Library of Economics and Liberty masthead logo 
David R. Henderson

GDP Fetishism

David R. Henderson*

"By focusing on GDP, proponents of the program have forgotten that GDP is not the same as wellbeing."
When economics professors teach the basics of Gross Domestic Product (GDP), we usually caution our students that it is not a good measure of welfare. Unfortunately, many economists go on to give GDP far more credit than it deserves. They tend to consider fiscal and monetary policy positive if these policies increase GDP, but they often fail to ask, let alone answer, whether those same policies increase or reduce welfare. I have a term for giving GDP such a sacred a place in economists' reasoning: GDP fetishism. If we return to some basic principles of economics, we will avoid GDP fetishism, do better economic analysis, and propose better policies.

For more on GDP, see National Income Accounts by Mack Ott in the Concise Encyclopedia of Economics.

To see why GDP is not the same as wellbeing [I use "welfare" and "wellbeing" interchangeably], consider the definition of GDP. One of the most careful definitions is in The Economic Way of Thinking, 10th edition, by Paul Heyne, Peter Boettke, and David Prychitko. They write: "The gross domestic product is the market value of all the final goods produced in the entire country in the course of a year."1 Most economists would agree with this definition. It turns out, though, as Heyne et al. point out, that even this careful definition does not accurately characterize GDP, let alone wellbeing. It is inaccurate in two ways. First, because there is usually no market for the things that government produces (the U.S. Postal Service being one of the exceptions), government spending on goods and services is valued at cost rather than at market prices. Second, because many goods and services are not bought or sold, even though they would have a market value if they were, these goods and services are not counted in GDP. In early editions of his best-selling textbook, Economics, the late Paul Samuelson gave his favorite example of this pitfall in GDP accounting. Samuelson pointed out that if a man married his maid, then, all else equal, GDP would fall.

These two inaccuracies in themselves mean that actual GDP is not a good measure of wellbeing. Take the first inaccuracy—the valuing of government-provided goods and services at cost rather than at market prices. Many government programs actually destroy value rather than create it. My "favorite" example is the systematic harassment of airline passengers by the Transportation Security Administration (TSA). The annual budget of the TSA is $6.3 billion.2 If you believe that this $6.3 billion is worth zero, then you would conclude that the TSA's contribution to GDP is not $6.3 billion, but zero. But even zero overstates the TSA's contribution to GDP because the TSA does not create zero value: it creates negative value, by wasting passengers' time in line and causing them to pack in ways they would not otherwise pack. The TSA also destroys value by causing passengers on the margin to shift to less-safe modes of travel. If you doubt that the TSA destroys value, then ask yourself how much you would pay to avoid the lines and the loss of privacy. I bet it is more than zero.

Of course, I rigged the question. You might hate to stand in line and hate the loss in privacy but still favor the TSA because you want them to hassle others who you fear will try to hijack or destroy an airplane with all on board. So then the question to ask yourself is how much you would pay for the TSA. If we could trust you and everyone else to come up with a right answer, then we could add up the positive amounts you and others would pay to have the TSA—call this X—and add up the amounts others and I would pay to eliminate the TSA—call this Y—and then subtract Y from X. I'm fairly confident that X-Y is a negative number. But I could be wrong.

This belief that the TSA could be making us safer may be legitimate, although, interestingly, the examples we have of terrorists on airplanes being defeated are all examples of victory by fellow passengers. Most people recognize, though, that governments at all levels are quite wasteful. That is, they take resources and use them in inefficient and, often, destructive ways. So, if you find my TSA example unpersuasive, then simply substitute your own.

Consider the other inaccuracy that makes our measure of GDP not quite live up to its definition: the failure to value non-market transactions at market prices. Take Samuelson's example of the man marrying his maid. Samuelson's point is that the new bride continues doing the housework without being paid. But that would not mean that the work suddenly had no market value. So, in this case, GDP actually understates the market value of all final goods and services because this particular service is no longer exchanged on the market. Because many valuable goods and services are not exchanged on the market, this inaccuracy imparts a downward bias to measured GDP as a measure of actual GDP.3

But even if our measure of GDP actually measured what the definition says it does, there would still be a major problem with GDP as a measure of wellbeing. That problem arises because GDP does not take into account the value of leisure.

To see why this is a problem, consider what would happen if the productivity of an economy magically doubled, so that each person could produce twice as much per hour of work. Consider two extreme responses to this change, although the likely response would be between the extremes. In extreme case one, everyone works half as much and produces the same amount. Real output remains the same and, if the government's price index is accurate, real GDP will be constant. But suddenly people's leisure has increased. People value leisure. Ask yourself whether you would be better off if you could buy the same goods and services as before by working half as much. Here, real GDP understates wellbeing. In extreme case two, people work the same number of hours but produce twice as much as before. Here, real GDP would double. But in the more likely in-between case, where people work somewhat less but produce more, real GDP would increase; but the increase, by leaving out leisure, would understate the increase in wellbeing.

Economist and Civil War historian Jeffrey Hummel points out4 a real case in which per capita GDP was a poor measure of welfare due to this failure to measure the value of leisure: the emancipation of black slaves in the U.S. south. The total labor supplied by former slaves was only two thirds as much as the labor they had "supplied" as slaves. It hadn't made sense for slave owners to take account of the value of slaves' leisure any more than it made sense for slave owners to take account of the value of horses' leisure. The slave owners regarded the slaves as assets. But the newly acquired freedom to participate in the labor market and make their own choices led the former slaves to choose to work one third less. The value of their production was lower—and they were better off.

It became clear to me during two nationwide discussions of economic policy over the last two years that focusing on GDP can lead us astray from sound economic reasoning. The first discussion was of the wisdom of Keynesian fiscal policy—having the federal government spend money to add to GDP. Advocates of such policy often argued that there was a substantial fiscal policy "multiplier." What they meant was that an added dollar of spending by the federal government would generate additional dollars of spending in the private sector. Many good economists have argued for this multiplier and many against. I am not enough of a macroeconomist to judge the size of this multiplier. But I'm enough of a microeconomist to see what's wrong with much of the discussion of the multiplier. It completely ignores the limits of GDP as a measure of wellbeing.

Take an extreme case: the government spends $10 billion to pay people to dig holes and then refill them. The result is that the $10 billion produces something worth zero. Assume—to give the advocates of such a program their best shot at making a case—that all of the people employed in digging and refilling holes would otherwise have been unemployed. Then, employing them in the hole business does not cause other goods and services to be foregone. Still, the employees value leisure somewhat and they give up this leisure by working. So there is an opportunity cost. If you doubt this, ask yourself whether you think the employees would have been willing to work for one penny per day. If not, then they value their leisure positively. If they are paid, say, $10 an hour in a government job and would have been willing to do the job for a minimum of, say, $6 an hour, then the cost of employing them is $6 an hour. So there is a gain from the program: the $4 per hour that the diggers receive over and above the wage for which they would have been willing to do the job. Assume, for simplicity, that there are no other costs of the $10 billion program besides the wages. This program, then, spends $10 billion to create 1 billion times $4, or $4 billion, of net value. Although GDP registers $10 billion, the real gain is only $4 billion. The multiplier may add to the gain. But my point is that by focusing on GDP, proponents of the program have forgotten that GDP is not the same as wellbeing. The $10 billion part of GDP due directly to the refilled holes represents a benefit of only $4 billion.

And, of course, we haven't even mentioned the loss in wellbeing due to two interrelated factors. First, the $10 billion must be financed somehow, typically with future taxes or with printing money. Second, whichever of these two methods is chosen, there is a cost beyond the $10 billion that economists call a deadweight loss. In the case of future taxes, the deadweight loss is the loss due to tax-avoidance behavior by taxpayers. In the case of printing money, the deadweight loss is caused by the fact that people who hold money will economize on their use of money to avoid the implicit tax that comes from holding money that loses value.

A more-specific recent program, besides the general program of added government spending, illustrates how economists can forget (or ignore) that GDP and wellbeing are not the same. That was the so-called "Cash for Clunkers" program that the federal government ran during the summer of 2009. Under this program, if people owned a car for at least a year, and if that car got fewer than 18 miles per gallon, they could turn in the car as trade toward a new car and get a payment of up to $4,500 from the dealer. The dealer then destroyed the engine to take it out of circulation and then was reimbursed by the government.

Here's what Harvard economist Martin Feldstein wrote a few months later:

"Cash for clunkers," for instance, was successful in raising auto buying and gave a temporary boost to GDP, since two-thirds of the third-quarter GDP rise was motor-vehicle production. The credit for first-time home buyers also gave a temporary boost to the housing market. But both programs just borrowed demand from the future.5

Do you see what's missing? Feldstein's only criticism of "Cash for Clunkers" is that it "borrowed demand from the future." This is true, but notice what he leaves out: that the purpose of the program was to generate GDP even at the cost of destroying wellbeing.6 The government actually paid money to get valuable goods from the private sector and destroy them. Thus, even for an economist as good as Martin Feldstein, GDP has become the Holy Grail. GDP has replaced wellbeing. That's GDP fetishism.

If, instead of seeking GDP, we ask of each government policy, "What will it cost and how much value will it create?" we will come up with better policies. The concept of GDP, handled carefully, can be useful. But for many people, and even many good economists, GDP has been used to judge wellbeing even when using it that way leads to highly misleading conclusions.


P. 352.


See the TSA bio of David R. Nicholson (accessed on February 23, 2010.)


Assuming this problem is constant over time, though, it does not bias the rate of growth of GDP. I thank Bob Murphy for pointing this out.


See Jeffrey Rogers Hummel, Emancipating Slaves, Enslaving Free Men: A History of the American Civil War, Chicago: Open Court, 1996, p. 322. At


Martin Feldstein, "Missing the Target,"Wall Street Journal, January 20, 2010.


For more on this, see my two EconLog posts, "Cash for Clunky Ideas," August 20, 2009 and "Cash for Clunkers in Action," August 25, 2009.

* David R. Henderson is a research fellow with Stanford University's Hoover Institution and an associate professor of economics at the Graduate School of Business and Public Policy at the Naval Postgraduate School in Monterey, California.

Return to top
Copyright ©2008
Liberty Fund, Inc.
All Rights Reserved