As policymakers and advisers thrash out policy options, the debate about the role of government is undercut by those who quickly jump to extreme cases–by those who argue that a policy that goes one more step in a certain direction will put society on a slippery slope to perdition. Granted, countries have engaged in death spirals of government involvement. Poor government intervention causes its own problems, which inspire calls for more intervention, and so on. Often only a crisis, such as a deep recession or manifestly poor performance over a long period compared to similar countries, can generate the political will to break the poor policy cycle.

Although an awareness of the dangers of such poor policies, including an awareness of the political difficulties of undoing policies that create their own political constituencies, is extremely important, it does little good to brand every proposal to increase the role of government as a step toward the slippery slope. Some are, but some are not.

An example of such a reaction is George Will’s op-ed in the Washington Post. here.  A few days before Will’s piece, Elizabeth Warren said that rich people do not achieve their wealth solely on their own merit, they have relied on society for security, for education for themselves and for the people they hire, for infrastructure, etc. Because the wealthy got so much more from society than the poor, she argued, effective federal income tax rates on high-income households should be increased. George Will agrees with the initial position, that the attainments of individuals are heavily influenced by society, but then he goes on to say that this does not “…entail a collectivist political agenda…” an agenda that is based on the idea that “…any individual’s achievements should be considered entirely derivative from society…” and that “…society is entitled to socialize–i.e. conscript–whatever portion it considers its share.”

Did Warren say that? She basically used one of the traditional “fairness” arguments for proportionately higher taxes on high-income families, and she was implying that the current federal income tax structure, which, by itself, levies extremely disproportionate taxes on high-incomes, is not disproportionate enough. This argument can be fleshed out by noting that although roughly the lowest 50 percent of households pay no federal income tax, they pay, compared to high-income households, a disproportionately large share of their income in payroll taxes.

Good arguments can be made against Warren’s proposal. One could point out that an increase in marginal tax rates on income distorts economic decisions and cause resources to be diverted away from their best uses (by encouraging even more people to spend time and money to avoid taxes instead of producing something, or discouraging work effort generally, or causing investment in plant and equipment to be influenced by tax law instead of estimates of the intrinsic merit). This would be a legitimate criticism of Warren’s proposal. Even under current law, a high-income household that lives in a jurisdiction that has high local marginal tax rates, such as New York, may have a marginal income tax rate close to 50 percent. Is that too high? Will that cause costly distortions?

But Will does not hone the debate by bringing up such issues. Instead, he uses Warren’s statement to remind readers that her argument, if carried to extremes, can be destructive. He brings in arguments not directly related to the proposal, but which instead are directed against a caricature of a socialist, which, I guess, he thinks she must be. He argues that people who use the fairness argument to raise taxes on the rich exemplify those who distrust the individual’s ability or willingness to do what is best for themselves, and that people use this argument to justify massive government involvement in the lives of its citizens. Will certainly agrees that there is a proper role for government, but he feels it is necessary to raise a warning that Warren’s argument can lead us down a slippery slope.

But do such warnings help? Do they help us determine which policies are good, or how policies should be designed? Will says that government keeps society strong by facilitating voluntary “…cooperation with roads, schools, police, etc.–and by getting out of the way. This is a sensible, dynamic, prosperous society’s ‘underlying social contract’”.

Sounds good, but what is in the “etc.?” How should government provide roads and schools? How should it raise revenue? Productive discussions of the role of governments focus on those specifics, on analyses that provide anchors for the design of good policies, not in repeated scaremongering of slippery slopes.

Levels of real per capita GDP are often cited to indicate living standards are higher in one country than another, and, by implication, that some systems of government or sets of government policies are better than others, but this measure of comparative living standards is so flawed as to be practically useless. It is commonly invoked when comparing economic systems. For example, a common reply to those who advocate more government involvement in the U.S. economy is to compare real GDP per capita in western European countries to that of the United States. In a comment to Paul Krugman’s argument that the social democracies of Western Europe may show that “…social justice and progress can go hand in hand”, Greg Mankiw said readers of Krugman’s column “…might find these figures useful to keep in mind.” here

Real GDP per Capita in US $
United States: 47,440
United Kingdom: 36,359
Germany: 35,539
France: 34,205
Italy: 30,631
Spain: 30,589

Mankiw’s implication, of course, is that European countries, which have high levels of government involvement in the economy, have lower living standards—France is 25 percent lower than the United States!—and therefore one should be careful about advocating more government involvement in the economy. However, such comparisons of levels of real GDP are extremely misleading—they muddy the waters rather than contribute something useful to the debate.

Such comparisons are poor for many reasons, but to illustrate the magnitude of the problem, it is sufficient to discuss only six:
1. GDP does not include every material good or service that contributes to living standards, particularly household production and much of the underground economy.
2. A country that actively chooses to consume more leisure than another will fair poorly in such a comparison, but the consumption of leisure definitely contributes to living standards.
3. GDP measures health services by their cost, but there are well-documented differences between health spending and health outcomes across countries.
4. Inter-country comparisons of GDP require price level adjustments, and the standard Purchasing Power Parity methods, though better than nominal exchange rates, may be highly inaccurate indicators of relative costs-of-living.
5. Individuals’ sense of their standard of living is affected by income distribution. In general, people prefer to live in a country in which poverty is low or non-existent, but GDP measures do not reflect income inequalities.
6. Government involvement in an economy can dramatically affect individuals’ economic freedom and opportunities, both for better or worse.

Household production and other economic activities that contribute to living standards are not counted in GDP. One set of estimates indicates household output was 50 percent of reported GDP in 1946, but only 36 percent in 1997. See here. Household production fell relative to the market-oriented production that is measured by GDP, so the nation’s true output grew less rapidly than GDP indicates over that time. Households now eat out more, purchase more prepared foods at the grocery store, hire out child-care services they once performed themselves, and so on, so more of the production that once took place in households now shows up in GDP.

The absence of measures of household production in GDP accounts clearly affects the usefulness of per capita GDP comparisons across countries that have vastly different shares of household production. Also, higher taxes (or burdensome regulations) on transactions encourage do-it-yourself and underground or ‘off-the-books’ activity. This tends to lower measured GDP (although GDP accounts may capture some underground activity) relative to true production. The magnitude of these effects is unknown, but a rough order of magnitude of at least the household production effect can be gleaned from the fact that about 10 percent more women ages 15-64 were employed in the U.S. in 2000 than in Germany or France. Household production is probably significantly lower relative to GDP in the United States than in those countries.

Choice between work and leisure. European countries in general have made a social choice favoring more leisure than has the United States. Leisure is a good that is consumed, but it is not measured in the GDP accounts. Therefore, a society that opts to have more leisure will tend to produce less GDP per capita, but it will not necessarily be worse off. It is impossible to make a reliable estimate of the relative value of the leisure choice across countries, but relative hours data provides a crude estimate of the magnitude of this problem. Hours worked per year in the larger west European countries are much lower than in the United States. This alone could account for much of the differences between GDP per capita. See chart 4 here, which indicates French income per hour worked is the same as in the United States.

Expenditures on health care vs. health care outcomes. Health care expenses, which are about 18 percent of GDP in the United States, are a poor proxy for comparing standards of living across countries because health outcomes are not closely related to health spending. It is difficult to say how much this undercuts the per capita GDP comparison, because there is no easy way to adjust the data for this effect, but it is clear that, compared to the United States, life expectancy and infant death rates are better in many European countries that spend less on health care. Therefore, the linkage between GDP and living standards in this major category of consumption spending is suspect.

Effect of income distribution on living standards. GDP ignores the degree of income equality in an economy, but virtually everyone’s sense of their own standard of living is affected to some degree by concerns about massive economic inequality (or, at least, by how many people are living in dire poverty). This is clear from the extent of charitable giving and from voters’ support of government programs that attempt to ameliorate poverty. In addition, a survey that asked people to give their views of ideal income patterns indicated a desire for a great deal of income equality. See here.

Therefore, when ranking the standard of living among countries, one should try to put a quantitative value on the value people receive from more ‘income equitable’ societies, or societies that provide a relatively sturdy safety net for those who are poor. Of course, there is no way to derive such a valuation, which is one of the reasons it is not included in GDP—those who compile GDP estimates strive to make them as objective as possible. However, European countries tend to have greater income equality than the United States.

The accuracy of Purchasing Power Parity (PPP) estimates is suspect. In order to compare GDP across countries with different currencies, economists traditionally convert currencies to an ‘international U.S. dollar’ standard by estimating what the price level of other countries would be if they used the U.S. dollar. This could accurately reflect relative standards of living if relative prices and tastes were similar and if the statistical agencies use similar methods for estimating prices, but those assumptions clearly do not hold, even for comparing European countries with the United States. Health care prices alone would severely undermine the assumption. Although PPP estimates are better than using market exchange rates to convert GDPs to the same currency unit, it is simply not known how accurate they are.

Economic Freedom and Economic Opportunity. Many of those who feel European economies are inferior do not argue on the basis of material output per capita or per hour worked, but instead argue that the large role of the government in European economies severely undercuts human freedoms–it is almost immaterial whether or not output is greater. This is not an argument that can be empirically tested to any convincing degree—it appeals more to deep-seated feelings, feelings that can vary radically from individual to individual within every country. This is not to say it should not be discussed, however. Although it generally appears that the United States would score higher on a freedom scale, one issue that is rarely, if ever, discussed is the degree to which poverty or the lack of support for children who are severely disadvantaged by malnutrition or emotional disorders induced by bad parenting may undercut many individuals’ freedom in the United States. How free is an 18-year old whose childhood did not prepare him to be a functioning adult?

There are many more problems with international comparisons of living standards (see here for an in-depth report), but the problems posed by these six issues show that discussions of the role of government are not illuminated by offhand appeals to per capita GDP rankings.