In his write-up of the new data on GDP, Wonkblog's Matt O'Brien noted that Gross Domestic Income grew considerably more rapidly (or more accurately, shrank less rapidly) than GDP in the first quarter. O'Brien sees this as evidence that the economy grew more rapidly than the GDP data indicate.

That is possible, but it is also possible that the GDI data are simply in error. In principle GDP and GDI should be the same. GDP measures everything that was produced based on the sales of goods and services. GDI measures all the incomes generated in the production process. As a practical matter, they never add to be exactly the same. The Bureau of Economic Analysis generally considers the GDP measures to be more accurate since its ability to measure the sales of goods and services is better than its ability to measure income.

However there are patterns to the divergences. When there are large run-ups in asset prices (i.e. stocks and housing), the GDI measure tends to show stronger growth than the GDP measure. There is a simple explanation as to why this would be the case. If a portion of the capital gain income from a run-up in asset prices ends up being recorded as ordinary income, then the larger the capital gains, the more income will be wrongly reported. (Capital gains or losses should not be counted in GDI.)

It is likely that this would be the case. While people pay lower taxes on long-term capital gains than ordinary income, they pay the same tax rate on short-term capital gains. This means that they have no reason to be careful to distinguish these capital gains from ordinary income on their tax returns. Since tax returns provide the ultimate basis for GDI data, insofar as income is misrepresented on these returns it will lead to a misreporting of GDI.

In fact, we find a consistent pattern where GDI grew more than GDP in both the stock and housing bubble and again in the last few years with the sharp run-up in stock prices. If the capital gains explanation is correct, it means that income in the national accounts is overstated. This means that the saving rates are substantially lower than the official data show. (Saving is defined as income minus consumption, if income is overstated by 2 percent, then the saving rate is overstated by 2 percentage points, which would be close to half in recent years.)

It  also means that we can take no special solace in GDI numbers that are stronger than GDP numbers. It's just a mistake.