There are three reasons this looks so different from the picture of flat wages and rising productivity that we are used to. First, it's the average wage rather than the median wage. 1/n
Second, it adjusts for inflation using the GDP deflator, which reflects the prices of everything produced in the economy, rather than the CPI or PCE, which reflect the prices of consumption goods. 2/n
Third, it loks at shares of net output rather than gross domestic product. That is, it subtracts depreciation from the capital share. 3/n
Since average wages have risen faster than median wages, the price of consumption goods have risen faster than non-consumption goods, and measured depreciation faster than output, all these adjustments close the gap between pay and productivity. 4/n
Individually these choices are defensible. We can debate whether this way of looking at it is more appropriate than the usual way - or rather, we can debate which questions each way of looking at it is better usited to answer. 5/n
To the extent that the gap is smaller using average rather than median pay, for example, that suggests that rising inequality is more about the distribution of wage income rather than division between wages and profits. That's a legitimate question! 6/n
But I think there is a sort of subtle point of logic that means you cannot combine these adjustments in this way. Adjusting for inflation using the GDP deflator means you are asking what share of all output, including capital goods, goes to workers' pay. 7/n
But if that's the question you're asking, I don't think you can subtract depreciation only from capital income. If the costs of maintaining capital goods fall only on capital owners, then logically the benefits of lower prices for capital goods must go only to them also. 8/n
I think you can deflate wages by the GDP deflator, or you can look at wages as a share of net product and capital income as the remainder. But I don't think you can logically do both. 9/n
To see this, consider a hypothetical example. Suppose we have a steady state economy where new investment each year just equals depreciation. Some fixed fraction of the capital stock is replaced each year. 10/n
Now suppose that the depreciation rate rises by 10 percent and the cost of capital goods (measured however, don't @ me) falls by 10 percent. What happens to the factor shares? 11/n
The economy is producing exactly the same basket of final gods in the first period and in the second one, with the exact same inputs of labor and capital. The flow of money incomes to capital-owners and workers is exactly the same in both periods. 12/n
The consumpion baskets of workers and capital-owners are exactly the same in both periods. So logically, we ought to agree that any measurement of the labor and capital shares ought to give the same result in both periods, right? 13/n
But if we use @DonFSchneider 's approach, we will find that the labor share has *risen*. Because we are now adjusting money wages by a lower deflator, as a result of the fall in the price of capital goods. In effect we are saying that real wages have risen. 14/n
@DonFSchneider But net output is the same, since the rise in the real output of capital goods is just equal to the higher deflation rate. So by this procedure, the capital share must have fallen. Even tho, again, both money incomes and respective consumption baskets are unchanged. 15/n
@DonFSchneider The conclusion is that if you are going to adjust wages for inflation using the output deflator, yu must also distribute depreciation proportionately across labor and capital income. (Or just dn't subtract it - the results will be the same.) 16/n
@DonFSchneider You can't implicitly assume labor exercises claim on whole product at one point in a calculation, and assume that spending on capital goods comes only out of capital income at another point in the same calculation. Both assumptions are reasonable, but only one at a time. 17/n
@DonFSchneider I think that this is about as close to a logical proof as one can usefully get in macro.
(And leaving aside the profound measurement problems which make net product not really a usable concept at all.)
18/18
(And leaving aside the profound measurement problems which make net product not really a usable concept at all.)
18/18
