Site Loader

Most economists and historians credit the origin of the unemployment-inflation
tradeoff to the British economist William Phillips, hence the name ‘Phillips
Curve’. His 1958 paper looked at data from three different time periods which
supported his hypothesis about a non-linear relationship between unemployment
and the rate of change of money. Even though it might have been Phillips
seminal Economica paper which brought this idea to the forefront of
policymaking and evaluation there were several economists and philosophers
before him who noticed the relationship and wrote about it. Throughout the 260
years that have passed since the tradeoff was first studied the empirical
evidence behind it has been constantly evolving and the data updated, however,
the rationale behind the original thinkers still holds until today.

Thomas Humphrey in his 1985 review The Early History of the Phillips Curve talked about the origins
and the thinkers behind the unemployment-inflation relationship. In 1752, David
Hume, an Scottish economist, argued that unemployment variations occurred
because of price perception errors. He believed that in order to promote
economic activity there needs to be a steady increase in prices and wages and
not a constant high price and wage level. Similar to Hume, Henry Thornton in
1802 argued that a monetary expansion stimulates employment by raising overall
price levels in his paper An Enquiry into
the Nature and Effects of the Paper Credit of Great Britain. Even though
his argument aligns with that of Hume’s in the sense that there needs to be a
steady increase and not a one-time shock in the money supply in order to
stimulate the economy, he disagreed with Hume’s policy applicability of their
discovery. Hume’s argument regarding price perceptions errors was later
reinforced by John Stewart Mill during the 1820’s. Mill believed that the
tradeoff occurs only when perceived and real prices are different since there
is a price adjusting period in which unemployment is impacted. Disputing Mill’s
work, during that same time period, Thomas Atwood argued that it is the role of
the central government to continuously depreciate their currency until full
employment is reached. Contrary to Mill, Atwood believed that full employment
was needed in order to achieve prosperity.

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!

order now

During the late 1700s’ and early 1800s’ most thinkers, including
Hume, based their arguments on perceptions of the economy, observations and
their own rationality. Even though some of their arguments still hold until
today and laid the groundwork for Phillips’s work, none of them presented
factual evidence to support their claims, mostly because of the lack of
availability. In 1926, almost 175 years after Hume’s work, Irving Fisher
published A Statistical Relationship between
Unemployment and Price Changes. His work provided factual evidence of a strong
bivariate correlation between unemployment levels and lagged price changes. Even
though at the time his paper “attracted little attention…because it was
published in the International Labour
Review” (Puttaswamaiah 1996) it was later republished in 1973 under the
title I Discovered the Phillips Curve
in the Journal of Political Economy. Due
to the relevance of his study and in an effort to understand his empirical
research we will pay considerable attention to Fisher’s 1973 paper.

Fisher looked at data of the United Stated for the 1915-1925
period and noticed a high correlation between price changes and employment. He
argued that “the correlation … is sufficiently high to say that for the period
considered, changes in the purchasing power of the dollar may very largely
explain changes in employment”. He followed a rather simple logic arguing that
when prices are rising, profits increase faster than expenses and firms are
willing to hire more labor therefore reducing unemployment. We will see later
on that this was also one of Phillips’s hypotheses. Similar to Hume and
Thornton, Fisher argued that it is not a constant high level of inflation which
creates unemployment but rather a steady increase in inflation which produces
unemployment. This belief is exemplified in his analysis in which he compared
the relationship between prices levels and price changes with employment. The
following graph represents his analysis of the fluctuations in price levels:

He plots the 1903-1925 price index, price change and price
change projected against each other. Fisher, himself being the thinker behind
lag theory, added the price change projected variable since according to him it
works “as a moving average of the price change” (Fisher 1973). He argues that since
the effects on employment are not felt immediately after changes in price, a
lagged variable must be created. This variable represents the “composite
effects of the rates of rise and fall of the price level” (Humphrey 1985). We
can also see graphically that there were abrupt price changes which empirically
allow for an easier comparison than it would have otherwise with constant
stable prices. In order to see the relationship between price changes and
employment, we turn to look at employment levels and the created price change
projected variable. The following graph represents such comparison using
employment data from the Harvard Committee of Economic Research:

We can graphically see that there is strong relationship
between employment and lagged price changes. Fisher also argued that periods
with less abrupt price changes demonstrated a lower correlation than those with
more abrupt price changes. In order to capture the tradeoff between employment
and lagged price changes Fisher performed three different statistical analyses:
1) employment on price change with a four month fixed lag, 2) employment on
price change with a three month fixed lag and 3) employment on price change
with a, according to him, properly distributed lag (shown on the above graph).
The first regression produced a 79 percent correlation, the second 62 percent
and the third 90 percent. For the third regression he adjusted the lagged price
change variable to a normal probability distribution since it was the one that
best adjusted to the price data.

Overall Fisher’s work reinforced previous literature, was
clear and concise and most importantly got the job done: statistically proved
that a tradeoff exists. Even though his work on monetary and financial
economics was groundbreaking there was not much attention given to this
specific piece of work probably because of its lack of policy applicability at
the time. From my perspective he lacks support for his adjustment of the price
change variable and he is also incurring in omitted variable bias. It is
important to point out here that Fisher argued for a causal relationship from
price changes to employment, not the other way around. However, in 1936, only
10 years after Fisher’s analysis, Jan Tinbergen argued that it is in fact
unemployment which produces inflation in his paper An Economic Policy for 1936. This conclusion contradicted all
previous literature since Tinbergen argued that the tradeoff worked as a function
of a reactionary wage instead of a function of prices like everyone before him.
Tinbergen used data for the 1923-1933 period for the Netherlands and according
to Humphrey he presented the “first econometric equation which estimated the
tradeoff”. Even though Tinbergen’s 1936 study omitted certain variables it challenged
established beliefs and by doing so added economists to the conversation. In
1951, he produced additional studies in which he included variables like the
cost of living and unionization in order to improve the statistical fit using
data for the United Kingdom (Humphrey 1985). Even with the added variables and
a better statistical model he reached the same conclusion. Both Fisher and
Tinbergen had contrasting beliefs as to the direction of the causality of the
tradeoff but empirically speaking both of them were able to empirically prove
that there is a tradeoff.

The literature, which started with Hume in the 1750s’, had
evolved over 200 years to a more macro-oriented policy tool. In 1955, Klein and
Goldberger followed Tinbergen’s work and looked at the tradeoff using data from
the United States for the 1929-1952 period. What was interesting about their
study was that they were the first to use data for the United States and that
the data for that period represented the years of the Great Depression. During
those years, starting in 1929 with the stock market crash, the level of
unemployment skyrocketed and the level of inflation remained low and volatile compared
to previous years. From my perspective, the high unemployment levels during
those years was not the driver of inflation since there were many coinciding
factors. However, the fact that Klein and Goldberger reached results similar to
those of Tinbergen with contrasting data samples (booming Dutch economy and a
crumbling economy in the United States) strengthened the idea of causality
moving from unemployment to inflation. In 1955 and 1957, A.J Brown and Paul
Sultan used data from the United Kingdom and the United States as empirical evidence
of the relationship. Both Brown and Sultan produced Phillips-type graphs and
analyses a couple years before Phillips published his work. Brown looked at
data for the United Kingdom during a period of booming foreign trade and the
Edwardian era, similar time period than Phillips. The following graphs
represent their analysis:


Almost 250 years since Hume’s work, in 1958, A.W Phillips
published his seminal paper The Relation
Between Unemployment and the Rate of Change of Money Wage Rates in the United
Kingdom, 1861-1957. In his analysis Phillips states three main hypotheses
which coincide with much of the previous literature. He argues that there are
three main factors which determine the rate of change of money:

1st Factor: rate of change of the price of labour
services (wages)

When the demand for labor is high and unemployment is low
then firms are willing to offer higher wages to attract the right talent and
they are willing to do it fast, causing wage rates to rise rather quickly. On
the other hand, when the demand for labor is low and unemployment is high
workers are not immediately willing to accept lower than preferred wages.
However, after some time of being unemployed and because of necessity they are
forced to accept lower wages. This transition causes wage rates to decrease
slowly. This logic demonstrates that the relationship between inflation and
unemployment is highly non-linear.

2nd Factor: rate of change of the demand for
labour (employment – unemployment)

When there is a rise in business activity employers are
likely to raise wages since their demand for labour is increasing due to a
needed increase in production. They can also afford this increase because their
revenue is also increasing. This rise in the wage rate happens rather fast. On
the other, when there is a slowdown in business activity employers will be
hesitant to raise wages due to a decrease in demand, unemployment would also be
increasing causing a slowdown of wage rates.

3rd Factor: rate of change of retail prices (consumer

Post Author: admin


I'm Erica!

Would you like to get a custom essay? How about receiving a customized one?

Check it out