Over the last few decades, has the typical American become: poorer? a little bit richer? quite a bit richer? much richer? Yes! Every answer is correct.
How can this be? It simply depends on how you decide to measure income and income growth, as Steven Rose shows in not one but two new important papers for the Urban Institute. Median income growth from 1979 to 2014 was negative, tepid or robust, according to different authors, using different methodologies:
Even the same authors can reach different conclusions depending on their approach. Compare for example the -8% growth reported by Piketty and Saez (2002) and the +33% growth reported by Piketty, Saez and Zucman (2018).
Rose painstakingly describes the different methodological choices that have such dramatic impact on results, namely:
- What counts as income? Pretax income, post-tax income, government cash transfers, all government transfers, health benefits, and employer benefits.
- Household-size adjustment. Some measures adjust for household size, while others don’t.
- Unit of analysis. Whose income is measured? Choices include tax units, households, families, or individuals.
- Inflation measure. Also known as a “price deflator.” Choices include the consumer price index (CPI-U-RS), the personal consumption expenditure price deflator (PCE), and the national income deflator.
The point here is simple but vital. Methods matter. The next time you hear a politician or pundit talking about what is happening to middle class incomes, you should ask not only how they are defining the middle class, but also how they are measuring income. The danger here is that people simply select the approach that best fits the story they want to tell.
 Different methods of household-size adjustment do exist. However, each method produces relatively similar results.