The Phillips Curve

Contributor Notes

Unemployment and inflation are two of the greatest economic problems of governments everywhere. An attempt to demonstrate a consistent, quantitative relationship between them, first launched some 15 years ago, has accordingly aroused immense interest and controversy among economists and has spilled into the general press.

This paper discusses the employment of women in developing countries in the light of recent changes in emphasis on the strategy and objectives of economic development. The paper highlights that in the vast majority of countries—both developed and developing—the role of women is still limited and their responsibilities restricted. This paper examines automated manufacturing techniques in developing economies. The operations and transactions of the special drawing account are discussed. The paper also analyzes Latin America’s prospects for overcoming historical attitudes and other constraints to achieve wider economic integration.

Abstract

This paper discusses the employment of women in developing countries in the light of recent changes in emphasis on the strategy and objectives of economic development. The paper highlights that in the vast majority of countries—both developed and developing—the role of women is still limited and their responsibilities restricted. This paper examines automated manufacturing techniques in developing economies. The operations and transactions of the special drawing account are discussed. The paper also analyzes Latin America’s prospects for overcoming historical attitudes and other constraints to achieve wider economic integration.

Anne Romanis Braun

In a celebrated article that appeared in 1958, Professor A. W. Phillips set out to illustrate how the law of supply and demand applied with respect to wage increases. Taking the unemployment percentage as a measure of the strength of demand relative to the available supply of labor, Phillips argued that

When the demand for a commodity or service is high relative to the supply of it we can expect its price to rise, the rate of rise being greater the greater the excess demand. Conversely, when the demand is low relative to the supply we expect the price to fall, the fall being greater the greater the deficiency of demand. It seems plausible that this principle should operate as one of the factors determining the rate of change of money wage rates, which are the price of labor services.1

When he plotted the data on changes in money wage rates and unemployment in the United Kingdom over almost a century up to 1957, Phillips found a significant and remarkably stable inverse relationship between the change in money wages and the level of unemployment. His diagram for the period 1861-1913 is reproduced as Chart I.

Chart I
Chart I

THE ORGINAL PHILLIPS CURVE

Unemployment and Wage Changes in the United Kingdom 1861…1913.

Citation: Finance & Development 0008, 004; 10.5089/9781616353056.022.A010

Each dot indicates the change in wage rates and the average unemployment percentage for a year.Reproduced from Economica November 1958 page 285.

As originally conceived, the Phillips curve illustrated the relation between the unemployment percentage (u) measured on the horizontal axis and the rate of increase-in wages measured on the vertical axis, as in Chart II (the symbol is frequently used for the change in wages). The concept of a Phillips curve has since been extended also to cover a curve showing the relation between the level of unemployment and the rate of increase in prices (). Phillips himself suggested that, assuming the annual growth in output per man-hour (symbol q˙) is constant, the rate of price increase corresponding to each level of unemployment could be derived from the unemployment/wages curve in the manner illustrated in Chart III. And many economists who have studied the Phillips curve for price increases rather than wage increases have also assumed, implicitly or explicitly, that the main thread of causation runs from increases in wage costs to price changes. It is, however, also usually assumed that a recent change in prices, especially in the cost of living, influences the increase in wages and this has been confirmed by the findings of econometric studies.

Chart II
Chart II

SITUATIONS OF NO TRADE-OFF

Citation: Finance & Development 0008, 004; 10.5089/9781616353056.022.A010

NOTE: These three diagrams show the case where there is a constant (linear) relation between u and ẇ or ṗ over a certain range of u. Chart I shows a changing (non-linear) relation.
Chart III
Chart III

WAGES AND PRICES CURVES: PHILLIPS’ VERSION

Citation: Finance & Development 0008, 004; 10.5089/9781616353056.022.A010

The Trade-Off Between Unemployment and Wage Increases

As it seemed to provide policymakers with a tool for quantifying the possible range of choices between the two major policy objectives of full employment and price stability, the Phillips curve immediately attracted great interest.

Since high levels of employment are certainly more conducive to increases in wages than are low levels, the Phillips curve will slope down from left to right, indicating that over some range of unemployment levels the greater the unemployment percentage the smaller the rise in wages; or, alternatively, that the smaller the unemployment percentage the greater the rise in wages (see lower diagrams in Chart II). This so-called trade-off appears to offer the policymaker a “menu of choices” between employment and inflation.

On this interpretation, the policymaker has only to look at the curve for his country, estimated on the basis of experience, to see whether maintenance of a slightly lower (or higher) level of unemployment than actually prevails would result in a markedly higher (or lower) rate of increase in wages, as it would if the curve were steeply sloping; or if a change in the level of unemployment maintained would have slight effect on the rate of inflation, as would be so if the curve were almost flat. Alternatively, the policymaker can ascertain from the curve what level of unemployment would be required in order to maintain a high degree of price stability.

The relation between unemployment and inflation had no place in the thinking of classical economists. Under their assumptions, full employment was continuously maintained as supply created its own demand in a perfectly competitive and frictionless system (Chart II a). The level of monetary demand, not the level of employment, determined the level of prices and wages.

There will be little or no trade-off in certain situations in the real world. At some high level of employment, the Phillips curve will begin to rise almost vertically because scarcity of labor and shortages of goods induce a wage-price spiral. The level of unemployment at which this will occur is likely to differ between countries, depending upon the level of recorded unemployment at which it becomes difficult for employers to hire workers. For example, in countries with heavy disguised unemployment in low productivity self-employment (such as peasant agriculture and small-scale trading), it will be possible for employers to hire additional labor when unemployment is very low. This will also be so if employers can readily hire foreign workers. In the opposite situation, employers may have difficulty in obtaining labor when recorded unemployment is heavy, if for any reason the unemployment figures include large numbers of persons who are not readily accessible for employment. This may be so for instance if unemployment is concentrated in regions where there are few employment openings and the unemployed are unable or unwilling to move to other parts of the country, or if the method of measuring unemployment causes many persons, such as pensioners and housewives, who are not actively seeking work, to be counted as unemployed.

There may also be little or no trade-off under conditions of high unemployment if the labor organizations are sufficiently powerful to enable employed workers to resist reductions in money wages and to make it difficult for employers to hire workers at less than the going money wage rates (although the unemployed themselves may be prepared to work for lower wages). Phillips himself suggested that this situation would cause the curve to flatten out after crossing the horizontal axis as in Chart I. If money wages were completely inflexible, the curve would coincide with the axis as in Chart II f.

Issues Raised

The Phillips curve raises several issues of vital interest to economists and policymakers. One question concerns the nature of the chain of causation that links the rate of change in wages or prices to the level of unemployment. This has an important bearing upon the question whether the relation between the level of unemployment and the rate of change in wages or prices in the short run that has been observed in the past can reasonably be expected to apply consistently, in widely different circumstances. Clearly, if the short-term relation is a stable one in this sense, it will constitute a useful tool for the policymaker seeking to evaluate the consequences of alternative policies; if not, it may be a dangerously misleading guide for action. Another point at issue is whether a strong association between the level of unemployment and the rate of change in wages or prices in the short run provides a valid reason for supposing that there is a long-run trade-off between the level of unemployment maintained and the rate of inflation experienced in any economy over a period of years.

Causation in a Closed Economy

The interpretation of the Phillips curve proper and of the relation between unemployment and prices poses some difficult and as yet unsettled questions concerning the process of price determination in economies where workers are strongly organized in labor unions, and oligopoly or other forms of imperfect competition prevail in markets for many products. These questions are too far-reaching to be discussed fully here. But one fundamental question is how far the relation between unemployment and wage or price increases actually represents the influence of conditions in the labor market. Now, to adopt one oversimplified view of the causation, one might suppose that an increase in aggregate demand leads to increased demand for labor, which causes employers to experience greater difficulty in hiring workers and strengthens the bargaining power of labor, thus wages tend to increase more rapidly than before, and this leads to a faster rise in prices. Alternatively one might assume that the increase in aggregate demand and consequent increase in pressure on productive resources lessens the incentives for price cutting and makes it easier for businessmen to raise prices, and that this change strengthens the bargaining power of labor, especially of the unions (this is because the extent to which capacity is fully utilized conditions the effectiveness of strike action, working to rule, etc.). In the latter case it is not the level of unemployment per se that influences the rate at which wage costs rise, but the extent of pressure on the productive resources in general, and more especially the degree of utilization of the existing capital stock. In effect then, the level of unemployment relative to that in recent periods serves as a good indicator of pricing conditions in product markets, and this is the underlying cause of the observed relation between unemployment and the rate of increase in wages.

In reality, of course, the truth lies between these two extremes, and both types of causation are likely to be at work at the same time within the economy as a whole. The connection between increases in wages and prices is further complicated both by feedback effects from recent price increases to wage increases owing to the effect of increases in the cost of living in provoking larger wage claims and settlements, and by the fact that increases in wages may give rise to price increases with a lag.

The Phillips Curve of an Open Economy

When one comes to consider an open economy, i. e., a single country in the international setting, it becomes necessary to consider the impact of changes in the level of demand and prices in other countries upon the rate of increase in domestic prices or wages. Then one must visualize not a single curve relating the change in prices to the level of unemployment, but a family of curves each corresponding to a given rate of increase in import prices or demand conditions abroad.

The rise in the general price level would be some weighted average of the increase in import prices and the rise in prices of domestic outputs, even if imports consisted solely of goods not produced in the country (such as mineral and agricultural commodities produced only in other areas of the world), and if increases in the price of such imports had no influence upon wages or the pricing of domestic production processes. However, in countries that are dependent upon imports of basic foodstuffs, fuels, and raw materials, changes in the price of such imports are likely to have an important impact on the cost of living, and are, therefore, liable to influence the rate of increase in wages. Furthermore, in practice, of course, most imports of manufactured goods into highly industrialized countries compete as more or less close substitutes with goods produced in the country. Therefore changes in the price and availability of imports tend to make it easier or more difficult for businesses to raise their prices at a given level of domestic output and unemployment. The strength of demand and the level of capacity utilization abroad also affect the pricing of domestic goods that are exported. Thus, the level of demand existing in other countries is liable to influence the pricing conditions for domestic production associated with any given level of unemployment, and is also liable to influence the size of wage increases that businesses are prepared to concede at that level of employment.

This brief exposition will suffice to indicate that the Phillips relation arises from many different causal links and is the outcome of a complex chain of interactions between the levels of activity, wages, and prices in response to changes in aggregate demand at home and abroad. Empirical estimates of the shape of the Phillips curve proper or of the unemployment/price inflation curve provide valuable insights into the strength and consistency of the association in the past between the current rate of increase in wages and prices and the level of unemployment and other variables (such as the change in unemployment, the movement of import prices, or the rate of increase in prices in preceding periods). But the interpretation of the findings remains difficult and raises a number of thorny problems for the advisor on policy decisions.

Is There a Consistent Trade-Off?

One basic question of course is whether the trade-off can be expected to hold good in differing circumstances. (This is not always clear from the consistency of the association in the past, since experience in the period covered often comprises only a limited range of circumstances.) There are two parts to this question. The first is whether, subject to the influences of other factors such as external demand conditions and lagged price effects, the relation between the level of unemployment and the rate of increase in wages or prices in the short run is likely to be consistently the same. The second is whether the short-term Phillips curve proper or the unemployment/price inflation curve would apply if a certain level of unemployment were to be maintained for several years.

One problem is that, as was suggested earlier, the rate of change in prices and wages may actually reflect the influence of the degree of capacity utilization rather than the effect of conditions in the labor market per se. In this case the amount of unemployment associated with any given increase in prices would depend on the proportion of the labor force that was unemployed when the existing capital stock was fully utilized or when there was a certain smaller or larger margin of unutilized capacity. The position of the curve could then be different at different times, depending on the employment potential of the existing capital stock.

Now the size of the capital stock is likely to reflect the level of aggregate demand experienced in the past. Hence, if the unemployment percentage in fact represents the current degree of underutilization or pressure on capacity, one may expect that an economy in which there has been relatively heavy unemployment for a number of years (such as the United States) will encounter rising prices at a relatively high level of unemployment. Another economy (for instance, Germany) in which, largely as a result of a rapidly expanding export sector, the capital stock has expanded to the point of requiring large-scale employment of foreign workers may encounter a similar degree of pressure on capacity and consequent tendency for prices to increase at a much lower level of unemployment than the former.

Development Planning: Lessons of Experience

4th Printing

by Albert Waterston

A two-part comparative study that assesses the results of planning experience in more than 100 mixed-economy and socialist countries in Africa, Asia, Europe, and the Americas.

Part I describes and analyzes the process in the countries under review, with special emphasis on the problems of implementing development plans. Part II deals with organizational and administrative structures and functions. Theoretical information is included, but the approach is essentially pragmatic.

Appendices contain a comprehensive listing of national development plans (which were improved and expanded in the third printing in 1968), a list of central planning agencies and their addresses, and a bibliography on development planning.

Mr. Waterston is also the author of Planning in Pakistan, Planning in Yugoslavia, and Planning in Morocco, and coauthor of The Economic Development of Mexico.

The U.S. edition of Development Planning is priced at $12.00.

Orders may be sent to

MEXICO

Centro Americano de

Libros Académicos

(CILA).

31 Sullivan-Bis, Mexico 4, D.F.

U.S.A.

The Johns Hopkins Press

Baltimore, Maryland 21218

U.S.A.

CANADA

The Copp Clark Publishing Co.

517 Wellington Street West

Toronto 2B, Ont., Canada

BRITISH COMMONWEALTH

The United Kingdom,

Continental Europe.

The Near and Middle East,

Africa

IBEG Ltd.

2-4 Brook Street

London W1Y-1AA, England

This book has also been published by Dunod, Paris: LA PLAN I FICATION DU DEVELOPPEMENT; by Casa Editrice Dott.

A. Giuffre, Milan: LA PROGRAMMAZIONE OELLO SVILUPPO. LEZIONI DE LL’ESPE RIENZ A; by Fondo de Culture.

Economica, Mexico: PLAN I FICATION DEL DESARROLLO. LECCIONES DE LA EXPER1ENCIA.

Throughout Rest of the World, orders can be sent to The Johns Hopkins Press, Baltimore, Maryland 21218.

AUSTRALIA, NEW ZEALAND, and SOUTHEAST ASIA

Australia and New Zealand Book Co. Pty. Whiting St., Artarmon, N.S.W., 2064, Australia.

JAPAN, United Publishers Services Ltd., Shimura Building, 4-1 Kojimachi Chiyoda, Tokio, Japan

In the first few years after Phillips article appeared, there was a tendency to interpret the empirical relationships in a rather simple manner. Phillips himself took his results to mean that if the average level of unemployment in the United Kingdom were kept at about 2¼ per cent, “the average increase in wages over a period of years could be expected to be about 2 per cent per annum, so that with the rate of increase of productivity of about 2 per cent per annum experienced since the war, the average level of prices would be almost constant.”2 Now, in the United Kingdom activity and employment have fluctuated rather regularly over a four–five year cycle during most of the postwar period. Thus Phillips’ reasoning depended on the implicit assumption that the average rate of increase in wages observed at 2¼ per cent unemployment in the postwar years, when there was usually less unemployment, could be taken as a valid indication of the rate at which wages would rise if unemployment were to be maintained at 2¼ per cent for several years. It seems only reasonable, however, to suppose that if experience has shown that unemployment of 2¼ per cent is unusually high and probably will not be maintained, wage claims will tend to be postponed to more favorable periods. It follows that one may expect wage increases to be bunched in the periods of high employment during the cycle. If so the rate of wage increase associated with, say, 1 per cent unemployment may well be higher, and the rate of wage increase associated with 2½ per cent unemployment lower, than they would be if either level of unemployment were expected to continue. This argument suggests that the trade-off between unemployment and wage increases estimated over a period of regular cyclical fluctuations is likely to differ considerably from the relation which would be found to apply in the absence of regular fluctuations. It also suggests that a country that has experienced regular fluctuations in employment of a few years’ duration will probably be found to have a steeper Phillips curve proper than a country that has not experienced such fluctuations; this would explain the steeper curve observed for the United Kingdom than for the United States or Germany.

In the last year or two it has become increasingly recognized that the relation between unemployment and wage increases observed in the short run will not indicate the relationship that would hold if a certain degree of employment were to be maintained for a period of years and was expected to continue indefinitely. Recognition of this difficulty has led, especially in the United States, to sophisticated efforts to estimate a longer-run “steady-state” Phillips curve from short-term models that include estimates of the influence of the change in the level of unemployment and of the lagged effects of price increases upon wage increases. But it is difficult to see how a meaningful “steady-state” relation can be derived from the short-term observations, if the association between the rate of increase in wages and each level of unemployment reflects not only the level of unemployment but also the expectation of continued cyclical fluctuations, in the manner described above.

The Adjustment of Price Expectations

A number of distinguished economists of the monetarist school believe that no lasting trade-off can be maintained because once the system settles down to any constant rate of price/wage inflation people will recognize the fact and will adjust their income demands accordingly.3 Hence, a steady rise in wages and prices will tend to generate an accelerating increase, and a steady decline an accelerating decline in prices. Therefore, the sloping Phillips curve would be inherently unstable and must, in the long run, inexorably pivot into a vertical position, as shown in Chart IV, implying that a single rate of unemployment would be maintained indefinitely.

Chart IV
Chart IV

LONG-TERM PHILLIPS CURVE: MILTON FRIEDMAN’S VERSION

Citation: Finance & Development 0008, 004; 10.5089/9781616353056.022.A010

See P.A. Samuelson Economics (8th edition) page 812.

Now everyone will agree that large changes in prices will produce a change in economic responses by weakening “money illusion.”4 But both the size of the price changes and the level of unemployment prevailing would seem to be important in this connection. A rapid rate of price increase at close to full employment will quickly provoke a shifting of the curve because wage earners are in a position to secure speedy compensation for price increases; and a large decline, occuring when unemployment is very heavy, may quickly result in a downward shifting of the curve because wage earners will be unable to prevent money wages from falling once employers expect to have to reduce their prices. Thus, one can visualize that the Phillips curve would become almost vertical if very low or very high unemployment levels were to be maintained; but the curve as a whole need not tend to pivot in the manner shown in Chart IV. While there would be limits to the range of unemployment levels that could be maintained indefinitely, there would be no single “natural” level of unemployment.

It was suggested earlier that, in the absence of cyclical fluctuations, the short-run Phillips curve would probably be relatively flat over a considerable range of unemployment levels. One could then imagine some range of unemployment levels at which money wages would rise somewhat faster than productivity, causing prices to rise by some 1-2 per cent per annum; and some considerable range below that at which money wages would tend to rise less than productivity, or to remain unchanged, and prices would be stable or declining very slowly. For the individual consumer or wage and salary earner such slight price movements would scarcely be distinguishable from price stability—bearing in mind the scope for variations in product quality. It is difficult to believe that this portion of the curve would tend to pivot because of weakening “money illusion.” But the curve would be liable to shift for other reasons.

While emphasizing gradual, long-run effects of changing price expectations, the monetarist approach ignores the institutional and structural consequences of maintaining a high or low rate of unemployment for a considerable period of time. One should probably assume that the organization of labor is impeded, and existing unions weakened by the continuance of heavy unemployment, and that the organization of labor proceeds more rapidly and its bargaining power becomes stronger with continued high employment. This implies that the position of the short-run and long-run Phillips curves will be shifting to the left or right when low or high levels of unemployment are maintained for lengthy periods.

It may be easiest to see the implications of this, and of the probable shape of the short-run curve, by considering a possible, not entirely unrealistic, historical sequence. Suppose that there has been a considerable period of heavy unemployment and disguised unemployment, and labor is relatively unorganized. Then, as the classical economists assume, wages in general will only start to rise at close to full employment, and when there is slight unemployment wage earners will be unable to maintain money wages if prices are tending to fall. (Chart V, curve A … A.) However, when a comparatively low level of unemployment is maintained for some time, the power of the unions increases, money wages tend to rise while there is still appreciable unemployment and wage earners are better able to secure larger wage increases if prices rise. Thus, the steep part of the longer-run Phillips curve shifts to the right. The unions also become increasingly able to maintain money wages at high levels of unemployment and to prevent wages falling in line with prices. If businessmen’s price expectations are more responsive to slowly declining prices than wage earners’ willingness to accept lower money wages, a situation could arise in which the development of declining price expectations would cause increasing unemployment rather than rising employment, as Friedman’s model implies. One could imagine that the level of unemployment would gradually increase up to the point where the unions could no longer maintain money wages. Prolonged unemployment and the consequent weakening of the unions would then cause the steeply sloping negative part of the longer-run Phillips curve to shift over to the left.

Chart V
Chart V

POSSIBLE SHIFTING OF THE SHORT RUN PHILLIPS CURVE FOR WAGES, DUE TO THE INFLUENCE OF THE LEVEL OF UNEMPLOYMENT MAINTAINED UPON THE BARGAINING STRENGTH OF LABOR ORGANIZATIONS.

Citation: Finance & Development 0008, 004; 10.5089/9781616353056.022.A010

However, if the unions have acquired the political power to prevent the occurrence of severe unemployment, this sequence of events will not take place. It may be more realistic, therefore, to suppose that the economy is continuously on the positive segment of the long-run curve. This creates a strong inducement for policymakers to seek, by means of incomes policy measures, to shift the Phillips curve to the left or to reduce the steepness of the slope at low levels of unemployment.

Some Conclusions

What conclusions can be drawn from this brief discussion of some of the issues relating to the Phillips curve? The first is that the observed trade-off between unemployment and price-wage increases is an essentially short-term phenomenon. The estimation of Phillips-type relations of the influence upon wages or prices of the level of unemployment and additional variables, such as changes in the level of unemployment, lagged price effects, and changes in import prices, provides the policymaker with a useful means of assessing the apparent strength and consistency of the association between these factors and the rate of increase in wages or prices in the past; and, provided that he interprets the underlying causation of that association correctly, this evidence may be most helpful in evaluating the likely short-term consequences of changes in the level of aggregate demand. It is doubtful, however, whether the observed relations will provide the policymaker with an adequate indication of the rate of inflation that would be associated with the maintenance of a particular rate of unemployment over a period of years. It is unlikely that any marked upward (or downward) movement of prices can be sustained for long, without tending to intensify as economic behavior changes in response to the expectation of continuously changing prices. However, quite a wide range of unemployment levels may consistent with long-term price stability or continued slight upward or downward price movement. Moreover, the Phillips curve proper and the unemployment/inflation curve may be liable to shift gradually over time if close to full employment or heavy unemployment continues for a number of years; so that the limits of the range of the unemployment levels consistent with relative price stability may be changing over time.

1

“The Relation Between Unemployment and the Rate of Change of Money Wages in hte United Kingdom, 1861-1957,” Economica, (London, November, 1958), p. 283.

2

“Employment, Inflation and Growth,” Economica (London, February 1962), p. 11.

3

A highly influential statement of this view was made by Professor Milton Friedman in his Presidential address to the American Economic Association in December 1967, see American Economic Review (March 1968).

4

The term “money illusion,” often used in connection with wage bargaining, reflects the idea that settlements are concluded in money terms. It is the illusion that the purchasing power of money will remain constant. A weakening of money illusion implies the increased tying of all kinds of contractual payments to the level of prices.

Finance & Development, December 1971
Author: International Monetary Fund. External Relations Dept.
  • View in gallery

    THE ORGINAL PHILLIPS CURVE

    Unemployment and Wage Changes in the United Kingdom 1861…1913.

  • View in gallery

    SITUATIONS OF NO TRADE-OFF

  • View in gallery

    WAGES AND PRICES CURVES: PHILLIPS’ VERSION

  • View in gallery

    LONG-TERM PHILLIPS CURVE: MILTON FRIEDMAN’S VERSION

  • View in gallery

    POSSIBLE SHIFTING OF THE SHORT RUN PHILLIPS CURVE FOR WAGES, DUE TO THE INFLUENCE OF THE LEVEL OF UNEMPLOYMENT MAINTAINED UPON THE BARGAINING STRENGTH OF LABOR ORGANIZATIONS.