Richard Burkhauser on the Middle Class

Episode Summary

Executive Summary: Russ Roberts and Richard Burkhauser examine why claims that the American middle class is disappearing depend heavily on definitions. Burkhauser argues that using tax units, household size adjustments, transfers, taxes, and noncash benefits produces a far more favorable picture of middle-class well-being than market income alone, and that apparent stagnation largely reflects narrow measurement choices rather than broad economic decline.

Main Topics: How income is measured changes the story (Priority: 5/5): The episode centers on how different definitions of income—market income vs. post-transfer income, pre-tax vs. post-tax, and cash vs. noncash compensation—produce very different conclusions about the middle class. Tax units vs. households (Priority: 5/5): Burkhauser explains that using tax units rather than households can create more low-income units and make median income growth look weaker, while households better reflect shared consumption and economic well-being. Economies of scale and household size adjustment (Priority: 5/5): The discussion covers why household income should be adjusted for family size and economies of scale, typically using the square root of household size, to better compare living standards across households. Contrasting findings on inequality and middle-class stagnation (Priority: 4/5): Burkhauser contrasts his results with Piketty and Saez, arguing that their focus on tax-unit market income supports a stagnation narrative that does not capture transfers, taxes, or household sharing. Role of taxes, transfers, and in-kind benefits (Priority: 5/5): The conversation shows that once government transfers, payroll/income taxes, and employer-provided health insurance are added, measured growth in median well-being rises substantially. Interpretation, ideology, and policy implications (Priority: 4/5): Both speakers emphasize that statistical choices are often political and that policy debates about inequality should account for incentives, work disincentives, and the broader counterfactual effects of redistribution.

Key Arguments: Narrow measures of income can falsely suggest the middle class is stagnant even when real well-being is improving. Households are a more realistic unit of analysis than individuals or tax units because people share resources and benefit from economies of scale. Median income is useful because it avoids distortion from top outliers, but the choice of income concept still matters greatly. Piketty-Saez style tax-unit market-income measures answer a different question than whether typical Americans are better off. Including transfers, lower tax burdens, and noncash compensation such as health insurance materially raises measured growth in middle-class living standards. Rising inequality in market income does not necessarily mean declining living standards, because taxes and transfers offset some of the market distribution. Policy changes alter behavior: transfer programs and tax rates affect work incentives and therefore observed market income. The increase in measured inequality around 1992-1993 partly reflects changes in Census measurement and top-coding methods, not only real economic change.

Data Points: Median tax-unit market income growth, 1979-2007: 3.2% - Using Piketty-Saez-style tax units and market income in the CPS, Burkhauser says median income barely changed. Median household market income growth, 1979-2007: 15.2% - Changing the unit from tax unit to household substantially raises measured growth. Median household pre-tax, post-transfer income growth, 1979-2007: 23.6% - Including government transfers further improves the picture of middle-class well-being. Median household income after taxes, 1979-2007: 29.3% - After simulating federal, state, and payroll taxes, median growth is higher because tax burden falls relative to income. Median household income including noncash benefits: 36.7% - Adding employer-provided health insurance and the insurance value of Medicare/Medicaid boosts measured gains further. Employer-provided health insurance coverage: Approximately 75% of workers - Burkhauser says most workers or their spouses receive health insurance through an employer. Household size equivalence scale: Square root of household size (0.5 elasticity) - Used to adjust income for economies of scale in consumption. Median household income level: About $50,000 - Russ Roberts cites the approximate median household income level in the discussion. Median person income level: About $60,000 - Burkhauser notes the median person has higher income than the median household because larger households tend to have more people and more income. Change in inequality captured by CPS around 1992-1993: Measurement break - Burkhauser says the Census improved its ability to capture top incomes, affecting apparent inequality trends. CPS income interview frequency: March each year - Detailed income questions are asked in March, while the survey also tracks employment monthly. Annual CPS tenure in sample: 16 months - Respondents remain in the CPS for a period of 16 months. Single mothers’ market income trend: Dramatic increase since 1996 - Attributed to TANF and earned income tax credits encouraging work over welfare dependence.

Pivotal Quotes: "the middle shrank. It disappeared. But where did all those people go? Well, they disproportionately became richer." — Richard Burkhauser: Explaining his early finding that the apparent disappearing middle class mostly moved upward rather than downward. "The issue is how do you determine who the median American is? How do you line them up?" — Russ Roberts: Pressing on the importance of choosing the right unit of analysis for median income comparisons. "the right answer depends on what the right question is." — Richard Burkhauser: Summarizing his view that conflicting inequality findings reflect different measurement goals rather than one definitive truth.

Implications: Listeners should be cautious about headline claims on inequality: income trends depend on definitions, units, and adjustments. For policy, transfers and noncash benefits matter, and market-income inequality alone is not a full measure of living standards.