September 27, 2014
Justin Wolfers is trying to sell gross domestic income (GDI) as a more accurate measure of growth than GDP. He notes that gross domestic income grew at a 2.2 percent annual rate in the first half of 2014 and says that this “more accurate” measure of the economy should be taken over the 1.2 percent annual rate shown by the more widely used GDP measure.
In principle GDI and GDP should show the same growth. GDI measures the economy by measuring the incomes generated in production (e.g. wages, profits, interest, and rents). GDP measures the economy by counting the goods and services sold (consumption, investment, government, net exports, and inventories). In principle they should show the exact same number, but due to errors in measurement they always differ and sometimes by a large amount. (The gap is defined as the statistical discrepancy, which is GDP minus GDI.)
While Wolfers tells readers that the GDI measure is more accurate the good folks at the Bureau of Economic Analysis generally say that the GDP numbers provide a better measure. The data (Table 1.17.6) show that GDI is far more erratic than GDP. For example, if we believe the GDI measure then the economy grew at a 7.2 percent annual rate in the 2nd quarter of 2012 and then slowed to a 0.6 percent rate in the third quarter. The GDI data also show the recovery barely budged in the second half of 2009 even as the stimulus kicked in, growing at just a 1.1 percent annual rate. Growth then surged to a 5.7 percent annual rate in the first quarter of 2010 before falling back to a 0.5 percent rate in the second quarter.
If that doesn’t sound like the economy you remember there is an alternative explanation for the erratic movements in GDI. David Rosnick and I did a paper regressing the changes in the statistical discrepancy against lagged measures of capital gains. We found a strong relationship with the GDI becoming a larger negative number (GDI rises relative to GDP) following periods of strong capital gains.
For this to be plausible we would need a story whereby some amount of capital gains income shows up as ordinary income. (Capital gains income is not supposed to be counting in GDI.) Since one of the sources for GDI is tax returns, this seems plausible. While long-term capital gains are taxed at a lower rate than ordinary income, short-term gains (assets held less than one year) are taxed at the same rate. This means that people filing tax returns have no particular reason to distinguish between capital gains income and ordinary income.
If we hypothesize that some amount of capital gains income always shows up as ordinary income, then we would expect that the amount of capital gains income showing up as ordinary income would be higher when people have lots of capital gains income. This means that when there is a big run-up in asset prices, we would expect the statistical discrepancy to become a larger negative number, as the data show. See, economics is simple and fun.
Comments