Economics 104B - Lecture Notes - Professor Fazzari

Topic VI: Inflation and Unemployment – The “Phillips Curve”

(Final Version: April 1, 2013)

  1. Historical Link Between Inflation and Unemployment

1. U.S. data provide many examples of procyclical inflation

  • As we discussed with our overview of U.S. macro data, inflation usually falls when the economy is weak, and rises when the economy is strong. Thus, inflation is procyclical.
  • Examples of this pattern:
  • The Great Depression: inflation fell so much in the disastrously weak economy that we actually got substantial deflation.
  • Japan in the 1990s: A weak economy let to moderate deflation
  • The early 80s in the U.S.: Inflation fell significantly between 1980 and 1984, due most likely to the deep recession and high unemployment during this period.
  • In the late 1960s in the U.S., a very strong economy seemed to cause inflation to rise.
  • Inflation declined in the middle of the Great Recession (2008 and 2009). Some measures actually showed deflation, although “core” inflation never became negative.
  • The graph below shows the rate of inflation (monthly data measured as the percent change from the same month in the previous year). Two measures are shown. The red line (PCEPI) is the “headline” price index. It includes all consumer goods. The blue line is the “core” deflator for personal consumption expenditures. This is the preferred measure of the Federal Reserve.
  • In this case, “core” means that volatile food and energy are excluded. Note that the core measure is clearly less volatile but it more or less tracks the same patterns as the core measure. Sometimes it is higher and sometimes it is lower.
  • You can see support for the examples presented above. Inflation almost always declines, sometimes a lot, during and right after recessions.

2. Exceptions: 1970s stagflation and, to some extent, 1990s boom
  • In the 1970s the U.S. economy experienced a rise in unemployment and a rise in inflation.
  • The opposite occurred in the 1990s. Technological improvements made possible a reduction in production costs even with low unemployment. Inflation fell modestly from 3 to 4 percent in the early 1990s to about 2 percent by the late 1990s
3. Explanations of procyclical inflation
a) Labor Market
  • If the economy is strong and it is difficult to find workers, competition among firms will likely bid up wages. When there is lots of unemployment, wages will rise more slowly due to the competition among workers looking for jobs. Wages may even fall if things get really bad.
  • Changes in wage costs are likely to be passed through to prices. (Firms set prices as a "mark up" over costs.) Thus, faster wage growth in a strong economy is likely to lead to higher price inflation and slower growth of wages in a weak economy will reduce price inflation.

b) Product Market

  • If firms perceive strong demand for their products they are more likely to raise prices, creating more inflation. Alternatively, if the demand for a firm's products is weak, the firm will probably not raise prices much, if at all. Firms might even cut prices in very weak markets. If this weakness appears across much of the economy, inflation will decline.
B. The Phillips Curve Graph

Inflation

()

Unemployment (u)

1. Inverse relation between inflation and unemployment

  • The graph above shows an "inverse relation" between inflation and unemployment. When the economy is weak and unemployment is high, inflation is low. When the economy is strong and unemployment is low, inflation is high.
  • This relationship was first studied in England by an economist named Phillips. The inflation-unemployment graph is therefore called the "Phillips Curve."
  • We will use the Greek letter pi () to represent the inflation rate.

a) Slope of Phillips Curve

  • Economists typically draw the graph with some "curvature," that is, the slope of the Phillips Curve changes at different levels of unemployment.

b) Possibility of deflation

  • Note that the Phillips Curve cuts through zero on the inflation axis. This captures the fact that deflation can happen, although it will normally be associated with high unemployment.

c) Steeper Phillips Curve at very low unemployment

  • We know from our earlier discussion that unemployment cannot really get to zero. Thus, the Phillips Curve gets very steep as u gets very low. This means that prices can begin to rise very fast in an overheated economy.

d) Flatter Phillips Curve at very high unemployment

  • At high levels of unemployment, further increases in u may have little effect on , so the Phillips Curve gets flat.
2. Shifts of the Phillips Curve
a) Unfavorable supply shocks and stagflation
  • If firms' costs rise, they are likely to pass these costs on to their customers in the form of higher prices (again, this is the mark-up pricing idea). Therefore, a "cost shock" will cause higher inflation, at least for a while.
  • Higher costs will raise inflation for a given level of unemployment. Therefore, the Phillips Curve will shift upward.

b) Historical examples

  • The classic example of this situation is the oil price increases of the 1970s. In the early 70s, war in the Middle East led OPEC (Organization of Petroleum Exporting Countries) to impose an oil embargo on the U.S. Oil prices rose dramatically. Because energy is used in so many industries, the higher oil prices caused big cost increases throughout the economy.
  • There was a similar problem when oil imports from Iran were reduced by the revolution that deposed the Shah of Iran (who was supported by the U.S.) in the late 1970s.
  • Because the Phillips Curve shifts upward, inflation is higher after the cost shock (or "supply shock"). During the early 70s in the U.S., there were also forces raising unemployment. Thus, the Phillips Curve diagram looked like this:

9 % = 1974B

5 % = 1973A

U

U1973U1974(rising energy prices)

(stable energy prices)

  • The combination of a rise in unemployment and a rise of inflation was called "stagflation," that is, stagnation at the same time inflation was rising.
  • The increase in oil prices and the resulting upward shift in the Phillips Curve implies that inflation will rise, but it doesn't tell us what will happen to unemployment. As a first approximation, you can think of unemployment as determined by aggregate demand and the Keynesian Cross model, while the Phillips Curve determines inflation given the unemployment rate.
  • In the mid 1970s, aggregate demand fell and unemployment increased for a number of possible reasons. The higher prices themselves may have reduced demand. In addition, there was a lot of political turmoil during this period that could have reduced business and consumer confidence, lowering investment and consumption.
  • Other examples of negative supply shocks could include bad weather, especially for agricultural economies, and possibly even rising health care costs in the U.S.

c) Positive supply shocks: lower costs and technological improvement

  • There have been favorable supply shocks as well that lead to the opposite of stagflation. For example, oil prices declined significantly in the mid 1980s.
  • Be careful with supply shock terms. A "negative" supply shock is bad for the economy, even though it causes inflation to rise. A "positive" supply shock is good for the economy and causes inflation to fall. Perhaps it would be better to use the terms "unfavorable" supply shock and "favorable" supply shock.
  • Improvements in the technology of production will cause favorable supply shocks. Better technology implies that firms can produce more at lower costs. Competition among firms will reduce inflation in this situation. Thus, the Phillips Curve shifts downward when technology improves.
  • This desirable situation helps explain what happened in the U.S. in the 1990s.
  • Aggregate demand was very strong, in part driven by investment in new technologies. Strong AD pushed output up and unemployment down.
  • Lower costs from better technology, especially in high tech industries, reduced inflation, a downward shift of the Phillips Curve.

C. Inflation Expectations

The Phillips Curve has been part of macroeconomics for nearly half a century. But economists significantly changed their understanding of the "trade-off" between inflation and unemployment in the 1970s. The key change resulted from the observation that changes in inflation expectations would shift the Phillips Curve. To understand these ideas, we need to look more deeply at how wages are negotiated.

1. Role of real wages in labor market bargaining

  • When workers bargain over wages, most economists agree that they really care about the purchasing power of their wage. A wage of $5 per hour would be considered very bad now, in fact it would be illegal because it is well below the minimum wage. Was such a wage high in 1970? To answer this question you need to know what $5 could buy in 1970 relative to what it buys today. Thus, you need to correct the wage for inflation (changes in the price level).
  • Economists define the real wage as the money wage / price index. (The money wage is often called the "nominal wage.") The real wage measures the purchasing power of wages.
  • Using the CPI as the relevant price index, adjusting $5/hour to current prices would give a real wage in 1970 of about $25/hour (in 2013 prices, using the personal consumption expenditure price index). So, $5 per hour in 1970 would be pretty good. (Indeed, Professor Fazzari started working part time in 1971 and would have been thrilled to get $5 per hour. He actually got $1.25 per hour!)
  • Note that if money wages rise as fast prices, the real wage stays constant. Both the numerator and denominator of the real wage grow at the same proportional rate.

2. Self-fulfilling inflation expectations through wage bargains

  • If workers really care about the real wage and they bargain over money wages in advance (say at the beginning of the year), they are going to evaluate any money wage offer relative to expected inflation.
  • If you are offered an 8 percent raise when inflation is expected to be 2 percent (like in recent years), that's a big raise in real terms.
  • But in 1980, when inflation was over 9 percent and people might reasonably have expected inflation to exceed 10 percent, an 8 percent raise would not even keep up with inflation. That is, an 8 percent raise would lead to the expectation that the real wage would fall.
  • Thus, a rise in expected inflation will raise workers' wage demand, and increase firms' costs.
  • Even if workers don't explicitly bargain over wages, firms know that workers are concerned about the purchasing power of their wage. If firms don't offer wage increases high enough to compensate for expected inflation, their workers are more likely to look for other jobs. Again, the point is that higher expected inflation increases wages, and therefore increases firm costs.
3. Shift of the Phillips Curve due to change in inflation expectations
  • Higher inflation expectations raise wages and costs. Therefore, higher inflation expectations lead to a situation very similar to the supply shocks discussed earlier.
  • If firms pay higher wages, they are likely to raise prices to cover their higher costs. Thus, price inflation will rise, even for a given level of unemployment. The Phillips Curve will shift upward:

1

0Phillips Curve (High E)

U0U

Phillips Curve (Low E)

  • Note that inflation expectations can be "self-fulfilling." If people expect higher inflation, they demand higher money wages, which raises firms' costs. To cover their costs, firms will raise prices, which creates actual inflation.
4. Persistent effects of supply shocks on inflation
a) Show how expectations reinforce supply shock effect on inflation
  • Consider a supply shock like the big oil price increase in the middle 1970s. This shock shifts the Phillips Curve upward. Actual inflation rises.
  • It is reasonable to assume that higher actual inflation for a year or two will raise inflation expectations. Thus, even if the supply shock dissipates, high inflation expectations might keep the Phillips Curve higher and inflation higher.
  • Thus, the effect of supply shocks on inflation can be persistent.
b) Examples
  • In the 1970s, higher oil prices created a supply shock that shifted the Phillips Curve upward and raised inflation.
  • The oil prices did not come down (until much later), but they stopped rising not too long after the initial shock. If the only effect on the Phillips Curve was the supply shock, one would expect the Phillips Curve to shift downward not too long after oil prices stopped rising. This would lead to lower inflation.
  • But, as discussed above, the higher inflation resulting from the supply shock likely raises inflation expectations. Higher inflation expectations keep the Phillips Curve and actual inflation high even after oil prices stabilize.
  • This seems to be what happened in the 1970s. Two oil price increases, due to the OPEC embargo and the revolution in Iran, generated supply shocks that would increase inflation temporarily, but the subsequent rise in inflation expectations made the inflation persistent. By the early 1980s, workers in the U.S. clearly had an "inflation mentality." Expected inflation was likely around 8 percent. Some wage contracts included Cost of Living Adjustments (COLAs) to raise wages automatically to offset the effect of inflation. The Phillips Curve was stuck at a pretty high level.
  • In October of 1979, Paul Volcker, the recently named chairman of the Federal Reserve, announced that the Fed would fight inflation aggressively. The result was the deep 1981-82 recession. Consider how these events can be captured by the Phillips Curve graph:

9.1 %= 1980B

6.1 % = 1982C

5 % = 1973A

U0U1982 = 9.7%U

  • The supply shocks plus higher inflation expectations pushed the economy from point A in 1973 to point B by 1980.
  • Higher interest rates pursued by the Volcker Fed pushed the economy into recession and caused a rise in unemployment to U1. Inflation did come down (point C), but at the cost of higher unemployment.
  • Inflation dropped surprisingly quickly in the early 1980s. By the mid 80s is was around 4 percent. Unemployment also recovered during this period. Because unemployment was falling as inflation declined, the best explanation for the middle 80s is a downward shift of the Phillips Curve caused by lower inflation expectations as depicted in the following graph.

9.1 %= 1980 A

6.1 % = 1982B

3 % = 1985C

7.2 % = U1980U1982 = 9.7%U

  • Thus, the Fed was successful at "wringing" inflation and inflation expectations out of the system in the 1980s, but lower inflation came at the cost of one of the worst recessions since the Great Depression.

c) Policy lesson: high cost of reducing inflation expectations

  • Monetary policy makers around the world took an important lesson from this experience. Do not let inflation expectations get too high, otherwise you might have to endure a nasty recession to get them back to an acceptable level. This is part of the reason that central banks seem to be so worried about inflation. They do not want to have to cause unemployment if inflation heats up and drags expectations higher.
  • As the discussion above shows, this lesson applies to supply shocks. But it also could apply to high demand. If demand pushes above Y*, even temporarily, inflation can rise. Reducing demand back to Y* would reduce the inflationary pressure. But if inflation expectations increase, the effect on inflation can be persistent even if the excess demand is temporary. The only way to bring inflation expectations down may be to create a recession. (See the next topic for further discussion.)
5. Pattern of inflation during recession and recovery

a) Effect of recession on unemployment, inflation, and inflation expectations

  • The experience discussed above is a fairly typical pattern around recessions. The recession causes unemployment to rise and inflation to fall.

b) Persistent reduction of inflation due to recession

  • As the economy sits in the recession, wage growth slows and actual inflation is lower. Inflation expectations begin to fall and the Phillips Curve shifts downward. When the economy recovers, it does so with a lower inflation rate because of lower expected inflation.
  • This kind of story is consistent with events in the middle 1980s. As discussed in class, it also explains the approximate behavior of the economy around the 1990-91 recession and subsequent middle 1990s recovery. Inflation fell during the recession, but stayed lower during the recovery. The table below presents the numbers associated with these events. Notice from the highlighted data that inflation remained lower in 1993 and 1994 even with a falling unemployment rate. This phenomenon is could be due to the decline of inflation expectations induced by the actual low inflation rates during the 1990-91 recession and the slow (“jobless”) recovery of 1992. (If you model such a change with Phillips Curve, you would need to shift the Phillips Curve downward after 1992.)

Year /  / U
1989 / 3.8 / 5.3
1990 / 3.7 / 5.6
1991 / 3.6 / 6.9
1992 / 2.2 / 7.5
1993 / 2.3 / 6.9
1994 / 2.1 / 6.1
  • The general pattern associated with these examples is that a recession causes a temporary increase in unemployment, but a permanent reduction in inflation. The reduction in inflation is permanent because inflation expectations are lower even as the economy recovers.
  • In the 1980s for sure, and probably even in the early 1990s, this pattern was somewhat desirable. No one wanted a recession, but the permanent drop in inflation was good for an economy that had an inflation rate higher than was desirable. (Even the 3 to 4 percent inflation prior to the 1990-91 recession was viewed as too high.)
  • In contrast, however, following the economic downturns and slow subsequent slow growth periods after the tech bubble burst and the 2001 recession, and especially after the Great Recession no one really wants inflation to go any lower. Many economists worried about deflation during these periods.