We do not believe the AFL-CIO in their estimate of the Labor Share
We are afraid that we do not believe this statistic:
The report notes workers’ share of US national income has fallen to the lowest level since the second world war.
The source is the AFL-CIO here. More specifically, this from the Bureau of Labor Statistics. Which is an estimation of labour (to revert to English spelling) productivity. Which requires estimating wages and output, of course. From which it is possible to multiply up to how much of GDP do the workers receive in wages and other compensation for going to work?
Now, we’re entirely willing to be told that we’ve got this entirely wrong. If so please do tell us where. If you’re going to then it needs to be in detail, not just who the heck are we to question official numbers?
We have long made the point about UK national income statistics that there are in fact four sectors, capital, labour, mixed and subsidies to production and taxes on consumption. It is true that the labour share has fallen in recent decades but it’s not because the capital share has risen so much as that the other two have.
The US figures, well, we’ve not seen them - which could be our data navigation skills at fault - laid out the same way. But that’s not the point at issue here. We really do not believe that the labour share of US GDP is 52.9%. It’s absolutely true that the capital share is not 47.1% but that’s that different issue from the previous paragraph.
It’s a bit odd to be measuring the labour share from the productivity statistics but that’s what is indeed being done. It’s also useful to find out where that GDP is going if it’s not to labour and it’s not to capital, which also isn’t being done here. Which is where we note this from BLS:
Labor Compensation
The measure includes accrued wages and salaries, supplements, employer contributions to employee benefit plans, and taxes. Estimates of labor compensation by major sector, required for measures of hourly compensation and unit labor costs, are based primarily on employee compensation data from the NIPA, prepared by the BEA. The compensation of employees in general government, nonprofit institutions and households are subtracted from compensation of employees in domestic industries to derive employee compensation for the business sector. The labor compensation of proprietors cannot be explicitly identified and must be estimated. This is done by assuming that proprietors have the same hourly compensation as employees in the same sector.
The way we read that is that this is not, in fact, the labour share at all. It’s the labour share of the private sector, or perhaps the business sector. The wages/benefits paid to government employees - in fact, to the wider lanyard class - are removed before the totals are calculated. So, - again, this is how we’re reading it, correct us if we’re in error - we think this is the labour share of private sector employees, not the labour share of the whole economy at all.
At which point where all the money’s going becomes obvious. It’s going off to the Lanyards.
Now, this is a big claim because this 52.9% labor share is a usual estimation. We just think it’s wrong, in that it’s not the labour share at all, it’s that private sector employees’ labour share. A very different thing.
We think the error stems from using the productivity statistics to try to calculate this labour share. For of course we cannot measure the productivity of the lanyard classes as they have no priced output. Which means their exclusion from the productivity statistics is just fine, but which then makes the productivity statistics invalid as a source of the labour share simply because we’ve excluded all the lanyards. Or, if we are to do it this way then we should be using this measure of the labour share as our measure of the costs of all those Lanyards. Which would be fun.
As we say, we could be wrong here at which point, well, where are we?
Tim Worstall