William D. Nordhaus: In a democracy, one might suppose, the decisions of elected politicians are to some extent influenced by public opinion. This is the motivation for the question, which Alan Blinder and Alan Krueger address in this paper, of how public opinion and, specifically, public opinion about economic affairs is determined.
Blinder and Krueger investigate the state of economic knowledge and preferences among the general public by tracking down a random sample and asking them a battery of fact questions (to determine their knowledge), personal questions (to determine their self-interest), and economic policy questions (to see where they stand on the issues of the day). No one should be surprised to learn that the authors have gone about this task in a thoughtful manner and have produced many interesting findings. The survey has been carefully administered, and the response rate is high. My comments will raise a number of questions about their questions, which leave me thinking that survey research is even harder than I originally imagined.
To begin with, one might ask whether the respondents behave like the textbook homo economicus classicus. Somewhere, someone probably believes that economic policy opinions are formed by continuous-time maximization of a consistent preference function maximizing the present value of the utility of consumption using Bayesian updating in light of the constant inflow of data from the Internet, several televisions tuned to the financial channels, and a live feed from the Brookings Panel. But anyone who still entertains that notion will quickly discard it after reading this paper.
What do we learn about the public's economic opinions? In reality, public opinions about the economy in a democracy are the outcome of a very complex process in which people try to sort through conflicting accounts and theories, often provided by unreliable narrators. The news contains much factual reporting (including the Dow Jones average to seven significant digits), but very little analytical reporting. Analyses are often shaded by political motives (if by public figures), by profit motives (if by companies), or by desire for publicity (if by private individuals). Even a good news source such as PBS's Jim Lehrer NewsHour feels it has to present a balanced point of view, where "balanced" means someone is there to defend supply-side economics.
Americans report that, of all professions, they trust nurses most and car salesmen and HMO managers least. (1) Canadians report that they trust pharmacists most and national politicians least (only 9 percent of Canadians trust the latter). (2) Yet people get much of their economic knowledge from national politicians and little from nurses and pharmacists. I suspect that most people view getting economic information somewhat akin to extracting a referral from their HMO. Given the difficulties of knowing whom to trust and the contentiousness of most policy discussions, it would not be surprising if the public is confused at best and ill informed at worst.
How do Blinder and Krueger attack this issue? They administer a battery of nine fact questions to see how their respondents have sorted through the varieties of conflicting opinions. I admit that I found some of the questions vaguely unsatisfying, partly because recall of numerical facts is not identical with sound economic analysis.
In addition, I found myself constantly worrying about the questions. For example, on taxes the question was
"About what percentage of the typical American family's income do you think goes to paying taxes--including all levels of government?"
This is actually a very hard question. Does "typical" mean median or average or what the respondent thinks is a typical nuclear family? Should I be economically sophisticated and include production taxes and business taxes in the numerator, as a good public finance theorist would recommend and a good Chicago economist would predict? (Blinder and Krueger would not.) Should I include imputed rent on my owner-occupied house in the denominator as the Bureau of Economic Analysis does? (Blinder and Krueger do, but I doubt that most Americans would.) What do we learn from an incorrect answer to this question? Apparently, survey respondents shared some of my confusions, for they grossly overestimated the typical share of taxes paid by households.
In addition to the knowledge questions, people were asked a number of public policy questions. For example:
"Are you aware that President Bush has proposed that part of Social Security be replaced by personal investment accounts?"
followed by
"Do you favor or oppose this idea, or are you undecided?"
I would be surprised if most people had much knowledge of this proposal. I, for one, have forgotten the details, and perhaps even President Bush would have trouble explaining it clearly. The term "personal investment accounts" has a nice ring to it, and I personally am in favor of them. In fact, I think I already have a few of them. More of a problem is that the term is embedded in a question that states that personal accounts are a Bush proposal, so the question is probably heard as, "President Bush has made a new proposal called personal investment accounts; how do you feel about President Bush and his proposals?" Under this interpretation, it is not a question about Social Security but about attitudes toward Bush.
Economists are always complaining about questions that are posed without the context of the trade-offs involved. The most careful work in this area comes from environmental economists who have designed contingent valuation studies to learn about the value of pristine Alaskan waters or of preserving whooping cranes. The lesson from these studies is that one must be very careful to specify both the "commodity" and the "price" to have any hope that the responses will be meaningful.
Blinder and Krueger are inconsistent in this respect. For example, when querying about drug benefits, they ask
"Would you favor or oppose adding a prescription drug benefit to Medicare for people who are not in the hospital, bearing in mind that it would have to be paid for somehow?"
It is definitely good to bear in mind that the proposed benefit would have to be paid for. But "somehow" is a little vague. It does not explain the payment method, and so it does not really confront the respondent with a realistic choice. The average respondent might assume that "somehow" means "by someone else," as in former Senator Russell Long's ditty, "Don't tax you, don't tax me, tax that fellow behind the tree."
Even worse in this regard is the question on national health insurance:
"Do you favor or oppose what is called 'universal health insurance coverage,' meaning that the government would make sure that every American is covered by a health insurance policy?"
It is not surprising that a great majority of people favor such a policy when no price tag or tax tag is attached. I favor universal health insurance. I would guess most of the Brookings Panel favor it, whatever it is. But it would be better to tease out preferences with something like:
"The next question is about health insurance. One proposal is to extend the Medicare program to all U.S. residents. Statisticians estimate that this proposal could be financed with a tax of $2,500 per family. Would you favor or oppose such a proposal?"
The questions on fiscal policy are potentially the most illuminating, given the widespread support of the American people for tax cuts. Here Blinder and Krueger begin with the following:
"Do you think taxes in the United States are generally too high, too low, or about right?"
They find that most people think taxes are too high (71 percent) and virtually none (3 percent) think they are too low. I agree with most people. I think my taxes are too high. I also think that airplanes are too crowded, teenagers are too noisy, cell phones are a nuisance, and there are too many SUVs. But these opinions, like the view that taxes are too high, are of little relevance because the relevant budget constraints are nowhere in sight.
I mentioned above that people overestimate the typical tax rate. Is it possible that people overestimate the size of things they dislike? Recall that various surveys have found that people want to reduce U.S. spending on foreign aid, but at the same time they estimate that U.S. spending on foreign aid is 20 percent of the budget (an overestimate by about a factor of twenty). (3) Do people think taxes are too high because they constantly hear that taxes are a menace to the Republic rather than that taxes are the price of a civilized society? People may want lower taxes in return for larger deficits, but would they want lower taxes in return for lower Social Security benefits, fewer police, dirtier streets, and more crowded classrooms?
Blinder and Krueger also ask whether respondents favored the Bush tax cuts. The most important determinants of the answers are political philosophy (liberal versus conservative) and whether the respondent thought taxes were too high or too low to begin with. Self-interest and demographic variables (in the form of income) play a very modest role. Instead, what can roughly be thought of as ideology is the key determinant of attitudes on central political issues such as tax cuts.
A similar pattern of responses is seen on the questions about the importance of the federal budget deficit as a problem and how to solve it. Whether people thought taxes were part of the solution depended chiefly on whether they were conservative or liberal and on whether they thought taxes were too high to begin with. Here I worry a little about causality. Am I conservative because I don't like taxes, or do I dislike taxes because I'm conservative? If the label I attach to myself describes my constellation of views, what do we learn from these regressions? Don't we really need to go a step further and find out why I dislike taxes? All this may suggest that the ideology variables are not useful right-hand-side variables.
There are many fascinating findings here, but I will end with one puzzle. The authors conclude that economic opinions and policy attitudes appear to be largely independent of self-interest. They find, for example, that lower-income people are more likely to think taxes are too high than higher-income people. In other areas, such as Social Security and deficit reduction, the coefficient on self-interest often has the right sign, but its importance is swamped by ideological variables. This does suggest, for one thing, that Democratic politicians like Howard Dean could never persuade the guy with a pickup truck sporting a Confederate-flag decal that repealing the death tax is a bad idea.
The authors' conclusion here is not heretical; it is supported by a substantial array of research in the political science literature, which concludes, for example, that there are "no strong effects of personal financial well-being on the vote." (4) This result is consistent with the view that a substantial fraction of voters favor abolition of the inheritance tax, even though its repeal is likely to benefit only about 1 percent of the population.
Here a paradox lurks. Economists may recall results from the "political business cycle" literature showing that election returns track the business cycle, particularly for presidential elections. My colleague Ray Fair did better than most polling firms (or the Supreme Court) in calling the 2000 election, largely on the basis of economic variables. (5) Moreover, aggregate studies over the last three decades, going back to seminal work by Gerald Kramer, (6) as well as studies in virtually every advanced democracy, make it clear that voters punish incumbents who are running for elections when unemployment is rising, real incomes are falling, and inflation is high. How is it that voters punish incumbents who bring on bad times but do not vote on the basis of their self-interest?
General discussion: Several participants discussed the authors' finding that self-interest was not a key factor shaping views on economic issues. Benjamin Friedman applauded the paper for its new evidence about how people form their economic opinions. But he questioned whether attitudes across income groups about tax burdens and tax cuts should be interpreted as indifference to self-interest. He noted that the actual burden of taxes differed from that implied by statutory income tax rates. Low- and middle-income people had to pay their income and payroll taxes out of their paychecks, whereas the well-to-do had many avenues for avoiding high tax rates. Furthermore, assuming declining marginal utility of income, a typical tax cut package might be expected to produce more utility for lower-income people. Shang-Jin Wei argued that if self-interest helped to predict ideology, the authors could be misattributing the effect of self-interest to ideology. Olivier Blanchard noted that both ideology and self-interest have been found by others to be related to voting behavior, and that all observers grant the importance of self-interest in voting on issues such as property taxes. Martin Baily suggested that people may not be truthful in responding to questionnaires, and he cited wide differences between how people say they voted in union elections and the actual vote totals.
A number of panelists discussed how other questions that the survey did not ask might have yielded illuminating results. Robert Gordon argued for questions that related an economic fact to a political interpretation; for example, did respondents attribute the current economic rebound to George W. Bush's tax cuts or to Alan Greenspan's low interest rates? William Dickens compared the present paper with a recent paper by Bryan Caplan in the Economic Journal. Caplan found broad agreement among economists on what the "right" answers were, and that more-knowledgeable people got more answers right (that is, they agreed with the consensus of economists). Dickens conjectured that economists would agree less about the questions in the authors' survey, which might explain the low importance of knowledge in predicting views on economic policy. William Gale, following up on William Nordhaus's comment, noted that survey questions that omit reference to actual budget constraints often produce very different answers than do questions framed around realistic alternatives. He reported that proposals to balance the budget or cut taxes typically receive strong support when asked without the constraint and strong opposition when asked with it.
Austan Goolsbee discussed whether the paper's results really undermined the homo economicus model. He noted that, because policies in a democracy are set by elected officials, self-interest may be better reflected in who people vote for than in their own answers to economic questions. Evidence that people's answers to survey questions were inconsistent with that model need not imply that economic policy will be inefficient. As for whether homo economicus was the relevant model for economists to use, he reasoned that what mattered was such things as whether agents knew their own tax rates, not whether they knew any facts about macroeconomic aggregates.
The great heterogeneity of beliefs among the respondents received considerable attention. Edward Glaeser noted that, since any one person's vote is irrelevant, the homo economicus model would predict that individuals will be ill informed about the issues and that beliefs will be widely dispersed. He saw this as a challenge to prevailing Bayesian models and a challenge to understanding how the market for ideas works and how beliefs are formed. In this connection, he reasoned that understanding the incentives of those who provide information as well as the framework in which people form their views was critical; for example, in school Europeans are taught in a Marxian framework whereas Americans are taught Horatio Alger. Friedman suggested analyzing the authors' data for content bias associated with different information sources. For example, how did the views of those who relied on economists for economic information differ from the views of those who relied on ministers? Andrei Shleifer pointed out that people generally seek out providers of information that confirm their own beliefs: conservatives read and listen to conservative media sources and liberals to liberal media sources. One consequence is that conservatives are most likely to be confused or wrong about those issues where the conservative media want them to be confused or wrong, and liberals are most likely to be confused or wrong about issues where the liberal media want them to be confused or wrong.
Peter Orszag was both surprised and heartened by the survey responses regarding the long-term Social Security deficit and how to reduce it. Recent political proposals have relied mainly or entirely on benefit reductions to eliminate the prospective deficits, but only 5 percent of survey respondents chose that approach. The largest group of respondents (34 percent) wanted a combination of tax hikes and benefit cuts, which
corresponds to the approach detailed in the recent book written by Orszag and Peter Diamond.
References
Barrels, Larry M. 2003. "Homer Gets a Tax Cut: Inequality and Public Policy in the American Mind." Working paper. Princeton University (August).
Black, Duncan. 1948. "On the Rationale of Group Decision-Making." Journal of Political Economy 56, no. 1: 23-34.
Blendon, Robert J., and others. 1997. "Bridging the Gap between the Public's and Economists' Views of the Economy." Journal of Economic Perspectives 11, no. 3: 105-18.
Blinder, Alan S. 1997. "Is Government Too Political?" Foreign Affairs 76 (November/December): 115-26.
Blinder, Alan S., and John Morgan. Forthcoming. "Are Two Heads Better than One? Monetary Policy by Committee." Journal of Money, Credit, and Banking.
Caplan, Bryan. 2002. "Sociotropes, Systematic Bias, and Political Failure: Reflections on the Survey of Americans and Economists on the Economy." Social Science Quarterly 83, no. 2: 416-35.
Card, David, and Alan B. Krueger. 1995. Myth and Measurement: The New Economics of the Minimum Wage. Princeton University Press.
Citrin, Jack, and Donald Philip Green. 1990. "The Self-Interest Motive in American Public Opinion." In Research in Micropolitics: Public Opinion, 1990: A Research Annual, vol. 3, edited by Samuel Long. Greenwich, Conn.: JAI Press.
Downs, Anthony. 1957. An Economic Theory of Democracy. New York: Harper and Brothers.
Fair, Ray C. 2002. "The Effect of Economic Events on Votes for President: 2000 Update." Yale University, Department of Economics. Available at fairmodel. econ.yale.edu/RAYFAIR/PDF/2002DHTM.HTM.
Feldman, Stanley. 1984. "Economic Self-Interest and the Vote: Evidence and Meaning." Political Behavior 6, no. 3:229-51.
Fiorina, Morris P. 1978. "Economic Retrospective Voting in American National Elections: A Micro-Analysis." American Journal of Political Science 22, no. 2: 426-43.
Fuchs, Victor R., Alan B. Krueger, and James M. Poterba. 1998. "Economists' Views about Parameters, Values, and Policies: Survey Results in Labor and Public Economics." Journal of Economic Literature 36, no. 3: 1387-1425.
Fuller, Dan, and Doris Geide-Stevenson. 2003. "Consensus Among Economists: Revisited." Journal of Economic Education 34, no. 4 (Fall): 369-87.
Gale, William G., and Jeffrey Rohaly. 2003. "Three-Quarters of Filets Pay More in Payroll Taxes Than in Income Taxes." Tax Notes (January 6), p. 119.
Hamilton, James T. 2003. All the News That's Fit to Sell: How the Market Transforms Information into News. Princeton University Press.
Hansen, Karsten T., James J. Heckman, and Kathleen J. Mullen. Forthcoming. "The Effect of Schooling and Ability on Achievement Test Scores." Journal of Econometrics.
Kahneman, Daniel. 1986. "Comments." In Valuing Environmental Goods: An Assessment of the Contingent Valuation Method, edited by Ronald G. Cummings, David S. Brookshire, and William D. Schulze. Totowa, N.J.: Rowman and Allanheld.
Keeter, Scott, and others. 2000. "Consequences of Reducing Nonresponse in a National Telephone Survey." Public Opinion Quarterly 64, no. 2: 125-48.
Kramer, Gerald H. 1971. "Short-Term Fluctuations in U.S. Voting Behavior, 1896-1954." American Political Science Review 65, no. 1: 131-43.
Lupia, Arthur. 1994. "Shortcuts Versus Encyclopedias: Information and Voting Behavior in California Insurance Reform Elections." American Political Science Review 88, no. 1: 63-76.
Mansbridge, Jane J., ed. 1990. Beyond Self-Interest. University of Chicago Press.
McDonald, Michael P. 2003. "On the Over-Report Bias of the National Election Study Turnout Rate." Political Analysis 11, no. 2: 180-86.
Monroe, Alan D. 1979. "Consistency between Public Preferences and National Policy Decisions." American Politics Quarterly 7, no. 1: 3-19.
Mullainathan, Sendhil, and Andrei Shleifer. 2003. "The Market for News." Working paper. Harvard University, Department of Economics (December).
Page, Benjamin I., and Robert Y. Shapiro. 1983. "Effects of Public Opinion on Policy." American Political Science Review 77, no. 1: 175-90.
Palfrey, Thomas R., and Keith T. Poole. 1987. "The Relationship between Information, Ideology, and Voting Behavior." American Journal of Political Science 31, no. 3:511-30.
Persson, Torsten, and Guido Tabellini. 2000. Political Economics: Explaining Economic Polio,. MIT Press.
Romer, David. 2003. "Misconceptions and Political Outcomes." Economic Journal 113, no. 484: 1-20.
Rosenstone, Steven J., John Mark Hansen, and Donald R. Kinder. 1986. "Measuring Change in Personal Economic Well-Being." Public Opinion Quarterly 50, no. 2 (Summer): 176-92.
Sears, David O., and others. 1980. "Self-Interest vs. Symbolic Politics in Policy Attitudes and Presidential Voting." American Political Science Review 74, no. 3: 670-84.
Silver, Brian D., Barbara A. Anderson, and Paul R. Abramson. 1986. "Who Overreports Voting?" American Political Science Review 80, no. 2:613-24.
Slemrod, Joel. 2003. "The Role of Misconceptions in Support for Regressive Tax Reform." Working paper. University of Michigan (November).
Zaller, John R. 1992. The Nature and Origins of Mass Opinion. University of Cambridge Press.
(1.) "Matter of Trust: HMOs Get Little," Managed Care (January 2004), www.managed caremag.com/archives/0401/0401.news_trust.html.
(2.) "So, Whom Do We Trust?" Toronto: Ipsos-Reid (January 22, 2003), www.acpa.ca/ press_news/2003/trust.pdf.
(3.) Program on International Policy Attitudes, University of Maryland, "Americans on Foreign Aid and World Hunger: A Study of U.S. Public Attitudes (February 2, 2001), www.pipa.org/OnlineReports/BFW/finding1.html.
(4.) Feldman (1984); also see Fiorina (1978) and Rosenstone, Hansen, and Kinder (1986).
(5.) Fair (2002)
(6.) Kramer (1971).
