Living Standards Rise Faster than Official Estimates

10/01/1999
Summary of working paper 7148
Featured in print Digest

Between 1890 and 1940 the average work week fell by 20 hours and retirement rates of men older than 64 rose by almost 30 percentage points.

In 1972-91, living standards were rising twice as fast as conventionally measured... the primary beneficiaries of these newly measured gains were lower income households.

Measuring living standards is difficult for economists. The usual method is to look at the evolution over time of per capita real income ( income adjusted for inflation.) But "real income is an imperfect measure of trends in living standards," cautions NBER Faculty Research Fellow Dora Costa. One reason is that the Consumer Price Index (CPI) that is used to adjust for inflation overstates the rise in prices. More importantly, "real income does not account for such goods as health that are not purchased in the marketplace, for quality changes, for revolutionary technological change, and for increases in leisure," Costa maintains.

An alternative technique for measuring living standards is to examine spending on recreation. This technique assumes that, after providing for food, clothing, shelter, and other necessities of life, people will use some of whatever income is left over on purchasing recreation. As the proportion of income needed to meet necessities declines, people will have more money for vacations, radios, TVs, CDs, and other recreational goods and activities that strike their fancy. Rising recreational expenditures -- "the quintessential luxury goods" -- are seen as an indication of rising living standards.

In American Living Standards: Evidence From Recreational Expenditures (NBER Working Paper No. 7148), Costa uses consumer expenditure survey data from 1888-90, 1917-19, 1935-6, 1972-3, and 1991 to determine whether trends in real income per capita are consistent with trends in recreational budget shares. She also looks at trends in inequality in recreational spending for those up and down the income ladder. She finds that changes in real total expenditures per capita are likely to underestimate the increase in living standards, particularly during times of innovation in consumer goods and reductions in working hours, such as the 1920s, the 1970s, and the1980s. In the late 1880s, less than 2 percent of household spending was devoted to recreation; by the mid-1930s, recreation's share had risen to 4 percent, and by 1991 to 6 percent.

Recreational expenditures also tend to rise with income levels. Richer households spend proportionately more on luxuries, including recreation. At the start of this century, pleasure spending was considered a luxury that could not be afforded by low-income households. But by the mid-1930s, low cost recreational activities, such as motoring, movies, and the radio had already spread among the people, especially for those who had jobs.

The government also invested heavily in recreational facilities. The number of public swimming pools more than tripled and the number of baseball diamonds more than doubled between 1921 and 1930 alone. There were more parks. And, listening to the radio was a favorite activity of all classes in the 1930s. Today the popular medium is television.

Hours of work also have declined over the decades. Between 1890 and 1940 the average work week fell by 20 hours and retirement rates of men older than 64 rose by almost 30 percentage points. This decline in hours, the spending on recreation suggests, was especially the case for lower income workers. They worked 11-hour days in the 1890s, while the well-to-do worked nine hours. After 1940, paid vacations, holidays, sick days, and personal leave increased; retirement rates continued to rise. But the length of the average work week remained unchanged.

Costa notes that real total expenditures per capita fell by 1.2 percent per year between 1919 and 1935 and rose by 1.8 percent per year between 1972 and 1991. In contrast, trends in the recreational expenditure share of total spending imply that between 1919 and 1935, real per capita total expenditures actually rose by 1.2 percent per year, and between 1972 and 1991 by 3.6 percent per year. So the conventional measures of real per capita total expenditures -- used as an indicator of living standards -- may be biased downward by 2.4 percentage points per year between 1919 and 1935 and 1.8 percentage points per year between 1972 and 1991. Even in the period that includes the Great Depression, living standards rose by this measure, at least for those with jobs.

"It may therefore be time to reassess the 1920s and 1930s," writes Costa. "Changes in the consumption bundle of households and declines in inequality in recreation suggest that improvements in living standards may have been much more rapid, particularly among lower income households, than suggested by income measures alone. These gains may have been so large that even the income declines of Great Depression were not enough to reverse them."

Further, in 1972-91, living standards were rising twice as fast as conventionally measured. Costa also finds that the primary beneficiaries of these newly measured gains in per capita total expenditures in both these periods were lower income households.

-- David R. Francis