16.4 Milton Friedman and the Permanent-Income Hypothesis
In a book published in 1957, Milton Friedman proposed the permanent-income hypothesis to explain consumer behavior. Friedman’s permanent-income hypothesis complements Modigliani’s life-cycle hypothesis: both use Irving Fisher’s theory of the consumer to argue that consumption should not depend on current income alone. But unlike the life-cycle hypothesis, which emphasizes that income follows a regular pattern over a person’s lifetime, the permanent-income hypothesis emphasizes that people experience random and temporary changes in their incomes from year to year.3
494
The Hypothesis
Friedman suggested that we view current income Y as the sum of two components, permanent income YP and transitory income YT. That is,
Y = YP + YT.
Permanent income is the part of income that people expect to persist into the future. Transitory income is the part of income that people do not expect to persist. Put differently, permanent income is average income, and transitory income is the random deviation from that average.
To see how we might separate income into these two parts, consider these examples:
Maria, who has a law degree, earned more this year than John, who is a high school dropout. Maria’s higher income resulted from higher permanent income because her education will continue to provide her a higher salary.
Sue, a Florida orange grower, earned less than usual this year because a freeze destroyed her crop. Bill, a California orange grower, earned more than usual because the freeze in Florida drove up the price of oranges. Bill’s higher income resulted from higher transitory income because he is no more likely than Sue to have good weather next year.
These examples show that different forms of income have different degrees of persistence. A good education provides a permanently higher income, whereas good weather provides only transitorily higher income. Although one can imagine intermediate cases, it is useful to keep things simple by supposing that there are only two kinds of income: permanent and transitory.
Friedman reasoned that consumption should depend primarily on permanent income because consumers use saving and borrowing to smooth consumption in response to transitory changes in income. For example, if a person received a permanent raise of $10,000 per year, his consumption would rise by about as much. Yet if a person won $10,000 in a lottery, he would not consume it all in one year. Instead, he would spread the extra consumption over the rest of his life. If we assume an interest rate of zero and a remaining life span of 50 years, consumption would rise by only $200 per year in response to the $10,000 prize. Thus, consumers spend their permanent income, but they save rather than spend most of their transitory income.
495
Friedman concluded that we should view the consumption function as approximately
C = αYP,
where α is a constant that measures the fraction of permanent income consumed. The permanent-income hypothesis, as expressed by this equation, states that consumption is proportional to permanent income.
Implications
The permanent-income hypothesis solves the consumption puzzle by suggesting that the standard Keynesian consumption function uses the wrong variable. According to the permanent-income hypothesis, consumption depends on permanent income YP; yet many studies of the consumption function try to relate consumption to current income Y. Friedman argued that this errors-in-variables problem explains the seemingly contradictory findings.
Let’s see what Friedman’s hypothesis implies for the average propensity to consume. Divide both sides of his consumption function by Y to obtain
APC = C/Y = αYP/Y.
According to the permanent-income hypothesis, the average propensity to consume depends on the ratio of permanent income to current income. When current income temporarily rises above permanent income, the average propensity to consume temporarily falls; when current income temporarily falls below permanent income, the average propensity to consume temporarily rises.
Now consider the studies of household data. Friedman reasoned that these data reflect a combination of permanent and transitory income. Households with high permanent income have proportionately higher consumption. If all variation in current income came from the permanent component, the average propensity to consume would be the same in all households. But some of the variation in income comes from the transitory component, and households with high transitory income do not have higher consumption. Therefore, researchers find that high-income households have, on average, lower average propensities to consume.
Similarly, consider the studies of time-series data. Friedman reasoned that year-to-year fluctuations in income are dominated by transitory income. Therefore, years of high income should be years of low average propensities to consume. But over long periods of time—say, from decade to decade—the variation in income comes from the permanent component. Hence, in long time-series, one should observe a constant average propensity to consume, as in fact Kuznets found.
496
CASE STUDY
The 1964 Tax Cut and the 1968 Tax Surcharge
The permanent-income hypothesis can help us interpret how the economy responds to changes in fiscal policy. According to the IS–LM model of Chapter 11 and Chapter 12, tax cuts stimulate consumption and raise aggregate demand, and tax increases depress consumption and reduce aggregate demand. The permanent-income hypothesis, however, predicts that consumption responds only to changes in permanent income. Therefore, transitory changes in taxes should have only a negligible effect on consumption and aggregate demand.
That’s the theory. But one might naturally ask: is this prediction actually borne out in the data?
Some economists say yes, and they point to two historical changes in fiscal policy—the tax cut of 1964 and the tax surcharge of 1968—to illustrate the principle. The tax cut of 1964 was popular. It was announced as being a major and permanent reduction in tax rates. As we discussed in Chapter 11, this policy change had the intended effect of stimulating the economy.
The tax surcharge of 1968 arose in a very different political climate. It became law because the economic advisers of President Lyndon Johnson believed that the increase in government spending from the Vietnam War had excessively stimulated aggregate demand. To offset this effect, they recommended a tax increase. But Johnson, aware that the war was already unpopular, feared the political repercussions of higher taxes. He finally agreed to a temporary tax surcharge—in essence, a one-year increase in taxes. The tax surcharge did not seem to have the desired effect of reducing aggregate demand. Unemployment continued to fall, and inflation continued to rise. This is what the permanent-income hypothesis would lead us to predict: the tax increase affected only transitory income, so consumption behavior and aggregate demand were not greatly affected.
While these two historical examples are consistent with the permanent-income hypothesis, it is hard to draw firm inferences from them. At any moment in time, there are many macroeconomic influences on consumer spending, including the overall confidence that consumers have in their own economic prospects. It is hard to disentangle the effects of tax policy from the effects of other events occurring at the same time. Fortunately, some recent research has reached more reliable conclusions, as discussed in the next Case Study. 
CASE STUDY
The Tax Rebates of 2008
When medical researchers want to know the effectiveness of a new treatment, the best approach is a randomized controlled experiment. A group of patients is assembled. Half of them are given the new treatment, and the other half are given a placebo. The researchers can then track and compare the two groups to measure the effects of the treatment.
Macroeconomists usually cannot conduct randomized experiments, but sometimes such experiments fall into our lap as accidents of history. One example occurred in 2008. As a result of a severe financial crisis that year, the economy was heading into a recession. To counteract the recessionary forces, Congress passed the Economic Stimulus Act, which provided $100 billion of one-time tax rebates to households. Single individuals received $300 to $600, couples received $600 to $1,200, and families with children received an additional $300 per child. Most important, because sending out many millions of checks was a long process, consumers received their tax rebates at different times. The timing of receipt was based on the last two digits of the individual’s Social Security number, which is essentially random. By comparing the spending behavior of consumers who received early payments to the behavior of those who received later payments, researchers could use this random variation to estimate the effect of a transitory tax cut.
497
Here are the results, as reported by the researchers who did the study: “We find that on average households spent about 12 to 30 percent of their stimulus payments, depending on the specification, on nondurable consumption goods and services (as defined in the consumer expenditure survey) during the three-month period in which the payments were received. This response is statistically and economically significant. We also find a significant effect on the purchase of durable goods and related services, primarily the purchase of vehicles, bringing the average response of total consumption expenditures to about 50 to 90 percent of the payments during the three-month period of receipt.”4
The findings of this study stand in stark contrast to what the permanent-income hypothesis predicts. If households were smoothing their consumption over time, as the permanent-income hypothesis assumes, they would have spent only a small fraction of the tax rebate in a three-month period, but the data show a large impact of the rebate on spending. In addition, if the permanent-income hypothesis were correct, those receiving the early checks would not have behaved any differently from those receiving the later checks because the permanent income of the two groups was the same. Yet the data show that the timing of the check’s arrival had a profound impact on the timing of a household’s spending.
The permanent-income theory may be correct in positing that permanent tax changes influence consumer spending more powerfully than transitory ones. But from the evidence from the 2008 experience, it seems wrong to conclude that the effects of transitory tax changes are insignificantly small. Even very transitory changes in tax policy can influence how much consumers spend.