Академический Документы
Профессиональный Документы
Культура Документы
Business Cycles
Learning Objectives
I. Goals of Part 3
A. What causes business cycles?
B. How should policymakers respond to cyclical fluctuations?
1. Classical economists see little need for government action
2. Keynesian economists think government intervention can smooth the business cycle
C. Coverage of Chapters 8 to 11
1. Business cycle facts and features (Ch. 8)
2. The basic IS–LM model (Ch. 9)
3. The classical model of the business cycle (Ch. 10)
4. The Keynesian model of the business cycle (Ch. 11)
Teaching Notes
I. What Is a Business Cycle? (Sec. 8.1)
A. U.S. research on cycles began in 1920 at the National Bureau of Economic Research (NBER)
1. NBER maintains the business cycle chronology—a detailed history of business cycles
2. NBER sponsors business cycle studies
Data Application
A major compendium of studies on the business cycle was produced by the NBER in 1986, The
American Business Cycle: Continuity and Change, edited by Robert J. Gordon, Chicago:
University of Chicago Press. It contains general discussions of the then-current state of
knowledge of the business cycle, research on components of expenditure and how they change
over the cycle, discussions of the role of fiscal and monetary policies, and research on how the
cycle has changed over time.
Another NBER volume that is a great resource on business cycle information is Victor
Zarnowitz’s book, Business Cycles: Theory, History, Indicators, and Forecasting, Chicago:
University of Chicago Press, 1992. Zarnowitz discusses theories and evidence on the business
cycle, including the NBER’s research role, research on the cyclical characteristics of cycles, an
evaluation of coincident and leading indicators, and a broad discussion of many aspects of
business cycle forecasting.
B. Burns and Mitchell (Measuring Business Cycles, 1946) make five main points about
business cycles:
1. Business cycles are fluctuations of aggregate economic activity, not a specific variable
2. There are expansions and contractions
a. Aggregate economic activity declines in a contraction or recession until it reaches a
trough (Figure 8.1)
Figure 8.1
Data Application
The NBER Business Cycle Dating Committee must wait for some time to pass before they can
declare the start or end of a recession. For example, in the latest recession, the committee
announced in November 2008 that the recession had begun in December 2007; that’s 11 months
after the recession began. And they announced in September 2012 that the recession had ended in
June 2009; that’s 15 months later. The long time lags are necessary because data are often revised
and because the committee wants to ensure that there is truly a turning point in the economy, not
just a temporary change in direction.
For a discussion of the committee’s reasons for picking the dates that begin and end
recessions, see the NBER web site at www.nber.org.
Theoretical Application
Should we even care about the business cycle? Robert Lucas doesn’t think so. In his provocative
book, Models of Business Cycles, Oxford: Basil Blackwell, 1987, he suggests that the cost of
business cycle instability since World War II is very low; in particular, the cost is one-fifth the
cost of having an inflation rate of 10%. So, if faced with the choice of eliminating all recessions
and having a 10% inflation rate, or having recessions the size we’ve had since 1945 and having
no inflation at all, Lucas argues we should take the latter. He suggests that we should move
toward a microeconomic view of the business cycle.
Data Application
New economic theories and statistical techniques may change somewhat the way in which we
look at data on the business cycle. The real business cycle (RBC) approach, which will be
discussed in more detail in Chapter 10, suggests alternative interpretations of business cycle data,
as Finn Kydland and Ed Prescott show in their article, “Business Cycles: Real Facts and a
Monetary Myth,” Federal Reserve Bank of Minneapolis Quarterly Review, Spring 1990, pp.
3–18. They provide a basic set of facts (viewed through the lens of RBC theory) about the
movement of economic variables over the business cycle. They also suggest that inflation is
not procyclical in the United States since World War II, but seems to be countercyclical.
II. The American Business Cycle: The Historical Record (Sec. 8.2)
A. Text Table 8.1 gives the NBER business cycle chronology
B. The pre–World War I period
1. Recessions were common from 1865 to 1917, with 338 months of contraction and
382 months of expansion [compared with 642 months of expansion and 122 months of
contraction from 1945 to 2009]
2. The longest contraction on record was 65 months, from October 1873 to March 1879
C. The Great Depression and World War II
1. The worst economic contraction was the Great Depression of the 1930s
a. Real GDP fell nearly 30% from the peak in August 1929 to the trough in March 1933
b. The unemployment rate rose from 3% to nearly 25%
c. Thousands of banks failed, the stock market collapsed, many farmers went bankrupt, and
international trade was halted
d. There were really two business cycles in the Great Depression
(1) A contraction from August 1929 to March 1933, followed by an expansion that
peaked in May 1937
(2) A contraction from May 1937 to June 1938
e. By May 1937, output had nearly returned to its 1929 peak, but the unemployment rate
was high (14%)
f. In 1939 the unemployment rate was over 17%
2. The Great Depression ended with the start of World War II
a. Wartime production brought the unemployment rate below 2%
b. Real GDP almost doubled between 1939 and 1944
D. Post–World War II business cycles
1. From 1945 to 1970 there were five mild contractions
2. A very long expansion (106 months, from February 1961 to December 1969) made some
economists think the business cycle was dead
3. But the OPEC oil shock of 1973 caused a sharp recession, with real GDP declining 3%, the
unemployment rate rising to 9%, and inflation rising to over 10%
4. The 1981–1982 recession was also severe, with the unemployment rate over 11%, but
inflation declining from 11% to less than 4%
5. The 1990–1991 recession was mild and short, but the recovery was slow and erratic
E. The long boom
1. From 1982 to 2001, there was only one brief recession, from July 1990 to March 1991,
which was not very severe
2. The volatility of many macroeconomic variables declined sharply, so the Long Boom was
the first part of the period known as the Great Moderation
Data Application
When the expansion of the 1990s became the longest in U.S. history in early 2000, the Wall
Street Journal ran a special series of articles, “A Century of Booms, and How They Ended” in its
February 1, 2000 issue.
b. The housing crisis was followed by a financial crisis that rivaled that of the Great
Depression (described in greater detail in Chapter 14)
2. The unemployment rose above 10% for the first time since 1982 and the Fed reduced
interest rates to near zero
3. Economic growth was sluggish even after the recession ended in 2009
Data Application
For an examination of how various macroeconomic variables, especially consumer spending,
performed during the Great Recession, see the article by Mariacristina DeNardi, Eric French, and
David Benson, “Consumption and the Great Recession,” Federal Reserve Bank of Chicago
Economic Perspectives First Quarter 2012, pp. 1–16.
5. New research has focused on the reasons for the decline in the volatility of U.S. output
a. Stock and Watson’s research showed that the decline came from a sharp drop in volatility
around 1984 for many economic variables; dubbed the Great Moderation
b. A plot of real GDP growth (text Figure 8.2) shows that the quarterly growth rate of GDP
was more stable after 1984
c. A plot of the standard deviations of GDP, consumption, and investment (text Figure 8.3)
confirms the decline in volatility
d. Stock and Watson found that the change from manufacturing to services was not a major
cause of the reduction in volatility
e. Stock and Watson showed that evidence that changes in how firms managed their
inventories, which some researchers thought was the main source of the drop in volatility,
was sensitive to the empirical method used, and thus not a convincing explanation
f. Improvements in housing markets may have contributed to the decline in volatility,
but cannot explain the sudden drop in volatility, as those changes occurred gradually
over time
g. Reduced volatility in oil prices was also not an important factor in reducing the volatility
of output
6. After showing that many theories for the reduced volatility in output were not convincing,
Stock and Watson found three factors that were important
a. Reductions in the volatility of food and other commodity prices account for about 15% of
the volatility in output
b. Reduced fluctuations in productivity were responsible for another 15% of the reduction in
output’s volatility
c. Improvements in monetary policy were the most important factor, accounting for 20% to
30% of the reduction in the volatility of output
d. The remaining reduction in output’s volatility remains unexplained–some unknown form
of good luck in terms of smaller shocks to the economy
7. It is not yet clear if the Great Recession implies that the Great Moderation has ended,
though the decline in volatility in 2014 suggests that perhaps the Great Moderation is
continuing
Data Application
Frank Diebold and Glenn Rudebusch argue that although the debate between Romer and others
about the volatility of business cycles before 1929 compared to after 1945 is unsettled, there is clear
and convincing evidence that the duration of expansions and contractions has changed. The average
length of the business cycle remains about the same, but expansions have become much longer and
contractions much shorter. See their article, “Shorter Recessions and Longer Expansions,” Federal
Reserve Bank of Philadelphia Business Review, November/December 1991, pp. 13–20.
Data Application
Your students may enjoy looking at macro data on their own to see what’s out there. Some good
Internet sites that can be used for free are: (1) Federal Reserve Bank of St. Louis FRED database
(research.stlouisfed.org/fred2); (2) Economic Report of the President (www.gpoaccess.gov/eop);
and (3) National Bureau of Economic Research (www.nber.org).
Analytical Problem 3 looks at whether output or total hours worked is more volatile, given that
average labor productivity is procyclical.
c. The job loss rate is the probability that someone who is employed one month will become
unemployed the next month (text Figure 8.9); it declines in expansions and rises in
recessions
d. An example (text Table 10.2) shows that small changes in the job loss rate may lead to
larger changes in the unemployment rate than larger changes in the job finding rate
e. But since the job loss rate applies to many more people, job loss is the main force in
increased unemployment rates during recessions
Analytical Problem 2 asks for an explanation of why expenditures on durable goods are more
volatile over the business cycle than expenditures on nondurables and services.
Data Application
For a basic set of facts about business cycles across countries, see the article by Mario J.
Crucini, M. Ayhan Kose, and Christopher Otrok, “What Are the Driving Forces of International
Business Cycles?” Review of Economic Dynamics, 2011, pp. 156–175.
10. After the fact, the index of leading indicators is revised and appears to have predicted the
recessions well
11. Stock and Watson attempted to improve the index by creating some new indexes based on
newer statistical methods, but the results were disappointing as the new index failed to
predict the recessions that began in 1990 and 2001
12. Because recessions may be caused by sudden shocks, the search for a good index of leading
indicators may be fruitless
F. In touch with data and research: the seasonal cycle and the business cycle
1. Output varies over the seasons: highest in the fourth quarter, lowest in the first quarter
2. Most economic data is seasonally adjusted to remove regular seasonal movements
3. Barsky and Miron’s 1989 study shows that the movements of variables across the seasons are
similar to the movements of variables over the business cycle
Data Application
A further discovery made by Barsky and Miron in looking at the seasonal cycle is surprising:
they found little production smoothing. Economic theory suggests that even if demand changes
over the seasons, production needn’t. Firms could instead produce steadily through the year,
building up inventories of goods in the first three quarters of the year and selling them off in the
fourth quarter. But Barsky and Miron find that this doesn’t happen; production and sales tend to
move together.
4. If the seasonal cycle is like the business cycle, and the seasonal cycle represents desirable
responses to various factors (Christmas, the weather) for which government intervention is
inappropriate, should government intervention be used to smooth out the business cycle?
Policy Application
Some economists have gone so far as far as to challenge the need for the Fed to change the
money supply over the seasons. If the Fed did not increase the money supply in the fall, for
example, the seasonal demand for currency due to holiday shopping would cause interest rates to
rise. Some economists see the rise in interest rates as a natural phenomenon that the Fed should
not prevent. But the case for seasonal monetary policy is based on preventing bank panics (as
occurred frequently from 1890 to 1910) and reducing transactions costs (which arise because
people expend effort to reduce money balances when interest rates rise). For a good discussion of
these issues, see the article by Satyajit Chatterjee. “Leaning Against the Seasonal Wind: Is There
a Case for Seasonal Smoothing of Interest Rates?” Federal Reserve Bank of Philadelphia
Business Review, March/April 1993, pp. 13–24.
Figure 8.2
(1) Short-run equilibrium: the aggregate demand curve intersects the short-run aggregate
supply curve
(2) Long-run equilibrium: the aggregate demand curve intersects the long-run aggregate
supply curve
C. Aggregate demand shocks
1. An aggregate demand shock is a change that shifts the aggregate demand curve
2. Example: a negative aggregate demand shock (Figure 8.3; like text Figure 8.17)
Figure 8.3
Figure 8.4
Policy Application
In his article, “Productivity Growth and the American Business Cycle,” Federal Reserve Bank of
Philadelphia Business Review, Sept./Oct. 1995, Satyajit Chatterjee examines the role of monetary
policy when supply shocks are the major causes of recessions.
Figure 8.7
2. Comovement means that many economic variables move together in a predictable way over the
business cycle. The business cycle facts presented in the chapter illustrate comovement among all the
variables listed in text Summary Table 10 that are either procyclical (moving in the same direction as
aggregate economic activity) or countercyclical (moving in the opposite direction as aggregate
economic activity). Only those variables listed as acyclical do not show comovement.
3. There is some question as to whether or not the business cycle has become less volatile over time.
Originally it was thought that the cycle had been moderated, especially since World War II, but
Romer challenged this notion. Further examination of the data by Balke and Gordon, however, shows
that there has been some moderation of the business cycle.
Whether the business cycle has become less severe or not is important, especially to economic
policymakers. Since World War II, both fiscal policy and monetary policy have been used to try to
smooth out business cycles to reduce their severity. If it were found that business cycles are no less
severe than they used to be, it would point to the failure of government policy to achieve its
objectives.
4. A variable that moves in the same direction as aggregate economic activity is said to be procyclical,
while a variable that moves in the opposite direction is countercyclical. If the peaks and troughs of a
variable occur before the peaks and troughs in aggregate economic activity, it is said to be a leading
variable. If a variable’s peaks and troughs occur at the same time as the peaks and troughs in
aggregate economic activity, it is said to be a coincident variable. If a variable’s peaks and troughs
come after the peaks and troughs of aggregate economic activity, it is said to be a lagging variable.
5. If the economy were entering a recession, you’d expect production, investment, average labor
productivity, and the real wage to decline because they are all procyclical, and the unemployment rate
to rise because it’s countercyclical.
6. The fact that some economic variables are known to lead the business cycle is used to develop an
index of leading economic indicators. The index is used to forecast economic turning points.
7. The two components of a theory of business cycles are: (1) A description of the types of factors
(called “shocks”) that have major impacts on the economy, such as wars, new inventions, harvest
failures, and changes in government policy; and (2) a model of how the economy responds to the
various shocks.
8. Keynesians and classicals differ sharply in their beliefs about how long it takes the economy to reach
a long-run equilibrium. Classical economists believe that prices adjust rapidly (within a few months)
to restore equilibrium in the face of a shock, while Keynesians believe that prices adjust slowly,
taking perhaps several years.
Because of the time it takes for the economy’s equilibrium to be restored, Keynesians see an
important role for the government in fighting recessions. But because classicals believe that
equilibrium is restored quickly, there’s no need for government policy to fight recessions.
Since classicals think equilibrium is restored quickly in the face of shocks, aggregate demand shocks
can’t cause recessions, since they can’t affect output for very long. So classical economists think
recessions are caused by aggregate supply shocks. Keynesians, however, think that both aggregate
demand and aggregate supply shocks are capable of causing recessions.
Analytical Problems
1. Figure 8.8 illustrates the business cycle. The current NBER method picks peaks and troughs in the
level of aggregate economic activity, which are points on the figure where the slope of the line is
zero. These are shown in Figure 8.8 as P1 (at the peak of the cycle) and T1 (at the trough of the cycle).
However, the older method picks peaks and troughs in detrended economic activity. This means the
peaks and troughs occur at points that are the farthest away from the trend line, which means those
points at which the slope of the line showing aggregate economic activity is the same as the slope of
the trend line. These points are shown in Figure 8.8 as P2 and T2. Note that the point P2 occurs before
P1, meaning that peaks in detrended economic activity are earlier than peaks in the level of economic
activity. Note also that the point T2 occurs later than T1, which means that troughs in detrended
economic activity are later than troughs in the level of economic activity. Since under the old method,
troughs occur later and peaks occur earlier, contractions appear to be longer and expansions appear to
be shorter using the pre-1927 method than using the current method. Thus the fact that after World
War II expansions were longer and contractions were shorter than before World War I is somewhat
illusory, since it’s based on two different accounting mechanisms. If expansions and contractions
were in fact equally long in both periods, the change in accounting method would mean that our official
dating of the business cycle would show longer expansions and shorter contractions after World War II
than before World War I.
Figure 8.8
2. Expenditure on durable goods is more sensitive to the business cycle than expenditure on nondurable
goods and services, because people can more easily change the timing of their expenditure on durables.
When economic activity is weak, and people face the danger of losing their jobs, they avoid making
durable goods purchases. Instead, they may drive their cars a little longer before buying new ones, get
the old washing machine repaired instead of buying a new one, and put off buying new furniture until
a new expansion indicates greater income security. So in a recession, durable purchases decline a lot,
but when an expansion begins, durable purchases pick up substantially. The exception was in the
business cycle that began in March 2001, when very low interest rates supported expenditures on
durable goods.
3. (a) In symbols, let A average labor productivity, Y output, and H total hours worked. By
definition, A Y/H, so in growth terms, A/A Y /Y H/H. Since all three are procyclical,
they all move in the same direction over the business cycle. If total hours worked varied more
than output in an expansion, then H/H would be greater than Y/Y, so that A/A would be
negative, and average labor productivity would be countercyclical. So it must be the case that
output varies more than total hours worked in an expansion. A similar argument holds in a
contraction.
(b) That average labor productivity is procyclical helps explain why the Okun’s Law coefficient is 2,
not 1. A one-percentage point increase in unemployment is approximately a one percent fall in
employment. Thus, if there were no change in average labor productivity, we might expect the
percentage fall in output to equal the number of percentage points that the unemployment rate
rises. But since average labor productivity moves in the same direction as output, it magnifies the
output effect of a given amount of unemployment.
4. Figure 8.9 illustrates the effects of a demand shock. The economy begins in equilibrium at point A,
where the LRAS, SRAS, and AD curves intersect. The demand shock shifts the aggregate demand
curve to the left to AD . In the short run, the equilibrium is at point B, where AD intersects SRAS.
This is a point at which output has declined (a recession), but the price level is unchanged. Over time,
the short-run aggregate supply curve shifts down to SRAS , restoring long-run equilibrium at point C.
At this point, output is back at its full-employment level and the price level has declined. Thus the
result of a demand shock on the price level is that the price level is unchanged in the short run and
declines in the long run. Since the 1973–1975 recession was one in which the price level rose sharply,
it must not have been due to a demand shock.
Figure 8.9
Figure 8.10 illustrates the effects of a supply shock. The economy begins in equilibrium at point A,
where the LRAS, SRAS, and AD curves intersect. The supply shock shifts the long-run aggregate
supply curve to the left to LRAS . The new equilibrium is at point B, where AD intersects LRAS . This
is a point at which output has declined (a recession), but the price level has risen. This matches what
happened in the 1973–1975 recession. Thus we conclude that the 1973–1975 recession was the result
of a supply shock, not a demand shock.
Figure 8.10
5. Growth that is “too rapid” most likely refers to a situation in which the aggregate demand curve has
shifted to the right and, in the short run, intersects the SRAS curve at a level of output that’s greater
than the full-employment level of output (Figure 8.11). This situation is associated with inflation
because, in the long run, prices will rise, shifting the SRAS curve up to intersect with the LRAS and
AD curves. The shock that is implicitly assumed to be hitting the economy is an aggregate demand
shock, since that’s the only shock that increases output in the short run and inflation in the long run.
Figure 8.11
1. The unemployment rate is a persistent variable because most changes by 0.2 percentage points or more
were followed in the subsequent quarter by another change in the same direction.
3. Real stock prices often decline in the middle of economic expansions, so not every decline in real stock
prices is followed by a recession.
4. The business cycles of the United States and Canada are most closely related. Business cycles are
somewhat related for the United States and Germany but not as closely as the United States and
Canada.