This is the main story. At a time, perhaps because of fixed exchange rates curbing the flow of speculative capital, perhaps not, productivity (viz., increasing automation) gains meant median wage growth - ordinary people participated in GDP growth.
Now this is no longer true, and we get pieces lamenting the death of blue collar jobs through automation. We'll blame robots, but perhaps we should be looking at the fundamentals of our economy - how we've set the free movement of capital above all else, to our detriment.
I disagree with this conclusion, and I tend to disfavor the PCE. The argument is that employers are spending more on benefits, therefore pay has not actually fallen. However, the largest component of the difference between the PCE and CPI is based on employer-provided healthcare; in the US, healthcare spending has been growing at faster than inflation for decades, and medical costs are one of the main avenues of wealth transfer (for example, $300 billion a year, 2% of GDP, goes in patent royalties to owners of pharmaceutical companies).
It is a bit crap to suggest that because we are paying people more so they can pay their doctors more, for worse outcomes than anyone in the rest of the developed world, they are somehow turning out better. In any case, the fact that real wages have remained essentially flat requires us to believe that workers are fine with not having made any gains in disposable income in exchange for better benefits, which seems unlikely.
Also, the fact that labor peaked in 1973, when the graph begins to diverge, should reinforce, not contradict, this argument.
The conclusion isn't that everything is hunky dory. What he does conclude is that labor's share of income hasn't much changed. What has changed is the distribution of income between laborers. As has been noted in about a million places, gains have increasingly been going to those at the top.
Your chart implies some sort of breakdown between productivity and income but does not demonstrate that productivity gains have been equally distributed. In fact, there are a lot of reasons to think that most productivity gains are going towards the top end of the labor pool as well. Due to the lever of technology, we see increasingly small (in terms of # of employees) firms serve increasingly large markets.
My point (well Scott Sumner's really) is that there hasn't been a breakdown between pay and productivity. People are still getting paid more if they are more productive. It's just that somewhere around 1970 or so the median worker stopped getting all that much more productive.
Now, maybe that's what you meant by your graph. In which case, fine. But most people look at that and they think that the median worker started getting cheated out of their productivity gains. And that's probably not what's happening.
I think that's exactly what happened. For example, what I described earlier, medical costs increasing at twice the rate of inflation in the US while outcomes did not improve much at all, is the median worker being cheated out of their productivity gains: it means they are paying more and getting less, while health care providers and people who own patents on drugs or medical devices get more. This change has absolutely zilch to do with changes in how productive anyone is and much more to do with how income is distributed, how property rights are enforced, how high-tech industry is regulated, etc.
Increased costs for health care(1) are a measure of where workers are spending their money. That's an orthogonal issue to measures of worker productivity.
1. It's a side issue, but I'd also dispute your statement that people are getting less. Life expectancy has increased ~10 years in the US since 1970.
People in the US are getting less than in other places, in terms of health care. Other countries also haven't seen the same gap in gdp/income.
Arguing that workers are less productive because they earn less is a tautology; it also requires us to believe that workers magically stopped being more productive in 1973 after decades of improvement.
1) Our health care system is certainly unusual and it might very well have problems. But, again, that has nothing to do with measures of worker productivity outside of the health care sector.
2) Other countries have significantly larger transfers from the rich to the poor. That certainly puts more spending power in the hands of the bottom half of the bell curve (which might be a good idea!), but it doesn't mean they're more productive.
3) I don't magically believe that increases in median worker productivity slowed in the 1970s, I believe it because it matches what I see in the world. GM used to employ 100s of thousands of people when it was one of the most valuable companies in America. Now Facebook holds that status with something like 15,000.
Technology has become this huge lever where fewer and fewer people are needed to make more and more stuff. So increasingly it's the people at the top end of the bell curve that are capturing these gains. A programmer can write software that gets used by more people than ever. A retail store employee is doing pretty much the same job now that they were doing 30 years ago.
> fewer and fewer people are needed to make more and more stuff.
You're arguing against yourself, here. The latter is productivity increase, the former is productivity decrease. Also, productivity growth is not new, it has been going on as long as the economy exists. Machine automation has been occurring since the industrial revolution started, it did not start in 1973.
What started in 1973 was that the high end (basically, the owners of capital) stopped sharing productivity gains - meaning that where previously, workers were able to bargain for a share of the increase, now they cannot, because of various mechanisms of taking. CEO pay did not go up 10X because CEOs were suddenly drinking tiger blood and winning the Fields medal.
My original contention was that one of these was the shift away from the Bretton Woods system, with fixed currency exchange rates based on the dollar; going from here to speculative international currency flows and floating exchange rates is one possible mechanism for making workers less able to bargain for those productivity increases.
What started in 1973 was that the high end (basically, the owners of capital) stopped sharing productivity gains - meaning that where previously, workers were able to bargain for a share of the increase, now they cannot, because of various mechanisms of taking.
NO! This is EXACTLY what I (really Scott Summers) am saying has not happened. The share of national income going to labor (vs the owners of capital) has remained almost exactly constant since 1973 (and before then as well).
I already commented on this; his assumption is that the PCE is "correct" and the CPI is "wrong". I made a criticism of PCE that you did not address; it is merely counting increased health care spending (a significant mechanism of wealth transfer) as "labor share" of national income.
This is not one of those “he said, she said” where reasonable people can disagree on whether the PCE or CPI is a better price index. This is a pay/productivity gap being invented by using the slowly moving price index (NDP, which is similar to the PCE) to make worker productivity look better, and the faster moving price index (CPI) to make real wages look lower. That’s not kosher. You need to use the same type of index for both lines on the graph.
There is no difference between using NDP and GDP; the shift emerges only when you switch between PCE and CPI. It is totally about whether the PCE or the CPI is a better index. We can see this clearly enough from the fact that there is a notch in the graph one way and not the other.
If you simply don't believe in wage stagnation, then I'm not sure what this whole conversation has been about.
Wage stagnation is definitely a thing, but it's due to increasing wage inequality between different employees not due to capital grabbing an increasing share of national income. Labor's share is remarkably stable over time:
Of particular interest to me is that if you look at the BEA numbers, they split 'labor share' into wages and non-wage compensation (benefits, SS, medicare, etc.) - the latter share has climbed over time to 20%, meaning much of the stability in the "wage share" is just increasing money being paid to Medicare.
It's not clear what's being measured and not in that chart, but, for example, this St. Louis Fed data shows both an increasing Dividend share of GDP and a declining wage share of GDP:
This is the main story. At a time, perhaps because of fixed exchange rates curbing the flow of speculative capital, perhaps not, productivity (viz., increasing automation) gains meant median wage growth - ordinary people participated in GDP growth.
Now this is no longer true, and we get pieces lamenting the death of blue collar jobs through automation. We'll blame robots, but perhaps we should be looking at the fundamentals of our economy - how we've set the free movement of capital above all else, to our detriment.
reply