Chapter 10
Classical Business Cycle Analysis:
Market-Clearing Macroeconomics

Goals of Chapter 10

A) Use the IS-LM model with rapidly adjusting wages and prices to present the classical model

B) Examine the relationship between money and the business cycle



I.   Business Cycles in the Classical Model (Sec. 10.1)

A) The real business cycle theory

1. Two key questions about business cycles

a. What are the underlying economic causes?

b. What should government policymakers do about them?

2. Any business cycle theory has two components

a. A description of the types of shocks believed to affect the economy the most

b. A model that describes how key macroeconomic variables respond to economic shocks

3. Real business cycle (RBC) theory (Kydland and Prescott)

a. Real shocks to the economy are the primary cause of business cycles

(1)        Examples:

a.   Shocks to the production function,

b.   the size of the labor force,

c.   the real quantity of government purchases,

d.   the spending and saving decisions of consumers (affecting the IS curve or the FE line)

(2) Nominal shocks are shocks to money supply or demand (affecting the LM curve)

b. The largest role is played by shocks to the production function, which the text has called supply shocks, and RBC theorists call productivity shocks

(1)        Examples:

a.   Development of new products or production techniques,

b.   introduction of new management techniques,

c.   changes in the quality of capital or labor,

d.   changes in the availability of raw materials or energy,

e.   unusually good or bad weather,

f.     changes in government regulations affecting production

(2) Most economic booms result from beneficial productivity shocks; most recessions are caused by adverse productivity shocks

c. The recessionary impact of an adverse productivity shock: Do this in class

(1)        Results from Chapter 3:

a.   Real wage,

b.   employment,

c.   output,

d.   consumption, and

e.   investment decline,

f.     while the real interest rate and price level rise.

(2) So an adverse productivity shock causes a recession (output declines), whereas a beneficial productivity shock causes a boom (output increases); but output always equals full-employment output

d. Real business cycle theory and the business cycle facts

(1) The RBC theory is consistent with many business cycle facts

(a)      If the economy is continuously buffeted by productivity shocks, the theory predicts recurrent fluctuations in aggregate output, which we observe

(b)     The theory correctly predicts procyclical employment and real wages

Assign this problem

Numerical Problem 1 looks at the relationship between real wages and employment over the business cycle and the issue of whether the labor-supply curve should be flat or steep to be consistent with the data.

(c)      The theory correctly predicts procyclical average labor productivity. If booms weren’t due to productivity shocks, we would expect average labor productivity to be countercyclical because of diminishing marginal productivity of labor

(2) The theory predicts countercyclical movements of the price level, which seems to be inconsistent with the data

(a)      But Kydland and Prescott, when using some newer statistical techniques for calculating the trends in inflation and output, find evidence that the price level is countercyclical.

(b)     Although the Great Depression appears to have been caused by a sequence of large, adverse aggregate demand shocks, Kydland and Prescott argue that since World War II, large adverse supply shocks have caused the price level to rise while output fell

(c)      The surge in inflation during the recessions associated with the oil price shocks of 1973–1974 and 1979–1980 is consistent with RBC theory

Data Application

The “bible” of empirical work on RBC models is by Robert King, Charles Plosser, and Sergio Rebelo of the University of Rochester, “Production, Growth and Business Cycles: Technical Appendix,” May 1987. This is an appendix to two articles published in the Journal of Monetary Economics in 1989 that have become the cornerstone for subsequent empirical work using the RBC approach. The appendix explains in detail how the RBC model is analyzed and calibrated, and it describes computer programs written for the software package MATLAB.


4. Application: Calibrating the business cycle

a. A major element of RBC theory is that it attempts to make quantitative, not just qualitative, predictions about the business cycle

b. RBC theorists use the method of calibration to work out a detailed numerical example of the theory

(1) First they write down specific functions explaining the behavior of people in the economy; for example, they might choose as the production function for the economy, Y = AKaN1–a

(2) Then they use existing studies of the economy to choose numbers for parameters like a in the production function; for example, a = 0.3

(3) Next they simulate what happens when the economy is hit by various shocks to different sectors of the economy

(4)        Prescott’s computer simulations (text Figures. 10.1 and 10.2) match post–World War II data fairly well




Data Application

The work on calibration has led to a major scientific debate within the economics profession about how to do empirical work. Economists working on RBC models, led by Prescott, believe strongly in calibration as the only way to do empirical work in macroeconomics. Others disagree, just as vehemently. An illuminating debate by many of the leading antagonists can be found in the symposium on “Computational Experiments in Macroeconomics” in the Journal of Economic Perspectives, Winter 1996.



5. Are productivity shocks the only source of recessions?

a. Critics of the RBC theory suggest that except for the oil price shocks of 1973, 1979, and 1990, there are no productivity shocks that one can easily identify that caused recessions

b. One RBC response is that it doesn’t have to be a big shock; instead, the cumulation of many small shocks can cause a business cycle (text Figure 10.3)



Assign this Problem


Numerical Problem 6 is a coin-flipping exercise to show that random shocks can lead to big aggregate movements.

6. Does the Solow residual measure technology shocks?

a. RBC theorists measure productivity shocks as the Solow residual

(1) Named after Robert Solow, the originator of modern growth theory

(2) Given a production function, Y = AKaN1-a, and data on Y, K, and N, the Solow residual is

                                    A = Y/(KaN1–a)                            (10.1)

(3) It’s called a residual because it can’t be measured directly

b. The Solow residual is strongly procyclical in U.S. data

(1) This accords with RBC theory, which says the cycle is driven by productivity shocks

c. But should the Solow residual be interpreted as a measure of technology?

(1) If it’s a measure of technology, it should not be related to factors that don’t directly affect scientific and technological progress, like government purchases or monetary policy

(2) But statistical studies show a correlation between these

d. Measured productivity can vary even if the actual technology doesn’t change

(1) Capital and labor are used more intensively at times

(2) More intensive use of inputs leads to higher output

(3) Define the utilization rate of capital uK and the utilization rate of labor uN

(4) Define capital services as uK K and labor services as uN N

(5) Rewrite the production function as

                Y = AF(uK K, uN N ) = A(uK K )a (uN N )1–a             (10.2)

(6) Use this to substitute for Y in Eq. (10.1) to get

                          Solow residual = A uKa uN1–a                 (10.3)

(7) So the Solow residual isn’t just A, but depends on uK and uN

(8) Utilization is procyclical, so the measured Solow residual is more procyclical than is the true productivity term A

(a)      Burnside-Eichenbaum-Rebelo evidence on procyclical utilization of capital

(b)     Fay-Medoff and Braun-Evans evidence on procyclical utilization of labor

i)  Labor hoarding: firms keep workers in recessions to avoid incurring hiring and firing costs

ii) Hoarded labor doesn’t work as hard, or performs maintenance

iii) The lower productivity of hoarded labor doesn’t reflect technological change, just the rate of utilization

e. Conclusion: Changes in the measured Solow residual don’t necessarily reflect changes in technology

7. Technology shocks may not lead to procyclical productivity

a. Research by Basu and Fernald shows that technology shocks are not closely related to cyclical movements in output

b. Shocks to technology are followed by a transition period in which resources are reallocated

c. Initially, less capital and labor are needed to produce the same amount of output

d. Later, resources are adjusted and output increases

8. Also, the critics suggest that shocks other than productivity shocks, such as wars and military buildups, have caused business cycles

Theoretical Application

For more on criticisms of the RBC theory and the RBC response to the critics, see the discussion in the Federal Reserve Bank of Minneapolis Quarterly Review, Fall 1986, and the Journal of Economic Perspectives, Summer 1989.

B) Fiscal policy shocks in the classical model

1. The effects of a temporary increase in government expenditures (Figure 10.1; like text
Figure 10.4)

Figure 10.1

a. The current or future taxes needed to pay for the government expenditures effectively reduce people’s wealth, causing an income effect on labor supply

b. The increased labor supply leads to a fall in the real wage and a rise in employment

c. The rise in employment increases output, so the FE line shifts to the right

d. The temporary rise in government purchases shifts the IS curve up and to the right as national saving declines

e. It’s reasonable to assume that the shift of the IS curve is bigger than the shift of the FE line, so prices must rise to shift the LM curve up and to the left to restore equilibrium

f. Since employment rises, average labor productivity declines; this helps match the data better, since without fiscal policy the RBC model shows a correlation between output and average labor productivity that is too high

g. So adding fiscal policy shocks to the model increases its ability to match the actual behavior of the economy

Analytical Problems 2, 3, and 5 deal with various aspects of the classical IS-LM model.

2. Should fiscal policy be used to dampen the cycle?

a. Classical economists oppose attempts to dampen the cycle, since prices and wages adjust quickly to restore equilibrium

b. Besides, fiscal policy increases output by making workers worse off, since they face higher taxes

c. Instead, government spending should be determined by cost-benefit analysis

d. Also, there may be lags in enacting the correct policy and in implementing it

(1) So choosing the right policy today depends on where you think the economy will be in the future

(2) This creates problems, because forecasts of the future state of the economy are imperfect

e. It’s also not clear how much to change fiscal policy to get the desired effect on employment and output

C) Unemployment in the classical model

1. In the classical model there is no unemployment; people who aren’t working are voluntarily not in the labor force

2. In reality measured unemployment is never zero, and it is the problem of unemployment in recessions that concerns policymakers the most

3. Classical economists have a more sophisticated version of the model to account for unemployment

a. Workers and jobs have different requirements, so there is a matching problem

b. It takes time to match workers to jobs, so there is always some unemployment

c. Unemployment rises in recessions because productivity shocks cause increased mismatches between workers and jobs

d. A shock that increases mismatching raises frictional unemployment and may also cause structural unemployment if the types of skills needed by employers change

e.   So the shock causes the natural rate of unemployment to rise; there’s still no cyclical unemployment in the classical model



Theoretical Application

A nice discussion of the classical view of unemployment is by Robert E. Lucas, Jr., Models of Business Cycles, Chapter V, New York: Basil Blackwell, 1987.

4. Davis and Haltiwanger show that there is a tremendous amount of churning of jobs both within and across industries

Data Application

For everything you’ve ever wanted to know but were afraid to ask about how jobs change over the business cycle, see the book by Davis, Haltiwanger, and Schuh, Job Creation and Destruction (Cambridge: MIT Press, 1996).

5. But this worker match theory can’t explain all unemployment

a. Many workers are laid off temporarily; there’s no mismatch, just a change in the timing of work

b. If recessions were times of increased mismatch, there should be a rise in help-wanted ads in recessions, but in fact they fall

6. So can the government use fiscal policy to reduce unemployment?

a. Doing so doesn’t improve the mismatch problem

b. A better approach is to eliminate barriers to labor-market adjustment by reducing burdensome regulations on businesses or by getting rid of the minimum wage


Numerical Problem 7 looks at the behavior of the unemployment rate due to a temporary productivity shock when there are many people in transition between being employed and unemployed.

D. Household production

1. The RBC model matches U.S. data better if the model accounts explicitly for output produced at home

2. Household production is not counted in GDP but it represents output

3. Rogerson and Wright used a model with household production to show that such a model yields a higher standard deviation of (market) output than a standard RBC model, thus more closely matching the data

4. Parente, Rogerson, and Wright showed that after household production is accounted for, income differences across countries are not as large as the GDP data show

Theoretical Application

In most macroeconomic models, including the IS–LM and AD–AS models, the key variables are economy-wide averages of income, the wage rate, wealth, money holdings, and so on. But some issues in macroeconomics are better addressed in models in which agents in the model (agents are decision-makers such households and business firms that decide how much to consume or invest) act in different ways or face different wages or have differing amounts of wealth; such models are heterogeneous-agent models. For example, to understand how the unemployment rate changes over time, a model of the demographics of the labor force (the number of workers of different ages, different levels of experience, and different levels of education) is useful. In recent years, more macroeconomists have begun building heterogeneous-agent models. (For a description of such models and how they are developed, see José-Víctor Ríos-Rull, “Models with Heterogeneous Agents, ” in Thomas F. Cooley, ed., Frontiers of Business Cycle Research (Princeton: Princeton University Press, 1995), pp. 98–125.)

   Some researchers have used heterogeneous-agent models to study the costs of business cycles, in terms of the reduced well-being of the agents. In recessions, people who do not lose their jobs are not affected as much as people who lose their jobs; heterogeneous-agent models can account for the differential impact on the well-being of different people. In addition, people who lose their jobs may not be able to borrow, so their consumption spending declines, making them worse off. Research shows that when people cannot borrow, the costs of business cycles are significantly larger than if people were able to borrow whenever they lose their jobs, and thus not have to reduce their spending.

  Researchers have also used heterogeneous-agent models to see if they can calibrate the real interest rate better than in other models. The real interest rate generated by RBC models is often several percentage points higher than is true in the data. But in RBC models with heterogeneous agents in which people face risk, such as the risk of becoming unemployed, and cannot borrow if they become unemployed, then the real interest rate is somewhat lower than in other RBC models without heterogeneous agents. The risk in such models also leads people to save more than they would if there were no such risk.

  So, RBC models with heterogeneous agents are able to match certain aspects of the economic data better than standard RBC models.

II. Money in the Classical Model (Sec. 10.2)

A) Monetary policy and the economy

Money is neutral in both the short run and the long run in the classical model, because prices adjust rapidly to restore equilibrium

B) Monetary nonneutrality and reverse causation

1. If money is neutral, why does the data show that money is a leading, procyclical variable?

a. Increases in the money supply are often followed by increases in output

b. Reductions in the money supply are often followed by recessions

2. The classical answer: Reverse causation

a. Just because changes in money growth precede changes in output doesn’t mean that the money changes cause the output changes

b. Example: People put storm windows on their houses before winter, but it’s the coming winter that causes the storm windows to go on, the storm windows don’t cause winter

c. Reverse causation means money growth is higher because people expect higher output in the future; the higher money growth doesn’t cause the higher future output

d. If so, money can be procyclical and leading even though money is neutral

Data Application

A review of empirical work testing the RBC theory of reverse causation is Shaghil Ahmed, “Does Money Affect Output?” Federal Reserve Bank of Philadelphia Business Review, July/August 1993. He finds mixed support for reverse causation, but does suggest that money growth is unlikely to be a major factor causing business cycles.

3. Why would higher future output cause people to increase money demand?

a. Firms, anticipating higher sales, would need more money for transactions to pay for materials and workers

b. The Fed would respond to the higher demand for money by increasing money supply; otherwise, the price level would decline

Theoretical Application

The early theoretical RBC models did not include a monetary sector at all—they assumed that money was unimportant for the business cycle. More recently, RBC theorists have been trying to incorporate money into their models. The focus so far has been trying to get the models to produce a liquidity effect, in which in increase in the money supply temporarily reduces nominal interest rates. See Lawrence J. Christiano, “Modeling the Liquidity Effect of a Money Shock,” Federal Reserve Bank of Minneapolis Quarterly Review, Winter 1991.

Analytical Problem 4 works out another example of how reverse causation could occur through firms’ demand for money for transactions and the Fed’s money-supply response.

C) The nonneutrality of money: Additional evidence

1. Friedman and Schwartz have extensively documented that often monetary changes have had an independent origin; they weren’t just a reflection of changes or future changes in economic activity

a. These independent changes in money supply were followed by changes in income and prices

b. The independent origins of money changes include such things as gold discoveries, changes in monetary institutions, and changes in the leadership of the Fed

2. More recently, Romer and Romer documented additional episodes of monetary nonneutrality since 1960

a. One example is the Fed’s tight money policy begun in 1979 that was followed by a minor recession in 1980 and a deeper one in 1981

b. That was followed by monetary expansion in 1982 that led to an economic boom

Theoretical Application

For a thorough overview of how money works to affect the economy in various models, see the symposium on The Monetary Transmission Mechanism, in the Journal of Economic Perspectives, Fall 1995.

3. So money does not appear to be neutral

4. There is a version of the classical model in which money isn’t neutral—the misperceptions theory discussed next

Numerical Problems 2 and 3 examine price level effects in the classical model.

III.  The Misperceptions Theory and the Nonneutrality of Money (Sec. 10.3)

A) Introduction to the misperceptions theory

1. In the classical model, money is neutral since prices adjust quickly

a. In this case, the only relevant supply curve is the long-run aggregate supply curve

b. So movements in aggregate demand have no effect on output

2. But if producers misperceive the aggregate price level, then the relevant aggregate supply curve in the short run isn’t vertical

a. This happens because producers have imperfect information about the general price level

b. As a result, they misinterpret changes in the general price level as changes in relative prices

c. This leads to a short-run aggregate supply curve that isn’t vertical

d. But prices still adjust rapidly

B) The misperceptions theory is that the aggregate quantity of output supplied rises above the full-employment level when the aggregate price level P is higher than expected

1. This makes the AS curve slope upward

2. Example: A bakery that makes bread

a. The price of bread is the baker’s nominal wage; the price of bread relative to the general price level is the baker’s real wage

b. If the relative price of bread rises, the baker may work more and produce more bread

c. If the baker can’t observe the general price level as easily as the price of bread, he or she must estimate the relative price of bread

d. If the price of bread rises 5% and the baker thinks inflation is 5%, there’s no change in the relative price of bread, so there’s no change in the baker’s labor supply

e. But suppose the baker expects the general price level to rise by 5%, but sees the price of bread rising by 8%; then the baker will work more in response to the wage increase

3. Generalizing this example, if everyone expects prices to increase 5% but they actually increase 8%, they’ll work more

4. So an increase in the price level that is higher than expected induces people to work more and thus increases the economy’s output

5. Similarly, an increase in the price level that is lower than expected reduces output

6. The equation Y = Y+ b(PPe) [Eq. (10.4)] summarizes the misperceptions theory

7. In the short run, the aggregate supply (SRAS) curve slopes upward and intersects the long-run aggregate supply (LRAS) curve at P = Pe (Figure 10.2; like text Figure 10.6)

Figure 10.2

Analytical Problem 1 connects the effects of a change in the future marginal product of capital in an RBC model to that in a misperceptions model.

C) Monetary policy and the misperceptions theory

1. Because of misperceptions, unanticipated monetary policy has real effects; but anticipated monetary policy has no real effects because there are no misperceptions

2. Unanticipated changes in the money supply (Figure 10.3; like text Figure 10.7)

Figure 10.3

a. Initial equilibrium where AD1 intersects SRAS1 and LRAS

b. Unanticipated increase in money supply shifts AD curve to AD2

c. The price level rises to P2 and output rises above so money isn’t neutral

d. As people get information about the true price level, their expectations change, and the SRAS curve shifts left to SRAS2, with output returning to

e. So unanticipated money isn’t neutral in the short run, but it is neutral in the long run

3. Anticipated changes in the money supply

a. If people anticipate the change in the money supply and thus in the price level, they aren’t fooled, there are no misperceptions, and the SRAS curve shifts immediately to its higher level

b. So anticipated money is neutral in both the short run and the long run

Data Application

Does the data support the misperceptions theory? Robert Barro, Unanticipated Money, Output, and the Price Level in the United States, Journal of Political Economy, August 1978, pp. 549–580, found support for the misperceptions theory; his results suggested that output was affected only by unanticipated money growth. But others challenged these results and found that both anticipated and unanticipated money growth seem to affect output. See Frederic S. Mishkin, Does Anticipated Monetary Policy Matter? An Econometric Investigation, Journal of Political Economy, February 1982, pp. 22–51.

D)      Rational expectations and the role of monetary policy

1. The only way the Fed can use monetary policy to affect output is to surprise people

2. But people realize that the Fed would want to increase the money supply in recessions and decrease it in booms, so they won’t be fooled

3. The rational expectations hypothesis suggests that the public’s forecasts of economic variables are well-reasoned and use all the available data

4. If the public has rational expectations, the Fed won’t be able to surprise people in response to the business cycle; only random monetary policy has any effects

5. So even if smoothing the business cycle were desirable, the combination of misperceptions theory and rational expectations suggests that the Fed can’t systematically use monetary policy to stabilize the economy

Numerical Problems 4 and 8 look at the misperceptions theory and unanticipated compared to anticipated changes in the money supply.

6. Propagating the effects of unanticipated changes in the money supply

a. It doesn’t seem like people could be fooled for long, since money supply figures are reported weekly and inflation is reported monthly

b. Classical economists argue that propagation mechanisms allow short-lived shocks to have long-lived effects

c. Example of propagation: The behavior of inventories

(1) Firms hold a normal level of inventories against their normal level of sales

(2) An unanticipated increase in the money supply increases sales

(3) Since the firm can’t produce many more goods immediately, it draws down its inventories

(4) Even after the money supply change is known, the firm must produce more to restore its inventory level

(5) Thus the short-term monetary shock has a long-lived effect on the economy

Theoretical Application

Though the text presents the theories in the reverse order, the misperceptions theory came first (being developed in the 1970(s) and the RBC theory came later (in the 1980s). Many classical economists moved away from the misperceptions theory because they weren’t convinced by its arguments for monetary non-neutrality; in particular, the information lag in observing money and prices didn’t seem long enough to cause much effect. For a broad review of how classical macroeconomic theory developed, as well as its relationship to keynesian theory, see Robert J. Barro, New Classicals and Keynesians, or the Good Guys and the Bad Guys, National Bureau of Economic Research Working Paper No. 2982, May 1989. Also, Bennett McCallum discusses the development of the classical approach in New Classical Macroeconomics: A Sympathetic Account, Scandinavian Journal of Economics, 1989, pp. 223–252.

E) Box 10.1: Are price forecasts rational?

1. Economists can test whether price forecasts are rational by looking at surveys of people’s expectations

2. The forecast error of a forecast is the difference between the actual value of the variable and the forecast value

3. If people have rational expectations, forecast errors should be unpredictable random numbers; otherwise, people would be making systematic errors and thus not have rational expectations

Data Application

If you examine a survey of forecasters, like the Livingston Survey, you’ll see that the forecasters made very bad forecasts of inflation around 1973 to 1974 and again around 1979 to 1980. Both time periods are associated with large rises in oil prices.

    Looking at data on interest rates, if you take nominal interest rates and subtract the expected inflation rate (using the Livingston Survey forecasts of inflation), the resulting real interest rates are nearly always positive. But if you subtract actual inflation rates from nominal interest rates, you’ll find negative realized real interest rates around the time of the oil price shocks. In fact, the real interest rate was as low as negative 5 percent at one point. So making bad inflation forecasts has expensive consequences in financial markets.

4. Many statistical studies suggest that people don’t have rational expectations

5. But people who answer surveys may not have a lot at stake in making forecasts, so couldn’t be expected to produce rational forecasts

6. Instead, professional forecasters are more likely to produce rational forecasts

7. Keane and Runkle, using a survey of professional forecasters, find evidence that these forecasters do have rational expectations

8. Croushore used inflation forecasts made by the general public, as well as economists, and found evidence broadly consistent with rational expectations, though expectations tend to lag reality when inflation changes sharply

Data Application

The survey used by Keane and Runkle was begun by Victor Zarnowitz of the University of Chicago in 1968 and was run by the American Statistical Association and National Bureau of Economic Research until 1990. At that time the survey was taken over by the Federal Reserve Bank of Philadelphia and christened the “Survey of Professional Forecasters,” See the article by Dean Croushore, “Introducing: The Survey of Professional Forecasters,” Federal Reserve Bank of Philadelphia Business Review, November/December 1993.