Aviation in the 21st century. I’m typing this sitting on a plane, en route to LA. The plane is a Boeing 737-800. The 737 is the best-selling commercial airliner on earth; reading its Wikipedia page should raise some serious doubts about the idea that we live in an era of accelerating technological change. I’m not sure how old the plane I’m sitting on is, but it could be 15 years; the 800-series was introduced in its present form in the late 1990s. With airplanes, unlike smartphones, a 20-year old machine is not dramatically — is not even noticeably — different from the latest version. The basic 737 model was first introduced in 1967. There have been upgrades since then, but to my far from expert eyes it’s striking how little changed tin 50 years. The original 737 carried 120 passengers, at speeds of 800 km/h on trips of up to 3,000 km, using 6 liters of fuel per kilometer; this model carries 160 passengers (it’s longer) at speeds of 840 km/h on trips of 5,500 km, using 5 liters of fuel per kilometer. Better, sure, but probably the main difference you’d actually notice from a flight 50 years ago is purely social: no smoking. In any case it’s pretty meager compared that with the change from 50 years earlier, when commercial air travel didn’t exist. The singularity is over; it happened on or about December 1910.
Unnatural rates. Here’s an interesting post on the New York Fed’s Liberty Street blog challenging the ideas of “natural rates” of interest and unemployment. good: These ideas, it seems to me, are among the biggest obstacles to thinking constructively about macroeconomic policy. Obviously it’s example of, well, naturalizing economic outcomes, and in particular it’s the key ideological element in presenting the planning by the central bank as simply reproducing the natural state of the economy. But more specifically, it’s one of the most important ways that economists paper over the disconnect between the the economic-theory world of rational exchange, and the real world of monetary production. Without the natural rate, it would be much hard to pretend that the sort of models academic economists develop at their day jobs, have any connection to the real-world problems the rest of the world expects economists to solve. Good to see, then, some economists at the Fed acknowledging that the natural rate concepts (and its relatives like the natural rate of unemployment) is vacuous, for two related reasons. First, the interest rate that will bring output to potential depends on a whole range of contingent factors, including other policy choices and the current level of output; and second, that potential output itself depends on the path of demand. Neither potential output nor the natural rate reflects some deep, structural parameters. They conclude:
the risks associated with monetary easing are asymmetric. That is, excessive easing can be reversed, but excessive tightening may cause irreversible damage to the economy’s potential output.
In the research described in this blog, we focus on the effect of recessions on human capital. Recessions may affect potential output through other channels as well, such as lower capital accumulation, lower labor force participation, slow productivity growth, and so forth. Our research would suggest that to the extent that these mechanisms are operative, a monetary policy that seeks to track measured natural rates—of unemployment, interest rates, and so forth—might be insufficiently accommodative to engineer a full and quick recovery after a large recession. Such policies fall short because in a world with hysteresis, “natural” rates are endogenous. Policy should set these rates, not track them.
Also on a personal level, it’s nice to see that the phrases “potential output,” “other channels,” “lower labor force particiaption,” and “slow productivity growth” all link back to posts on this very blog. Maybe someone is listening.
More me being listened to: Here is a short interview I did with KCBS radio in the Bay area, on what’s wrong with economics. And here is a nice writeup by Cory Doctorow at BoingBoing of “Disgorge the Cash,” my Roosevelt paper on shareholder payouts and investment.
Still disgorging. Speaking of that: There were two new working papers out from the NBER last week on corporate finance, governance and investment. I’ve only glanced at them (end of semester crunch) but they both look like important steps forward for the larger disgorge the cash/short-termism argument. Here are the abstracts:
Lee, Shin and Stultz – Why Does Capital No Longer Flow More to the Industries with the Best Growth Opportunities?
With functionally efficient capital markets, we expect capital to flow more to the industries with the best growth opportunities. As a result, these industries should invest more and see their assets grow more relative to industries with the worst growth opportunities. We find that industries that receive more funds have a higher industry Tobin’s q until the mid-1990s, but not since then. Since industries with a higher funding rate grow more, there is a negative correlation not only between an industry’s funding rate and industry q but also between capital expenditures and industry q since the mid-1990s. We show that capital no longer flows more to the industries with the best growth opportunities because, since the middle of the 1990s, firms in high q industries increasingly repurchase shares rather than raise more funding from the capital markets.
And:
Gutierrez and Philippon – Investment-less Growth: An Empirical Investigation
We analyze private fixed investment in the U.S. over the past 30 years. We show that investment is weak relative to measures of profitability and valuation… We use industry-level and firm-level data to test whether under-investment relative to Q is driven by (i) financial frictions, (ii) measurement error (due to the rise of intangibles, globalization, etc), (iii) decreased competition (due to technology or regulation), or (iv) tightened governance and/or increased short-termism. We do not find support for theories based on risk premia, financial constraints, or safe asset scarcity, and only weak support for regulatory constraints. Globalization and intangibles explain some of the trends at the industry level, but their explanatory power is quantitatively limited. On the other hand, we find fairly strong support for the competition and short-termism/governance hypotheses. Industries with less entry and more concentration invest less, even after controlling for current market conditions. Within each industry-year, the investment gap is driven by firms that are owned by quasi-indexers and located in industries with less entry/more concentration. These firms spend a disproportionate amount of free cash flows buying back their shares.
I’m especially glad to see Philippon taking this question up. His Has Finance Become Less Efficient is kind of a classic, and in general he somehow seems to manages to be both a big-time mainstream finance guy and closely attuned to observable reality. A full post on the two NBER papers soon, hopefully, once I’ve had time to read them properly.
“Sets” how, exactly? Here’s a super helpful piece from the Bank of France on the changing mechanisms through which central banks — the Fed in particular — conduct monetary policy. It’s the first one in this collection — “Exiting low interest rates in a situation of excess liquidity: the experience of the Fed.” Textbooks tell us blandly that “the central bank sets the interest rate.” This ignores the fact that there are many interest rates in the economy, not all of which move with the central bank’s policy rate. It also ignores the concrete tools the central bank uses to set the policy rate, which are not trivial or transparent, and which periodically have to adapt to changes in the financial system. Post-2008 we’ve seen another of these adaptations. The BoF piece is one of the clearest guides I’ve seen to the new dispensation; I found it especially clarifying on the role of reverse repos. You could probably use it with advanced undergraduates.
Zoltan Pozsar’s discussion of the same issues is also very good — it adds more context but is a bit harder to follow than the BdF piece.
When he’s right, he’s right. I have my disagreements with Brad DeLong (doesn’t everyone?), but a lot of his recent stuff has been very good. Here are a couple of his recent posts that I’ve particularly liked. First, on “structural reform”:
The worst possible “structural reform” program is one that moves a worker from a low productivity job into unemployment, where they then lose their weak tie social network that allows them to get new jobs. … “Structural reforms” are extremely dangerous unless you have a high-pressure economy to pull resources out of low productivity into high productivity sectors.
The view in the high councils of Europe is that, when there is a high-pressure economy, politicians will not press for “structural reform”: there is no obvious need, and so why rock the boat? Politicians kick every can they can down the road, and you can only try “structural reform” when unemployment is high–and thus when it is likely to be ineffective if not destructive.
This gets both the substance and the politics right, I think. Although one might add that structural reform also often means reducing wages and worker power in high productivity sectors as well.
Second, criticizing Yellen’s opposition to more expansionary policy,which she says is no longer needed to get the economy back to full employment.
If the Federal Reserve wants to have the ammunition to fight the next recession when it happens, it needs the short-term safe nominal interest rate to be 5% or more when the recession hits. I believe that is very unlikely to happen without substantial fiscal expansion. … In the world that Janet Yellen sees, “fiscal policy is not needed to provide stimulus to get us back to full employment.” But fiscal stimulus is needed to create a situation in which full employment can be maintained…. if we do not shift to a more expansionary fiscal policy–and the higher neutral rate of interest that it brings–now, what do we envision will happen when the next recession arrives?
This is the central point of my WCEG working paper — that output is jointly determined by the interest rate and the fiscal balance, so the “natural rate” depends on the current stance of fiscal policy. Plus the argument that, in a world where the zero lower bound is a potential constraint — or more broadly, where the expansionary effects of monetary policy are limited — what is sometimes called “crowding out” is a feature, not a bug. Totally right, but there’s one more step I wish DeLong would take. He writes a lot, and it’s quite possible I’ve missed it, but has he ever followed this argument to its next logical step and concluded that the fiscal surpluses of the 1990s were, in retrospect, a bad idea?
Farmer on government debt. Also on government budgets, here are some sensible observations on the UK’s, from Roger Farmer. First, the British public deficit is not especially high by historical standards; second, past reductions in debt-GDP ratios were achieved by growth raising the denominator, not surpluses reducing the numerator; and third, there is nothing particularly desirable about balanced budgets or lower debt ratios in principle. Anyone reading this blog has probably heard these arguments a thousand times, but it’s nice to get them from someone other than the usual suspects.
Deviation and trend. I was struck by this slide from the BIS. The content is familiar; what’s interesting is that they take the deviation of GDP from the pre-criss trend as straightforward evidence of the costs of the crisis, and not a demographic-technological inevitability.
Cap and dividend. In Jacobin, James Boyce and Mark Paul make the case for carbon permits. I used to take the conventional view on carbon pricing — that taxes and permits were equivalent in principle, and that taxes were likely to work better in practice. But Boyce’s work on this has convinced me that there’s a strong case for preferring dividends. A critical part of his argument is that the permits don’t have to be tradable — short-term, non transferrable permits avoid a lot of the problems with “cap and trade” schemes.
Why teach the worst? In a post at Developing Economics, New School grad student Ingrid Harvold Kvangraven forthrightly makes the case for teaching “the worst of mainstream economics” to non-economists. As it happens, I don’t agree with her arguments here. I don’t think there’s a hard tradeoff between teaching heterodox material we think is true, and teaching orthodox material students will need in future classes or work. I think that with some effort, it is possible to teach material that is both genuinely useful and meaningful, and that will serve students well in future economics class. And except for students getting a PhD in economics themselves — and maybe not even them — I don’t think “learning to critique mainstream theories” is a very pressing need. But I like the post anyway. The important thing is that all of us — especially on the heterodox side — need to think more of teaching not as an unfortunate distraction, but as a core part of our work as economists. She takes teaching seriously, that’s the important thing.
Apple in the balance of payments. From Brad Setser, here’s a very nice example of critical reading of the national accounts. Perhaps even more than in other areas of accounts, the classification of different payments in the balance of payments is more or less arbitrary, contested, and frequently changed. It’s also shaped more directly by private interests — capital flight, tax avoidance and so on often involve moving cross-border payments from one part of the BoP to another. So we need to be even more scrupulously attentive with BoP statistics than with others to how concrete social reality gets reflected in the official numbers. The particular reality Setser is interested in is Apple’s research and development spending in the US, which ought to show up in the BoP as US service exports. But hardly any of it does, because — as he shows — Apple arranges for almost all its IP income to show up in low-tax Ireland instead. To me, the fundamental lesson here is about the relation between statistical map and economic territory. But as Setser notes, there’s also a more immediate policy implication:
Trade theory says that if the winners from globalization compensate the losers from globalization, everyone is better off. But I am not quite sure how that is supposed to happen if the winners are in some significant part able to structure their affairs so that a large share of their income is globally (almost) untaxed.
Regarding the second post of DeLong about the Fed working with the government in order to get both full employment and a high enough Federal Funds Rate.
Although I agree that a large enough FFR would be helpful in the next recession, I think we do not take too much into account the fact that, although the interest rate channel does not have a strong impact on cyclical expenditures (such as fixed investment), it does have a substantial impact on the exchange rate as well as on the financial/economic conditions of countries long on the dollar (either because they peg their currency or because they borrow heavily in dollars).
As a result, even if fiscal policy was used to push the economy to “full employment” (however you want to define full employment), this does not mean that quickly raising the short-term interest rate in the current global environment would not have a strong impact on the US exchange rate as well as on global economic conditions.
That might result in China having to make some hard choices, the US current account eating growth from the economy, growing global imbalances and entities with credit in dollars throughout the world finding it increasingly difficult to continue servicing their liabilities.
Unfortunately the pass-through of the FFR on the exchange seems to have become quite significant, especially compared to its pass-through on domestic demand.
I agree with you in general but I think that supports DeLong’s view. As long as the dollar remains the global reserve currency, and as long as national governments are discouraged from imposing capital controls, it’s very hard for most countries other than the US to sustain current account deficits, without risking balance of payments of payments crises. So in the interests of stable international payments it’s better for the US to run deficits, so the rest of the world can accumulate dollar reserves. So another argument for the high-interest, high-deficit path to full employment is that it implies larger current account deficits than the low-interest, low-deficit path.
the pass-through of the FFR on the exchange seems to have become quite significant
This does seem to be true. It’s a bit of a puzzle to me why cross-border financial flows are so much more sensitive to the policy rate than domestic flows.
To me it seems to be a combination of the effect on treasury yields (relative to other interest rates) and the knock on effects on LIBOR and other offshore dollar interest rates.
As someone with an engineering background, I would agree that the main functional changes in aircraft occurred earlier. The hardware improvements are leading to things like more fuel efficiency, but this would not be visible to people riding the planes.
The control systems (my engineering field) would be much improved. This would not be visible to outsiders, although pilot fatigue is reduced. This may have led to safety record improvements. That said, air travel was quite safe even by the 1960s.
More generally, my feeling is that to what extent technology is improving our lives, you need to look at advances in medicine. Or if you like watching kittens riding roombas, in which case the world is immeasurably better over the past decade.
Actually I would say medicine is one of the clearest examples for the technological slowdown argument. Although in that case the inflection point is perhaps a bit later. If you go back to 1900 and look for the leading causes of death, two of the top three are tuberculosis and gastrointestinal diseases. By 1955, both of these, along with diphtheria (also in the top ten in 1900) have essentially vanished. But if you look at the lists for 1955 and today, they’re basically the same. The big improvements in infant mortality also came in the first half of the century.
I’m afraid the case for accelerating technological change really comes down to the kittens on roombas.
Regarding the “singularity”, I didn’t like the idea because I think we don’t have a real way to quantify technological progress, but recently thinking about Ricardo’s concepts of “necessities” and “leisure goods” I came up with this idea: that technological progress leads to an increase in quantity of “necessities”, and in quality of “leisures”, however as productivity increases the share of the economy that is made of necessities falls, and the quantitative effects of technological progress fade.
In Ricardo’s theory there is a distinction between “necessities” and “leisure goods”, because worker’s wages are supposed to be at susistence level and subsistence depends only on “necessities”, whereas landlors were assumed to consume only “leisures”. A change in technology has a different effect if it happens in necessities or leisures (according to the LTV):
1) suppose that there are 2 goods: bread (necessities) and wine (leisures);
Suppose that the time of production of 1Kg of bread and of 1lt of wine is 1h, that corrsponds to 1$;
Suppose that workers need 3Kg of bread a day and work, on average, 5 hours a day.
Workers have a wage of 3$, and bread production makes up to 60% of total production, whereas wine production makes up for the remaining 40%.
The average worker produces 3Kg of bread and 2lt of wine every day.
2) suppose that in country A the productivity in terms of bread increases, so that now in 1h a workers produces 2Kg of bread.
Now the price of bread falls to 0.5$, and worker’s wages fall to 1.5$ (as they stay at subsistence level). But landlords drink much more wine, so that now the average worker produces (in 5 hours):
3Kg bread (1.5$)
3.5lt of wine (3.5$)
and bread represents now just 30% of the economy, whereas wine now is 70%. Basically the increase of productivity in bread shows up as an increase in wine producted.
3) on the other hand, suppose that in a nearby country B technology goes the other way, and it is wine making that becomes more productive so that now in 1h one worker produces 2lt. of wine. The wage stays the same because the subsistence level isn’t influenced by the change, but in country B the average worker produces, in 5h:
3Kg bread (3$)
4lt wine (2$)
(note the funny effect that a middle class guy who makes 5$ has an higer “real” income in country B, but is farther from subsistence, and therefore has an higer wage in ricardian terms, in country A).
Now, two addtional assumptions:
1) let’s ignore the level of wages and just assume that workers, if their wages are above subsistence, will also consume some leisure goods;
2) let’s assume that people don’t drink enormous quantities of wine but just scale up to more expensive wine.
Then, the effects of technologic improvements in “necessities” tend to be smaller and smaller with time, as necessities become a smaller and smaller share of the economy: if necessities are just 10% of the economy, and productivity doubles, they fall to 5%, but the amount of “other stuff” that becomes available is very small (from 90 to 95).
On the other hand, the effects of increased productivity in “leisures” are mostly in terms of quality, that is hard to quantify, for example if there is an improvements in productivity of movies special effects we will have better SF movies but this doesn’t really show up in productivity measures.
So the “singularity” might be past not because technology is now improving more slowly than in the past, but because the percentage of social production made of “necessities” is becoming increasingly small.
To make my point in a simple way:
suppose that we produce only necessary, subsistence goods, and take all the excess productivity as free time.
At time 1, it takes 10 hours to produce the minimum amount of stuff to carry on.
At time 2, productivity doubles, so people only have to work 5 hours a day, and gained 5h of free time.
At time 2, productivity doubles again, and people have to work only 2.5h a day; but now the net gain is only of 2.5 h of free time.
Each time productivity doubles, the gain is smaller, so even if productivity increases constantly, the amount of free time gained falls over time.
I’m quite sympathetic to your argument here. But in many specific cases we can measure technological progress in straightforward physical terms, e.g. aircraft speed. Of course we can’t aggregate across different technologies in physical terms, and that’s where the issues you raise become important. But if we see slowdowns in a broad enough range of technologies we may reasonably describe that as a slowdown in technological change in general, even if we can’t quantify it.