Limitations of GDP Statistics

All macroeconomic statistics have limitations. What are some of the limitations of the GDP numbers?

With a few exceptions, GDP only counts goods and services that pass through markets. Production that is not bought or sold does not generally get counted. When one looks at the per capita GDP (total GDP divided by population) of various countries, one finds some countries that have per capita GDPs of only one or two hundred dollars. This statistic may mean that the inhabitants of such countries are truly impoverished, or it may mean that most transactions are not market transactions. If a nomadic or agricultural society is self-sufficient within small groups, there may be considerable production unrecorded in the GDP statistics because none of it reaches the market. This omission also means that a rapid rise in GDP may not reflect more production, but a shift in production from a nonmarketed form to a marketed form.

Although GDP is often used to measure how well off people are in a material sense, it has serious deficiencies as a measure of economic welfare. In the late 1960s several economists decided to adjust GDP (actually GNP, which was the measure then) to obtain a better measure of economic welfare, which they called Measure of Economic Welfare, or MEW. They noted that leisure was valuable, but this value was not recorded in GDP. If people work 60-hour weeks rather than 40-hour weeks, GDP will be higher, but people may be worse off. They also noted that many "bads" are included in GDP--higher crime rates lead to more expenditures on police, international tensions lead to more expenditures on arms, more disease leads to more medical spending, etc.--and thus GDP overstates welfare. Further, the production of some goods also brings with it some bads--pollution is an example.

Though these adjustments are sensible, putting them into numbers involves serious problems. When MEW was computed for the Great Depression, the large drop in production was significantly offset by a large increase in leisure and a drop in pollution. Hence MEW suggested that conditions in the 1930s were not really so bad after all. As a result, measures such as MEW have not replaced GDP as a measure of how severe recessions and how strong recoveries are, even though GDP has deficiencies as a measure of welfare.¹

Another problem with GDP that has fascinated economists involves the size of the "underground economy," the economic activity unreported to the government because those engaged in it are attempting to avoid taxes. When a waitress takes tips that she does not report, or a plumber offers to work for less if paid in cash rather than by check, or a farmer who sells vegetables at a roadside stand understates his revenues to the IRS, each is part of the underground economy.

There is no way to accurately know how important this activity is because it is unreported. All estimates of its size must be done by indirect ways. Some economists have looked at the large amount of currency in circulation and concluded that only the existence of large amounts of unreported transactions can explain why in 1984 the public held $150 billion in cash.² Other economists believe that the rising percentage of the population that is employed in the visible economy suggests that the underground economy has not been increasing in size and may not be very large. The dispute is a reminder that there can be measurement error of unknown size in economic statistics.

The problem of measurement error in deriving economic statistics was highlighted in a Wall Street Journal article of November 22, 1983. The article reported that the Federal Reserve had estimated personal savings in the second quarter of 1983 at an annual rate of $209.3 billion, and the Commerce Department had estimated personal savings for the same period at an annual rate of only $92.3 billion. Thus, if one wanted to compute the rate at which Americans were saving, one had two very different figures from which to choose and could get two very different savings rates.

The Federal Reserve computed its savings statistic in the process of computing the sources and uses of funds raised in U.S. financial markets. Savings is found by adding the increase in financial assets households hold to investment such as new homes, and by subtracting the increase in personal debt. The Commerce Department computed its savings as part of the national income accounts. Savings is the difference between disposable (after-tax) income and consumption expenditures.

There are two ways to explain the huge discrepancy between these two estimates. One possibility is that the Commerce Department overlooked the "underground" economy, and thus income was understated. If income was understated but consumption was not, savings would be underestimated. Another possibility is that the Federal Reserve was not identifying foreign investment in the United States, but was including some of it as private savings, and thus overstating savings. Regardless of the cause, the discrepancy illustrates that there can be very large measurement errors in our economic statistics.

We finish this set of readings with a quick look at Leading Indicators.

¹Attempts to develop measures of welfare seem to have declined, replaced by happiness research. There are many articles and books that try to explain results of surveys asking people of how happy they are. Here are two examples from two economics blogs of the sorts of things discussed in this literature:
http://freakonomics.blogs.nytimes.com/2008/04/16/the-economics-of-happiness-part-1-reassessing-the-easterlin-paradox/
http://www.informationpile.com/econlog/brooks-hidden-secrets-of-happiness-by-bryan-caplan/

²Checks leave trails behind because images of all checks are recorded. The safest way to avoid detection of transactions that one does not want the government to know about is to use cash.
Copyright Robert Schenk