I mentioned theory-dependent 'facts' the other day, and here is a good example using a Brad DeLong post from yesterday. Here is RGDP and the trends DeLong sees:
There is a 'fact' of two cases of RGDP/productivity slowdown. The information transfer (IT) model trend tells a different story with different 'facts' (this graph is the same result from this post):
In this view, there hasn't been any falling productivity, just a trend towards a lower average RGDP growth rate. However, the more fundamental (and therefore better) measure in the IT model is NGDP:
I don't want to leave the impression that I am saying: Ha, ha, DeLong is wrong! Maybe the IT model is wrong (although the NGDP path is pretty excellent, I must say). I just wanted to illustrate how different theoretical models lead to different interpretations of data -- different 'facts'. And different theoretical models lead to different questions (and therefore research). Why has productivity slowed down? is a question you'd ask if you drew the trend lines DeLong drew; it's not a question that follows naturally from the IT model trend.
PS The NGDP path includes a demography slowdown component as changes in the growth rate of the labor force are the primary source of non-monetary changes. See here.
Sorry folks, but I do not see any fact at all about productivity in any of these graphs. And productivity is measurable. We are not in aetherland. Also, as far as I can tell, all of these graphs are consistent with a productivity slowdown.
ReplyDeleteNow, DeLong has drawn three trend lines to the RGDP data, with no known justification or methodology, and that is a problem, while the IT trend curves fit the data quite well. :) But I do not see any facts about productivity at all.
Hi Bill,
DeleteDivide by total hours worked.
Falling Growth of RGDP per hour is productivity slow down.
Basically dominated by RGDP growth slow down.
By that definition, productivity has risen, but the rate of growth of productivity has decreased. And that is so in all three graphs. WTP? (What's the problem?)
DeleteBTW, despite the label, GDP is not a direct measure of productivity. A lot of productivity is not reflected in GDP. For instance, the computer has made a lot of office workers more productive, but that productivity increase does not show up in GDP because of competition, so that the workers' productivity increases are not reflected in their wages. The Red Queen phenomenon.
To be clear, the productivity increases may not show up in wages, because of competition among workers, and it may not show up in prices, because of competition among businesses.
DeleteYes, it is a productivity slowdown as mentioned in the title, not actually falling productivity.
DeleteProductivity is a technical measure to account for the difference between real output given various inputs, so real output (RGDP) is part of the measure of productivity. Also the productivity we're talking about here doesn't have to show up in wages. We're talking about overall productivity, not labor productivity (which has also slowed down).
If computers have made it cheaper (or less time consuming) to produce an extra widget of output, then they are productivity enhancing. It doesn't matter if that additional output per unit input is partially reflected in wages or stock prices.
I agree that it seems much of the productivity gains have been reflected in the salaries of CEOs, stock dividends, etc rather than wage increases. Some even theorize that the reason for the overall productivity growth slowdown is because productivity increases don't result in higher wages. Workers don't try to enhance productivity when they don't get any benefit from it. Makes sense to me.
But in the IT model above, there isn't any productivity slowdown -- there really isn't anything called productivity that changes:
http://informationtransfereconomics.blogspot.com/2014/12/the-information-transfer-solow-growth.html
In the sense that computers are productivity enhancing in the IT model it's that they open up new states in the economic state space and speed up transitions into them.
However -- we have a competing effect of the multi-dimensional state space not growing as quickly relative to a single dimensional measure like GDP as the size of the economy increases. Computers may be opening huge swaths of state space, but since the space is already so large, the fractional increase is less.
Maybe that is why computers (famously) don't show up in the productivity stats ...
"If computers have made it cheaper (or less time consuming) to produce an extra widget of output, then they are productivity enhancing. It doesn't matter if that additional output per unit input is partially reflected in wages or stock prices."
DeleteThat's what I would say. However, GDP is measured in money, so it does not measure the additional output in physical goods or increased quality if the price does not change because of competition.
"Some even theorize that the reason for the overall productivity growth slowdown is because productivity increases don't result in higher wages. Workers don't try to enhance productivity when they don't get any benefit from it. Makes sense to me."
That's a micro argument. One macro argument is that without wages rising with productivity, the workers, who are also consumers, cannot afford to buy the same share of what they produce, and so the economy as a whole slows down. This ties in with John Stuart Mills's explanation of recessions, which were apparently general gluts, which were theoretically impossible. Mill pointed out that they could be explained as an insufficiency of money. So if workers produce more because of rising productivity, but they are not paid enough money to keep up with the additional production, then we get a relative insufficiency of money, even if there is not an official recession.
"But in the IT model above, there isn't any productivity slowdown -- there really isn't anything called productivity that changes:"
Right. I have been with you on that all along. :)
But if the graph is **by definition** one of productivity slowdown, the the, ahem, fact that the IT trend fits the graph is a happy accident. There is no need for any further explanation of the graph, or of a slowdown of productivity.
When I was living in Japan as a young man I had to duck my head to avoid hitting store awnings. But the next generation of Japanese were taller, because of eating more protein, as I heard. The failure of the Japanese to continue growing taller does not have to do with their failure to eat more protein. ;) That does not mean that their rate of growth in height as a group did not slow down.
Hmmm. I wonder if the assumption that GDP is a measure of productivity rests upon the assumption that wages rise in proportion to productivity increases. Otherwise, how can GDP, measured in money, keep up with productivity, measured in different ways? IIUC, the assumption that wages rise in proportion to productivity was normally made by economists until the 1980s, when it proved to be false.
DeleteYou see that same bent-RGDP curves in Japan and Switzerland: it's not as unusual as Delong suggests -- only really unusual for postwar US. Nevertheless, this sort of RGDP breakdown has happened before in the US, most notably (pre-tribulation-war) in the 1840s and 1930s. A growing proportion of the below-trend pace of NGDP growth can now be attributed to RGDP only since 2008.
ReplyDeleteWhy can't we simply attribute this to tight monetary policy (which would manifest also in low interest rates and high debt levels)? Tight monetary policy since 1980 gradually reduced NGDP to below-trend, and is now reducing RGDP similarly.
Monetary policy would appear to be the occam razor solution to below trend RGDP, rather than this seemingly random grab bag of unsystematic anecdotal causes.
Actually, that is pretty much the explanation of the bend in the 1980s for the US:
Deletehttp://informationtransfereconomics.blogspot.com/2015/02/what-if-us-inflation-had-been-2-since.html
There are three pieces to the IT model:
-- A general trend towards lower growth vs monetary expansion for all countries
-- A change to quasi-inflation targeting in the 1980s (which is why the bend is more pronounced)
-- Demographics of a slowing civilian labor force
The first one doesn't have a temporal component -- it doesn't tell us at what time it happens (it doesn't tell us how fast we follow the NGDP-M0 path). The latter two give us the specific temporal behavior.
Yes, the bend unwound, with 60-70 inflation mean reverting into 90-00 deflation. The '10s brought us below trend - presumably to be unwound with future inflation.
ReplyDeleteThe three pieces might be due to 1) growing debt to the point of overleverage, 2) tight monetary policy, 3) a slowing workforce is the symptom, not the cause.