Phillips curve

The Phillips curve , also known as Phillips curve , is a graph that describes the relationship between changes in wages or price levels on the one hand and the unemployment rate on the other. The Phillips curve was published in 1958 by the English statistician and economist Alban William Housego Phillips in the journal Economica . It has been modified several times since then, for example by Paul A. Samuelson and Robert Merton Solow in 1960 to the so-called extended Phillips curve. This establishes a connection between unemployment and the change in the inflation rate. However, there are further definitions of the Phillips curves in the literature.

history

Older posts

As early as 1926, Irving Fisher had pointed out the relationship between wage changes and unemployment rates for the USA in an essay. Altogether, representations of the relationship between wage increase rates and the unemployment rate before Phillips can be found in other authors: John Law (1671–1729), David Hume (1711–1776), Henry Thornton (1760–1815), Thomas Attwood (1783–1856), John Stuart Mill (1806–1873), Jan Tinbergen , Lawrence Klein and Arthur Goldberger , AJ Brown and Paul Sultan.

Rate of Change of Wages against Unemployment, United Kingdom 1913-1948 from Phillips (1958)

Original Phillips curve 1958

The original Phillips curve from 1958 only graphically depicts a historical-empirical correlation between average nominal wage increases and the unemployment rate . The data were for the period from 1861 to 1957 in Great Britain.

The curve served to illustrate the following two-digit inverse relationship : the higher the unemployment, the lower the wages (and logically implied: vice versa).

Phillips interpreted the correlation as stable because it was legal by introducing the hypothesis that employees with a high level of employment have “greater bargaining power” and can thus enforce higher wages.

This hypothesis follows from the model adopted profit principle of the worker, according to which he wants to maximize his wages, and the minimax principle of the entrepreneur, according to which he has to reduce wage costs in competition with other competitors.

The demand for labor when the supply is falling (fewer jobseekers unemployed) leads to wage increases, because the workers know that the employer cannot find a cheaper replacement for them or the employer can recruit additional workers from another company to activate the hidden reserve or to motivate an unemployed person to pay a higher price for work than what was previously sufficient.

The relationship described was originally expressly not interpreted in such a way that pronounced wage increases lead to higher employment. The causal factor was seen solely in the size of the labor reserve and in the consequent negotiating position of the workers. An employment policy influence on the wage level to increase the number of employees and to reduce unemployment was outside of the research lead interest. Phillips did not yet base his research on a macroeconomic model that incorporated such employment and other macroeconomic issues.

Modified Phillips curve 1960

This only happened through Paul A. Samuelson and Robert M. Solow with the development of the modified Phillips curve . The wage increase rate from the original Phillips model has been replaced by the inflation rate and a fixed, equidistant relationship between changes in nominal wages and changes in price levels has been assumed.

This replacement of wage increases by inflation was based on the theory that entrepreneurs pass on the higher wages to be paid to customers by increasing the price of their products (wage-price spiral), so that a higher wage level leads to a higher price level in the medium term, unless productivity increases Can make production cheaper. Money therefore loses value (inflation) due to price increases, which in turn leads to new demands for wage increases.

Employment policy interpretation of the modified Phillips curve

This conclusion does not contradict the interpretation of the original Phillips curve; However, the negative inverse correlation does not imply any causality and Phillips also rejected a double causality in his interpretation: Higher wages / inflation were not understood as a cause or an instrument for higher demand for employees, i.e. a politically desired decrease in the unemployment rate, but merely as Interpreted result of a low unemployment rate. The modified model, however, was interpreted as an economic policy "menu" that allows politicians to exploit the trade-off , i.e. the correlation between inflation and unemployment, as they wish.

This showed the political explosiveness of the newly interpreted model. Helmut Schmidt , for example, said : “Five percent inflation is easier to bear than five percent unemployment”.

This interpretation of the modified Phillips curve is based not only on economic application but also on the political science model of the party difference hypothesis, which assumes that "left" oriented parties are more likely to accept inflation as a condition for combating unemployment.

stagflation

In the 1970s and 1980s, however, it became apparent that the assumed mutual interdependency of unemployment and inflation did not correspond to reality, because the corresponding economic policy of increasing the money supply by lowering interest rates to generate currency devaluation did not lead to success. Stagflation - the "two-headed monster" - in the form of combined inflation and high unemployment took hold.

Critique of the employment policy interpretation

As early as the late 1960s, Milton Friedman and Edmund S. Phelps independently attacked the misinterpretation of the extended Phillips curve: A causal relationship between a nominal variable such as inflation and a real variable such as unemployment cannot last in the long term because it depends on the Neutrality of money must be assumed. Only with a permanent illusion of money value , i.e. the workers' idea that there would be no inflation, would inflation offset the wage increase again in the long term. However, if workers correctly anticipate inflation, which is generally assumed, inflation has no real impact because workers add the rate of price increases to their wage increase requests. This criticism was hardly heeded until the beginning of the stagflation phase.

Expectation-modified Phillips curve

The criticism, after being confirmed by the economic development, led to the "expectation-modified Phillips curve". This includes the wage earners' inflation expectations.

Keynesian Phillips curve

The Keynesian Phillips curve is linked to the modified Phillips curve, which is based on the relationship between the inflation rate (instead of wage increases) and the unemployment rate. It adopts the justification for the fact that lower unemployment rates tend to be associated with higher inflation rates, in the Phillips principle, but supplements it with the consideration that an increase in employment not only increases the bargaining power of the workers, but also the position of the providers on the Goods markets is strengthened.

The decisive factor for the shape of the Phillips curve is the possibility that employees on the labor markets and companies on the goods markets have to pass the negative consequences of inflation on to the other group in an inflationary process. From this it follows: When the price level rises, the employees try to enforce an inflation adjustment in the wage negotiations, while the companies for their part endeavor to pass on the increase in unit labor costs resulting from wage increases above the increase in labor productivity to the consumers, i.e. the employee households, by means of correspondingly higher prices. This results in a reciprocally driven wage-price-wage spiral, i.e. creeping inflation. Since these pass-on opportunities increase in both groups as the unemployment rate falls, the result is a falling Phillips curve.

The closer the economy approaches full employment, the higher the inflation rate. The Phillips curve only becomes vertical when the positions of the providers on the labor markets (that is, the workers and unions) and the providers on the goods markets are so strong that both can pass on the full burden that they have to face. The quoted sentence of the then Federal Chancellor Helmut Schmidt, on the other hand, is based on the Phillips curve, which falls when only partially passed over. Such an economic policy use assumes, however, that the Phillips curve does not shift over time. This can happen if the announcement by economic policy that it wants to create higher employment through expansionary monetary and fiscal policy encourages companies and unions to increase wage and price increases because they believe that the expansionary monetary and fiscal policy that has been announced guarantees them theirs Jobs and their sales of goods.

Instead of the actual inflation rate, one can also use the expected inflation in the argument. It is then assumed that the wage and price setters try to orientate themselves on this. This alone does not make the Phillips curve perpendicular - it must always be the case that the groups involved are able to pass on the full amount.

Blanchard / Illing (2004, p. 244) make this clear by listing not only the variant in which the actual inflation rate always corresponds to the expected one, but also the variant in which the actual inflation does not reach the expected level because the price and wage-setter do not achieve a full pass-on.

In the subsequent derivation of the unemployment rate, in which the inflation rate remains constant, they tacitly only use the first-mentioned variant (ibid, p. 246). Other textbooks use these variants straight away and thus arrive at the monetarist Phillips curve expanded by expectations.

From a Keynesian point of view, this tacit simplification is to be criticized.

In the empirical verification of the Phillips curve, one can either use the actual inflation rate or - to exclude the effect of particularly volatile prices, the fluctuations of which have nothing to do with the unemployment rate - the core inflation rate - which does not take into account the prices of food and energy sources become.

Monetarian Phillips Curve

The monetarists around Milton Friedman, Karl Brunner and Allan Meltzer criticize both the modified and the Keynesian Phillips curve as inadequate. They argued that monetary and fiscal policy can only affect inflation - not the level of employment. The reason for this is that, from a monetarist point of view, monetary policy has no real effects in the long term (actually economically correct: medium term), but only causes inflation.

From Keynesian side monetarist Phillips curve has undergone a lot of criticism - means the representation of the monetarists but that, above all, monetary policy is not to stimulate the economic growth could be used, but on the maintenance of price stability should focus, anyway nothing worth striving for by a monetary policy that does not pursue strict price stability. The critics see it

a wasted economic potential or

a preliminary economic decision and therefore

Reason for the accusation that the denial that monetary policy has an impact on unemployment, wage level and the share of labor income in national income is based on the motivation that there is no real interest in increasing the wage level, the share of labor income in national income, national income itself and the reduction in unemployment.

However, the message of the monetarists also contains an optimistic content: an economic policy geared towards disinflation does not have to live with the problem of severe job losses.

The monetarist standard model of the Phillips curve looks formally as follows:

{\ displaystyle \ Delta \ omega (t) = a \ cdot \ left (U _ {\ mathrm {n}} -U _ {\ mathrm {tat}} \ right) + b \ cdot \ pi _ {\ mathrm {e} } (t) \ quad {\ text {with}} \ quad a> 0, b> 0}

{\ displaystyle \ pi (t) = \ omega (t)}

{\ displaystyle \ Delta \ pi _ {e} (t) = \ tau \ cdot \ left (\ pi (t_ {0}) - \ pi _ {e} (t_ {0}) \ right) \ quad {\ text {with}} \ quad 0 \ leq \ tau <1}

Here are

{\ displaystyle \ Delta \ omega (t)}

the rate of wage growth with

{\ displaystyle \ Delta \ omega (t) = {\ frac {\ Delta \ omega} {\ Delta t}}}

{\ displaystyle U _ {\ mathrm {n}}}

the natural unemployment rate

{\ displaystyle U _ {\ mathrm {tat}}}

the actual unemployment rate

{\ displaystyle \ Delta \ pi (t)}

the inflation rate with

{\ displaystyle \ Delta \ pi = {\ frac {\ Delta \ pi (t)} {\ Delta t}}}

{\ displaystyle \ pi _ {\ mathrm {e}} (t)}

the expected inflation rate , each for the period .

{\ displaystyle \ Delta t = t-t_ {0}}

Whereby the unemployment rate or unemployment rate (new unemployed / time): denotes. ${\ displaystyle \ tau}$ ${\ displaystyle \ Delta \ pi _ {\ mathrm {e}} (t) = {\ frac {\ pi (t_ {0}) - \ pi _ {\ mathrm {e}} (t_ {0})} { \ tau \ cdot}}}$

The extended Phillips curve

Modified Phillips curve: To see the level of inflation is not accelerating unemployment ( English Non Accelerating Inflation Rate of Unemployment , in short NAIRU )

The extended Phillips curve (or modified Phillips curve ) supplements the considerations of the Phillips curve on the relationship between inflation and the unemployment rate . Here, the change in inflation is related to the unemployment rate.

The following explanations relate to the definition according to Blanchard / Illing. The reason for this is the better overall overview of market developments when considering inflation.

The Phillips curve expanded by expectations

In order to determine the nominal wages for the next year, wage setters must forecast the inflation rate during the next year. The following formula shows that given the expected price level, the same as in the previous year, lower unemployment leads to higher nominal wages.

${\ displaystyle \ pi _ {\ mathrm {t}} = \ pi _ {\ mathrm {t}} ^ {\ mathrm {e}} + (\ mu + z) - \ alpha \ cdot u _ {\ mathrm {t }}}$

With

${\ displaystyle \ pi _ {\ mathrm {t}}}$	= Inflation rate for the year under review
${\ displaystyle \ pi _ {\ mathrm {t}} ^ {\ mathrm {e}}}$	= expected inflation rate
${\ displaystyle \ mu}$	= Profit mark-up factor for prices above wages
${\ displaystyle z}$	= Factors that influence wage setting
${\ displaystyle \ alpha}$	= Effect of the inflation rate on the unemployment rate given inflation expectations
${\ displaystyle u _ {\ mathrm {t}}}$	= Unemployment rate for the year under review

A higher nominal wage leads to a higher price level. Thus, lower unemployment leads to a higher price level compared to the price level from the previous year, i.e. inflation. This is known as the wage-price spiral . Consequently, low unemployment leads to high nominal wages. As a result, the companies raise their prices and the price level rises. Due to rising price levels, employees want higher nominal wages the next time they are set. This results in constant wage and price inflation.

However, if the inflation rate for the year under consideration is zero, it is logical to expect an inflation rate of zero for the forecast year as well.

In the current situation in Germany, mostly positive inflation can be observed. H. the average inflation rate is 3.1%. In the model introduced by Phillips, Samuelson, and Solow, the average inflation rate was close to zero.

Justification for enlargement

Inflation and Unemployment, 1959–1967

The figure on the right shows the relationship between the inflation rate and the unemployment rate in the years between 1959 and 1967. In those years the prognosis for the Phillips curve agreed with the actual values. In the years with high inflation there was a low unemployment rate. Again, there was a low inflation rate in the years with a high unemployment rate. At the beginning of the 1970s, however, there was no connection between the unemployment rate and the inflation rate.

The reason for this was the change in the formation of expectations among wage setters in the course of the 1960s due to a change in inflation. The inflation rate has always been subject to certain fluctuations; sometimes it was positive, sometimes negative. But in the 1960s the inflation rate was consistently positive. In other words, the likelihood that a high inflation rate would be followed by a higher inflation rate in the next year grew. Due to this, the expectations of the wage setters changed. This changed the shape of the relationship between unemployment and inflation.

The following formula, assuming that the expectations are formed as follows, should clarify the relationship:

${\ displaystyle \ pi _ {\ mathrm {t}} ^ {\ mathrm {e}} = \ theta \ cdot \ pi _ {\ mathrm {t-1}}}$

With

${\ displaystyle \ pi _ {\ mathrm {t}} ^ {\ mathrm {e}}}$	= expected inflation rate
${\ displaystyle \ theta}$	= how much inflation rate is taken into account in the formation
${\ displaystyle \ pi _ {\ mathrm {t} -1}}$	= Inflation rate of the previous year

The larger , the more wage setters will raise their inflation expectations. As long as inflation was around 0, it could be expected that the price level in the current year would roughly correspond to the forecast year. During the period considered by Samuelson and Solow, it was therefore close to 0. ${\ displaystyle \ theta}$ ${\ displaystyle \ theta}$

From 1970 onwards, the wage setters changed their expectations due to changes in the inflation rate. From then on, they assumed a steadily rising inflation rate in the following years, whereupon it also rose. ${\ displaystyle \ theta}$

If you insert the above formula into the first formula, you get:

{\ displaystyle \ pi _ {\ mathrm {t}} = \ overbrace {\ theta \ cdot \ pi _ {\ mathrm {t} -1}} ^ {\ pi _ {\ mathrm {t}} ^ {\ mathrm {e}}} + (\ mu + z) - \ alpha \ cdot u _ {\ mathrm {t}} \,}

.

If one assumes , then one obtains ${\ displaystyle \ theta = 0}$

{\ displaystyle \ pi _ {\ mathrm {t}} = (\ mu + z) - \ alpha \ cdot u _ {\ mathrm {t}} \,}

.

If it is positive , the inflation rate is just as dependent on the unemployment rate as it is on the inflation rate of the previous year ${\ displaystyle \ theta}$

{\ displaystyle \ pi _ {\ mathrm {t}} = \ theta \ cdot \ pi _ {\ mathrm {t} -1} + (\ mu + z) - \ alpha \ cdot u_ {t} \,}

.

The formula looks like this, for one after subtracting the inflation rate of the last period on both sides: ${\ displaystyle \ theta = 1}$

{\ displaystyle \ pi _ {\ mathrm {t}} - \ pi _ {\ mathrm {t} -1} = (\ mu + z) - \ alpha \ cdot u_ {t} \,}

Consequently, the unemployment rate does not change the inflation rate, but the change in the inflation rate . That means high unemployment leads to falling inflation, low unemployment leads to an increase in inflation. ${\ displaystyle \ theta = 1}$

This explains what has happened since the 1970s. rose from 0 to 1 and thereupon a connection was formed between the unemployment rate and the change in the inflation rate. ${\ displaystyle \ theta}$

Changes in the inflation rate and unemployment rate in Germany

The diagram opposite shows the relationship between changes in the inflation rate and the unemployment rate for the years since 1980 for Germany. A negative correlation between the unemployment rate and the change in the inflation rate can be seen.

It can be seen from this that the change in inflation is positive when unemployment is low; conversely, the change in inflation is negative when the unemployment rate is high.

The extended Phillips curve thus describes the relationship between unemployment and the change in inflation. It is also often referred to as the modified Phillips curve, the Phillips curve expanded to include expectations, or the accelerating Phillips curve.

Those involved in the wage-setting process changed their expectations for the inflation rate, whereupon the Phillips curve name changed. The insight gained is that the relationship between unemployment and inflation is likely to change with the level and persistence of inflation.

Phillips curve modified by expectations

A further modification of the Phillips curve is reached by considering the inflation expectations of the economic agents. These play an essential role in the effectiveness of monetary policy. If a central bank pursues an expansive monetary policy, according to the modified form this would have to lead on the one hand to higher inflation (monetarist perspective) and on the other hand to a stimulation of the economy and thus to employment growth via the lower interest rates (movement from (1) to (2)) .

According to this understanding, however, the higher employment is only due to the fact that with rising prices and (initially) constant nominal wages, the real wages of employees have declined, which is why companies are hiring more workers. Since workers do not foresee this, this is referred to as surprise inflation . The Phillips curve modified by expectations thus corresponds to the modified one, at least for a short time.

In the medium term, however, employees realize that their wages have not adjusted to the current inflation trend, which is why they are demanding nominal wage increases from their employers to compensate for inflation losses. According to this, nominal wages ultimately rise to the same extent as inflation, which is why employment (if inflation remains the same) returns to its original level (3). Since this occurs with every economic policy influence on inflation, the Phillips curve is vertical in the medium term, according to the monetarist perspective.

The model is based on the assumption of adaptive expectations , i.e. In other words, the economic agents assume that the previous economic policy will also be maintained in the future. However, if one assumes that economic agents have all the relevant information available (assuming rational expectations, see Robert E. Lucas , Thomas Sargent , Robert J. Barro and Neil Wallace ), they will anticipate the surprise inflation induced by the central bank and Demand higher nominal wages at the same time, so that the detour via the short-term perspective is no longer necessary - the Phillips curve would then also be vertical in the short term.

literature

Olivier Blanchard , Gerhard Illing: Macroeconomics. 4th updated and expanded edition. Pearson Studium, Munich et al. 2006, ISBN 3-8273-7209-7 ( Wi - Wirtschaft ).
Bernhard Felderer , Stefan Homburg : Macroeconomics and New Macroeconomics. 7th improved edition. Springer, Berlin et al. 1999, ISBN 3-540-66128-X ( Springer textbook ).
Helge Majer : Modern Macroeconomics. Holistic view. Oldenbourg Wissenschaftsverlag, Munich et al. 2001, ISBN 3-486-25549-5
Hans-Rudolf Peters : Economic Policy. 3rd completely revised and expanded edition. Oldenbourg Wissenschaftsverlag, Munich et al. 2000, ISBN 3-486-25502-9 .

Individual evidence

↑ Olivier Blanchard and Gerhard Illing: Macroeconomics . 3rd edition, Munich 2004
^ Phillips: The Relation Between Unemployment and the Rate of Change of Money Wage Rates in the United Kingdom, 1861-1957
↑ Peters: Economic Policy . P. 90; Bernhard Felderer and Stefan Homburg : Macroeconomics and New Macroeconomics . P. 265; Majer: Macroeconomics . P. 377
↑ Fisher, Irving. 1926. A Statistical Relation between Unemployment and Price Changes. International Labor Review 13 (6): 785-792. Reprinted as "Lost and Found: I Discovered the Phillips Curve." 1973. Journal of Political Economy 81 (2): 496-502.
↑ Hume, David. “Of Money” (1752). Reprinted in his Writings on Economics. Edited by Eugene red wine. Madison: University of Wisconsin Press, 1955.
↑ THE EARLY HISTORY OF THE PHILLIPS CURVE by Thomas M. Humphrey, ECONOMIC REVIEW, SEPTEMBER / OCTOBER 1985
↑ PHILLIPS (1958), The Relation between Unemployment and the Rate of Change of Money Wages in the United Kingdom, 1861-1957; Economica, Vol. 25, pp. 283-299
↑ Samuelson and SOLOW (1960), Analytical Aspects of Anti-Inflation Policy; American Economic Review, Papers and Proceedings, Vol. 50, pp. 177-194
↑ "It seems to me that the German people - to a point - can tolerate a 5% rise in prices rather than 5% unemployment", Süddeutsche Zeitung, July 28, 1972, p. 8, also quoted as: "Better five percent inflation than five percent unemployment ", Attributed by www.spiegel.de (status 04/07)
↑ Olivier Blanchard and Gerhard Illing: Macroeconomics. 3rd edition, Munich 2004.

Web links

Commons : Phillips Curve - collection of pictures, videos, and audio files

[ReferenceA-1] Olivier Blanchard and Gerhard Illing: Macroeconomics . 3rd edition, Munich 2004

[2] Phillips: The Relation Between Unemployment and the Rate of Change of Money Wage Rates in the United Kingdom, 1861-1957

[3] Peters: Economic Policy . P. 90; Bernhard Felderer and Stefan Homburg : Macroeconomics and New Macroeconomics . P. 265; Majer: Macroeconomics . P. 377

[4] Fisher, Irving. 1926. A Statistical Relation between Unemployment and Price Changes. International Labor Review 13 (6): 785-792. Reprinted as "Lost and Found: I Discovered the Phillips Curve." 1973. Journal of Political Economy 81 (2): 496-502.

[5] Hume, David. “Of Money” (1752). Reprinted in his Writings on Economics. Edited by Eugene red wine. Madison: University of Wisconsin Press, 1955.

[6] THE EARLY HISTORY OF THE PHILLIPS CURVE by Thomas M. Humphrey, ECONOMIC REVIEW, SEPTEMBER / OCTOBER 1985

[7] PHILLIPS (1958), The Relation between Unemployment and the Rate of Change of Money Wages in the United Kingdom, 1861-1957; Economica, Vol. 25, pp. 283-299

[8] Samuelson and SOLOW (1960), Analytical Aspects of Anti-Inflation Policy; American Economic Review, Papers and Proceedings, Vol. 50, pp. 177-194

[9] "It seems to me that the German people - to a point - can tolerate a 5% rise in prices rather than 5% unemployment", Süddeutsche Zeitung, July 28, 1972, p. 8, also quoted as: "Better five percent inflation than five percent unemployment ", Attributed by www.spiegel.de (status 04/07)

[10] Olivier Blanchard and Gerhard Illing: Macroeconomics. 3rd edition, Munich 2004.