söndag 30 juni 2013

Bartlett om fallande löneandelen, och Krugman om rents

Bruce Bartlett skriver på New York Times ekonomiblogg Economix om att löneandelen minskat senaste decennierna i de rika länderna:
The report [OECD 2012] identifies the substitution of capital for labor in many industries as a cause of labor’s declining share. A shorthand term for this is “automation,” but it really goes beyond simply replacing workers with machines. It also includes the spread of technology, like computers and the Internet, that has increased the productivity of some workers, allowing them to do the work of several workers in the recent past, but not others.

But as a recent report from the International Labor Organization points out, the gains to higher productivity resulting from technological innovation are not going entirely to workers. It says that since 1999, average labor productivity in a cross-section of countries has increased twice as much as wages.

A new study by Loukas Karabarbounis and Brent Neiman of the University of Chicago’s Booth School of Business identifies the low cost of capital as a key source of labor’s woes. Because of low interest rates and low taxes on investment, companies have been encouraged to substitute technology, machinery and equipment for labor, which explains about half the decline in labor’s share of income, according to their estimate.

Interestingly, this means that rising interest rates, which have roiled the bond market in recent weeks, are part of the solution to labor’s declining share of income. By raising the cost of capital, higher rates will make labor relatively more attractive vis-à-vis capital. Higher taxes on capital would also have that effect as well.

As Professor Neiman put it in an e-mail to me,
Rising real interest rates, higher prices of investment goods, higher depreciation rates, or even increases in corporate tax rates would (everything else equal) push labor’s share upward as they all generate increases in the cost of capital.
This is important because many economists routinely assert that more capital investment is critical to increase productivity, which, in turn, will automatically lead to higher wages. While this is undoubtedly true up to a point, we may have passed the point where it is still true in economically advanced countries such as the United States."
Bruce Bartlett, "Labor's Declining Share is an International Problem", NYT Economix 25 juni 2013

-- annat nytt ur USA-debatten om löneandelar och liknande: Bruegel ger en finfin översikt över bloggdebatten utifrån Paul Krugmans inlägg om ökade vinster utan ökade investeringar, med svar från DeLong, Buttonwood, Owen Zidar, och Nick Rowe. Krugmans förklaring till att vinstandelen kan öka utan att det leder till högre investeringar är att inslaget av rents, avkastning på monopoltillgångar/fördelar som inte kan multipliceras, har ökat i ekonomin:
"the growing importance of monopoly rents: profits that don’t represent returns on investment, but instead reflect the value of market dominance. Sometimes that dominance seems deserved, sometimes not; but, either way, the growing importance of rents is producing a new disconnect between profits and production and may be a factor prolonging the slump." 
Paul Krugman, "Profits Without Production", New York Times 20 juni

T ex Justin Wolfers har påpekat (på Twitter) att Krugman inte presenterar några empiriska bevis för sin hypotes, utan mer använder sig av anekdoter. Och det är sant, och kanske inte helt förvånande från en opinionstext i NYT och några blogginlägg. Det är alltså i allra högsta grad en hypotes, inget bevisat faktum, att det är rents som ligger bakom vinst-investerings-divergensen som diskuterats i flera år nu. Men en intressant hypotes. (Man kunde också snacka om löneledd tillväxt, också det en intressant om än spekulativ hypotes.)

I ett blogginlägg förklarar Krugman också lite hur han tänker kring en makroekonomisk modell som ligger bakom ledaren "Profits Without Production" som jag citerat ovan. Krugman förklarar skissen till modell så här:
"OK, imagine an economy in which two factors of production, labor and capital, are combined via a Cobb-Douglas production function to produce a general input that, in turn, can be used to produce a large variety of differentiated products. We let a be the labor share in that production function.

The differentiated products, in turn, enter into utility symmetrically with a constant elasticity of substitution function, a la Dixit-Stiglitz; however, I assume that there are constant returns, with no set-up cost. Let e be the elasticity of substitution; it’s a familiar result that in that case, and once again assuming that the number of differentiated products is large, e is the elasticity of demand for any individual product.

Now consider two possible market structures. In one, there is perfect competition. In the other, each differentiated product is produced by a single monopolist. It’s possible, but annoying, to consider intermediate cases in which some but not all of the differentiated products are monopolized; I haven’t done the algebra, but it’s obvious that as the fraction of monopolized products rises, the overall result will move away from the first case and toward the second.

So, with perfect competition, labor receives a share a of income, capital a share 1-a, end of story.

If products are monopolized, however, each monopolist will charge a price that is a markup on marginal cost that depends on the elasticity of demand. A bit of crunching, and you’ll find that the labor share falls to a(1-1/e).

But who gains the income diverted from labor? Not capital — not really. Instead, it’s monopoly rents. In fact, the rental rate on capital — the amount someone who is trying to lease the use of capital to one of those monopolists receives — actually falls, by the same proportion as the real wage rate.

In national income accounts, of course, we don’t get to see pure capital rentals; we see profits, which combine capital rents and monopoly rents. So what we would see is rising profits and falling wages. However, the rental rate on capital, and presumably the rate of return on investment, would actually fall.

What you have to imagine, then, is that some factor or combination of factors — a change in the intellectual property regime, the rise of too-big-to-fail financial institutions, a general shift toward winner-take-all markets in which network externalities give first movers a big advantage, etc. — has moved us from something like version I to version II, raising the profit share while actually reducing returns to both capital and labor.

Am I sure that this is the right story? No, of course not. But something is clearly going on, and I don’t think simple capital bias in technology is enough."

-- Nick Rowe har också ett väldigt fint blogginlägg där han svarar på ett inlägg från Bruce Bartlett om finansialisering, löneandelen och ojämlikhet (det är lite fint och lite sorgligt att se hur mainstreamekonomer nu närmar sig analyser som postkeynesianer gjort i typ tio år, fast utan att referera till PK!). Rowe kommenterar i princip att Bartletts modell verkar funka fint för USA, men inte lika bra för Kanada.

Uppdatering 30 oktober 2014
"I caught a bit of CNBC in the locker room this morning, and they were talking about stock buybacks. Oddly — or maybe not that oddly, given my own experiences with the show — nobody brought up what I would have thought was the obvious question. Profits are very high, so why are companies concluding that they should return cash to stockholders rather than use it to expand their businesses?
After all, we normally think of high profits as a signal: a profitable business is one people should be trying to get into. But right now we see a combination of high profits and sluggish investment
What’s going on? One possibility, I guess, is that business are holding back because Obama is looking at them funny. But more seriously, this kind of divergence — in which high profits don’t signal high returns to investment — is what you’d expect if a lot of those profits reflect monopoly power rather than returns on capital.
More on this in a while."
Paul Krugman, "The profits--investments disconnect", Conscience of a Liberal 24 oktober
Brad Delong svarar: "Is there really a profit--investment disconnect?", 25 oktober

DeLong menar i grund och botten att de låga investeringarna nu är ett resultat av krisen:
"As a result of the housing bubble, the mortgage frauds, the attempts at regulatory arbitrage on their balance sheets by the money-center universal banks, the financial crisis, et sequelae, U.S. real GDP today is now 12% below what we back in 2007 expected it to be now. Since the post financial-crisis trough U.S. real economic growth has proceeded at 2.24%/year, compared to the 3.00%/year growth rate we saw between 1990 and 2007.
So far there are no signs anywhere that the gap between today and the pre-2007 trend in levels will be made up. So far there are no signs anywhere that the gap between today and the pre-2007 trend in growth rates will be made up.
That means that, come 2024 a decade hence, we can now expect a U.S. economy to be 19.5% smaller than the economy we confidently projected as of 2007 we would have.
And, with a capital-output ratio of roughly 3, that means that between 2007 and 2024 cumulative net investment will be lower than projected back in 2007 by 58.5%-point years of GDP–and cumulative gross investment considerably lower.
Given this extraordinary shift in our long-run growth trajectory and in the investment requirements consistent with that trajectory, is it really surprising that investment in this “recovery” is not matching previous patterns?
Paul Krugman says that because profits are high, the marginal return on capital is high, and that means that firms ought to be eager to add to their capital stocks via investment in order to achieve high returns and further boost their profits. This seems to me to rely on an identification of average-Q with marginal-Q that I have always found suspect. Remember: Profits are not high now because demand is high, throughput is high, and capacity is being fully used. Profits are high now because the labor share is unusually low. Firms almost surely, given the collapse in the labor share over the past fifteen years, operating with too much capital and too little labor along the isoquant to be profit maximizing. Why shouldn’t we presume that–just as after 1973 and 1979 they shifted to more energy-intensive mixes, and productivity growth was thus lower than previous experience would have predicted–firms are now shifting to more labor- and information-intensive mixes, and that investment (in everything except real investments in information technology) is lower than previous experience would have predicted?
I do see a puzzle needing explanation in the extraordinary shortfall in housing investment–in the now 8 million people who ought to be out on their own in apartments and houses but who are instead living in their sisters’ or other relatives’ basements.
I still have to be convinced that there is any shortfall or puzzle needing explanation in non-housing investment..."

Inga kommentarer: