctance to raise taxes often means that increased expenditures are deficit financed. When the resultant borrowing requirements threaten to raise interest rates and depress prices of government securities, the monetary authority is called upon to support the government bond market. The deficit is then accompanied by monetary expansion. Even when balanced budget expansion is the case, there is some net aggregate demand increase that results from the operation of the balanced budget multiplier.

When wage adjustment in response to increased taxation is present, the problem of noninflationary resource transfer to the public sector becomes even more difficult. Government increases its purchases of goods and services and finances this by a tax increase. But, if a wage adjustment occurs, the price level is both pulled up by the increased expenditure and pushed up by the wage adjustment. The movement of the price level may therefore be highly unfavorable and, as shown below, the combined policies may even be accompanied by a fall in output and employment. There is danger, then, that even resource transfers to the public sector collectively approved by a majority of the electorate may be obstructed because of the threat to the price level by the transfer process. Assar Lindbeck’s observations on the situation in Sweden are appropriate in this connection:

… it has also been suggested … that the allocation policy—the attempt to reallocate resources toward the public sector—has inflationary effects. The idea is that labor market organizations have tried to “retaliate” not only against progressive taxation but against tax increases in general.23

Lindbeck points out that, as a consequence,

Average wage increases often have not been consistent with the political decisions about expenditure on goods and services…. The result of these attempts by the public and private sectors together to obtain more than 100 per cent of national income has been a more or less permanent tendency to cost inflation.24

In studying the problem of resource transfer, it will be assumed that overall equilibrium exists initially and that it is therefore sufficient to study the comparative statics of the problem. Accordingly, the basic model is the same as that in Section II. Moreover, the problem is explored by means of numerical examples that utilize the same parameter values as those used above.

Table 2 shows the effects of three fiscal policies. Columns A and B show the effect of an increase in personal taxes. They are duplicated from columns A and B of Table 1 for the convenience of the reader. Column C shows the effect of an increase in government purchases of one nominal unit. Column D shows the balanced budget multiplier effect of a simultaneous rise in nominal government purchases and taxes when wage adjustment is absent. Finally, column E shows the effect of balanced budget expansion when the wage adjustment coefficient is assumed to have a value of unity.

Conventional balanced budget expansion increases the equilibrium levels of output, employment, prices, and the rate of interest. As shown in column D, the additional resources acquired by the government are obtained in part by increased output (item 7), and in part by reduced real consumption (item 9) and real investment (item 10).

When balanced budget expansion is accompanied by fully compensated wage adjustment, the result, as shown in column E, is something of a nightmare. The price level is both pulled by the aggregate demand effect of the policies and pushed by the wage adjustment, and therefore a very pronounced rise occurs in the price level. Furthermore, wage adjustment tends to reduce output, and the reduction in real consumption and real investment will therefore be magnified by wage adjustment, even though it is the intention of wage adjustment to prevent a fall in real consumption.

Table 2.

Changes Due to Three Fiscal Policies1

article image

See the text for explanation of symbols.

In order to prevent wage adjustment from occurring, personal taxes must not be allowed to rise. Thus, the noninflationary means of financing a government purchase is by borrowing from the private sector and combining this with further restrictive monetary policy. This follows from considerations of appropriate assignment. The price level effects of the increased expenditure should be neutralized by restrictive monetary policy. If neither a change in government purchases nor a change in the money supply provokes wage adjustment, the two policies will be perfect substitutes in the sense in which that term has been used in this paper. Therefore, if monetary restriction prevents excess demand from arising in the commodity market, such restriction also automatically averts disequilibrium in the labor market. Consequently, if a high priority is attached to the prevention of a rise in the price level, personal taxation should continue its previously assigned role of maintaining labor market equilibrium, and this implies that such taxes should not be changed at all.

The difficulty with such a mix of policy is that the entire resource transfer is at the expense of investment, presuming, of course, that consumption is interest inelastic. In this event the rate of interest rises by an amount that reduces real and nominal investment by exactly the amount of the increase in government purchases. The combination of policies prevents output, employment, the price level, and disposable income from changing. Consequently, investment becomes the residual that varies when the magnitude of the public sector is changed.

In other words, when wage adjustment accompanies a rise in personal taxes, it will be possible to transfer resources to the public sector without a rise in the price level only if the transfer is at the expense of investment. Transferring resources at the expense of consumption can be effected, but it will involve a rise in the price level. Again, in connection with developments in Sweden, Assar Lindbeck notes that

Almost ten percentage points of GNP have been moved from private consumption to the public sector in the post-war period. Thus, it seems that even if shifting of higher taxation to average nominal wages has succeeded, the shifting has not succeeded in real terms; the public sector rather than the household sector has managed to sustain its spending plans in real terms.25

V. Results, Implications, and Directions for Research

This paper has developed a short-run comparative static model that allowed for wage adjustment in response to, and in the same direction as, a change in personal taxes. Comparisons were made between the wage adjustment model and the “conventional” Keynesian model in which it is assumed that the money wage rate remains fixed when various policy changes are introduced. The model is a closed economy model, and it was assumed that government had available three instruments of policy—changes in government purchases, changes in personal taxes, and changes in the money supply. The following results emerged:

(1) In the conventional model an increase in personal taxes lowers the equilibrium level of employment as well as the equilibrium level of prices. If, instead, the tax increase is accompanied by upward wage adjustment, the loss in employment is magnified while the fall in the equilibrium price level is reduced. If upward wage adjustment is substantial enough, the equilibrium price level may increase. Wage adjustment initially impinges on profits but then results in adjustment of output, employment, and the price level.

(2) In the conventional model the restrictive effects of a personal tax increase are in part offset by the response of the monetary system, because the fall in the price level raises the real value of the money supply. This, in turn, lowers the rate of interest and raises the level of investment. However, when a tax increase is accompanied by upward wage adjustment, the monetary cushion is blunted because of the tendency of wage adjustment to prevent the price level from falling. If wage adjustment is sufficiently powerful, the price level may rise, the real value of the money supply may therefore fall, and investment may decline rather than increase.

(3) In the conventional model a tax change that exactly offsets the price level effect of a change in the money supply will simultaneously exactly offset the employment effect. However, when changes in personal taxes are accompanied by wage adjustment in the same direction, this will no longer be the case. In that event changes in the money supply cause a relatively greater change in the equilibrium price level, whereas changes in taxes cause a relatively greater change in the equilibrium levels of output and employment.

(4) As a consequence of this comparative advantage, the simultaneous attainment of full employment and price stability becomes a problem of finding the correct mix of policy. Failure to find the correct mix may result in inflation or recession, and it may also produce “stagflation.” The wage adjustment model suggests that the existence of stagflation may reflect a mix of policy in which the money supply and the level of personal taxes are both too high. It therefore suggests that the appropriate way to deal with the problem is to change the mix in favor of a combination of lower taxes and tighter money.

(5) The existence of comparative advantages between policies implies that the values of the various instruments should be set in a coordinated manner designed to attain overall equilibrium. However, if policy is difficult to coordinate, equilibrium may be attained by assigning each policy instrument exclusive responsibility for that target over which it has the greatest relative influence. If there are two instruments—personal taxation and the money supply—the presence of wage adjustment implies that monetary policy should seek to achieve equilibrium in product markets while personal taxes should be adjusted to maintain equilibrium in the labor market.

(6) Equal increases in government purchases and taxes raise the equilibrium levels of output, employment, prices, and the rate of interest. The additional resources acquired by the government are obtained in part by increased output and in part by reduced consumption and investment. If these policies are accompanied by upward wage adjustment, output and employment may fall, while the increases in price level and interest rates may be magnified. Because wage adjustment tends to cause output to fall, the reduction in real consumption and real investment will be magnified, even though it is the presumed intention of wage adjustment to prevent a fall in real consumption. Balanced budget expansion causes the price level to be pulled up by the increase in government purchases and pushed up by the wage adjustment. Consequently, the transfer of resources to the public sector is apt to be accompanied by a substantial rise in the level of prices.

In closing, it seems appropriate to consider some avenues for further inquiry. First, the empirical basis for the wage adjustment model needs to be more firmly established, and an attempt to measure the wage adjustment coefficient for different countries under different circumstances needs to be undertaken. Closely related to such an investigation would be an attempt to delineate other sources of comparative advantage between policy instruments. The differential impacts of the various kinds of taxes should be studied. Also, an attempt should be made to amend the wage adjustment model by incorporating endogenous wage taxation.

Finally, the balance of payments implications of the wage adjustment model are likely to be of some significance. As was noted in Section IV, the model suggests that balanced budget expansion might lower the equilibrium level of real output. If this is the case, such expansion might lower the demand for imports, so that a set of policies that is usually thought to be expansionary might have the effect of improving the current account of the balance of payments.


Mr. Dernburg, an Assistant Division Chief in the Fiscal Affairs Department, is a graduate of Swarthmore College and received his doctorate from Yale. He has taught at Purdue University, the University of Michigan, and Oberlin College, and has served on the staff of the Council of Economic Advisers.


Economic Council of Canada, Third Annual Review (Ottawa, 1966), pp. 223–24.


See Robert J. Gordon, “Inflation in Recession and Recovery,” Brookings Papers on Economic Activity, No. 1 (1971), pp. 105–66.


The Swedish case is among the most interesting. It has been the policy of the Government not to interfere directly in the wage determination process. In these circumstances, Bent Hansen suggested an indirect method of controlling wage increases in line with estimated productivity growth. In the event that the acual settlement exceeded the Government’s recommendation, income taxes would be punitively raised to an extent that would reduce real disposable income below the level that would have been obtained if the Government’s guidelines had been heeded.

The irony of the situation is that while Hansen’s proposal was never put into practice, the message that it conveyed was not lost on the unions, which reasoned that if taxation can retaliate against wage increases, the reverse can just as equally be the case. For an English language discussion of the problem in Sweden, see Assar Lindbeck, “Theories and Problems in Swedish Economic Policy in the Post-War Period,” American Economic Review, Vol. 58, Pt. 2, Supplement, (June 1968), pp. 1–87. For a discussion of labor market policies in Sweden, see Ekhard Brehmer and Maxwell R. Bradford, “Incomes and Labor Market Policies in Sweden, 1945–70,” Staff Papers, Vol 21 (March 1974), pp. 101-26. Bent Hansen’s proposals can be found in his book, The Economic Theory of Fiscal Policy (London, 1958), Chapter 17.


Dudley Jackson, H. A. Turner, and Frank Wilkinson, Do Trade Unions Cause Inflation? Two Studies: with a Theoretical Introduction and Policy Conclusion, Occasional Paper No. 36 (Cambridge University Press, 1972), Chapter 3.


Ibid., p. 65.


Ibid., pp. 67–68.


Ibid., p. 68.


Similar attempts to develop such comparative static models are those of Alan S. Blinder, “Can Income Tax Increases Be Inflationary? An Expository Note,” National Tax Journal, Vol. 26 (June 1973), pp. 295-301, and John H. Hotson, “Neo-Orthodox Keynesianism and the 45° Heresy,” Nebraska Journal of Economics and Business, Vol. 6 (Autumn 1967), pp. 34-49.


For a pioneering effort to examine the effect of income taxes on the workleisure choice, see Richard Goode, “The Income Tax and the Supply of Labor,” Journal of Political Economy, Vol. 57 (October 1949), pp. 428–37.


Jackson, Turner, and Wilkinson, Do Trade Unions Cause Inflation? (cited in footnote 4), p. 64.


Gordon, “Inflation in Recession and Recovery” (cited in footnote 2), p. 113.


Yoram Barzel and Richard J. McDonald, “Assets, Subsistence, and the Supply Curve of Labor,” American Economic Review, Vol. 63 (September 1973), pp. 627-28.


William John Fellner, Competition Among the Few: Oligopoly and Similar Market Structures (New York, 1949), pp. 252-81.


Observe that when both βT0G0 and γ equal zero, Δ reduces to Kλδ2X0p0w0. Consequently, equation (17) implies that dp=εdTλX0=N0dwλX0 Using the profit maximizing relation w/p = λ(X/N), it then follows that dp/p=dw/w


Robert A. Mundell has employed this type of device a number of times. The best-known example is his “The Appropriate Use of Monetary and Fiscal Policy for Internal and External Stability,” Staff Papers, Vol. 9 (March 1962), pp. 70–79. However, see also “The Monetary Dynamics of International Adjustment Under Fixed and Flexible Exchange Rates,” Quarterly Journal of Economics, Vol. 74 (May 1960), pp. 227–57. These papers are reprinted in Mundell’s book, International Economics (New York, 1968). These applications deal with “external” and “internal” balance. Another ingenious application within the international payments context is Trevor Swan, “Longer-Run Problems of the Balance of Payments,” in H. W. Arndt and W. M. Corden, eds., The Australian Economy: A Volume of Readings (Melbourne, 1963), pp. 384–95.


The change in the slope of the NN’ function with respect to a change in the wage-adjustment coefficient is given by d(dMdT)ndε=(βT0G0)δ2+γM0γλp0X0>0.


The change in the slope of the pp’ function with respect to a change in the wage adjustment coefficient is:



It is interesting to consider the values of the slopes of the NN’ and pp’ functions under the extreme special assumptions that have occupied a great deal of attention in the literature. Keynes’ liquidity trap implies that monetary policy cannot affect either employment or the price level, so that both NN’ and pp’ are vertical. The classical case, in which the demand for money in real terms is proportional to the level of real output, is more interesting. Under the conventional assumptions the NN’ and pp’ curves would be horizontal, with too much money implying inflation, too little implying deflation, and the level of taxes being irrelevant. An increase in taxes has no effect on expenditure, since the depressing effect on consumption is offset by a rise in investment in response to a fall in the rate of interest. In the wage adjustment model, a tax increase also raises wages and prices. Since this reduces the real value of the money supply, output must fall. The model therefore predicts that nominal national income remains constant, while the proportionate increase in prices equals the proportionate reduction in output. The money supply would have to be increased to offset the output effect, so that NN’ is positively sloped in the classical case. Since the tax increase raises prices, the money supply would have to be reduced to offset this, so the pp’ curve in this case would be negatively sloped.

These results are easily verified by inspection of equations (20) and (21). In the Keynesian case δ2→ – ∞ so that both the iso-employment and iso-price level functions have infinite slopes. In the classical case, δ2 = 0, so that the equations reduce to (dMdT)n=M0ελp0X0>0(dMdT)p=δ1ελ1<0.


Mundell, “The Appropriate Use of Monetary and Fiscal Policy” (cited in footnote 15). See also his book, International Economics (New York, 1968), Chapters 14 and 21.


The anatomy of success is described by Arthur M. Okun, “Measuring the Impact of the 1964 Tax Reduction,” in Walter W. Heller, ed., Perspectives on Economic Growth (New York, 1968), pp. 27-49.


Gordon, “Inflation in Recession and Recovery” (cited in footnote 2), p. 132.


Ibid., p. 115.


Lindbeck, “Theories and Problems in Swedish Economic Policy” (cited in footnote 3), p. 72.




Ibid., p. 72.

IMF Staff papers: Volume 21 No. 3
Author: International Monetary Fund. Research Dept.
  • View in gallery

    Iso-Employment and Iso-Price Level Loci Without Wage Adjustment

  • View in gallery

    Iso-Employment and Iso-Price Level Loci With Wage Adjustment

  • View in gallery

    Iso-Employment and Iso-Price Level Loci When Wage Adjustment Raises the Price Level