A recent Congressional Budget Office (CBO) report found household incomes still have not recovered from the recession. What has gone wrong? Some liberal analysts believe they know. They point to data from the Bureau of Labor Statistics (BLS) that track’s the proportion of income going to labor (instead of capital). This chart shows labor’s share falling to historic lows. Ergo, they conclude bargaining power has shifted from labor to capital, holding down incomes. Paul Krugman once picked this as his chart of the year.
Had Krugman dug into these numbers more, he would have made another selection. The BLS numbers have shifted for reasons unrelated to the bargaining power of workers vs. employers.
The BLS chart is problematic for two reasons. First, it measures workers’ share of gross income — before accounting for depreciation. However, depreciation has increased substantially over the past generation. Computers and software wear out (or become obsolete) relatively quickly. Consequently, businesses today must spend more on replacing worn out equipment.
The BLS counts these depreciation expenses as capital income, so this change mechanically reduces labor’s share of gross income. But business owners don’t get to consume this money. Looking at labor’s share of net income — what remains available to workers and business owners after depreciation — makes much more sense.
Second, the BLS changed how it measures self-employment income around the turn of the century. Allocating self-employment income between labor and capital is tricky: The self-employed work for themselves.
Until 2000, the BLS methods treated self-employment income as being roughly 80 percent labor and 20 percent capital. Then BLS switched methods and now splits self-employment income roughly 50–50 between labor and capital. This change also mechanically reduced BLS’s measured labor share. However, it does not reflect a real economic shift.
It turns out these seemingly minor choices matter. In a recent report, I examined labor share excluding depreciation and measuring self-employment income consistently. This changes the picture dramatically. The BLS headline chart show’s labor’s share of income dropping about 8 percentage points since 1948. The more accurate measure shows labor’s share virtually unchanged over that period.
The economy has many problems. But had Krugman looked more deeply into the data he would have found “redistribution from labor in general to capital” is not among them.