Tuesday, March 12, 2013

"Recession Is a Time of Harvest"

Noah and Seth say pretty much everything that needs to be said about this latest #Slatepitch provocation from Matt Yglesias.  [1] So, traa dy lioaur, I am going to say something that does not need to be said, but is possibly interesting.

Yglesias claims that "the left" is wrong to focus on efforts to increase workers' money incomes, because higher wages just mean higher prices. Real improvements in workers' living standards -- he says -- come from the same source as improvements for rich people, namely technological innovation. What matters is rising productivity, and a rise in productivity necessarily means a fall in (someone's) nominal income. So we need to forget about raising the incomes of particular people and trust the technological tide to lift all boats.

As Noah and Seth say, the logic here is broken in several places. Rising productivity in a particular sector can raise incomes in that sector as easily as reduce them. Changes in wages aren't always passed through to prices, they can also reflect changes in the distribution between wages and other income.

I agree, it's definitely wrong as a matter of principle to say that there's no link, or a negative link, between changes in nominal wages and changes in the real standard of living. But what kind of link is there, actually? What did our forebears think?

Keynes notoriously took the Yglesias line in the General Theory, arguing that real and nominal wages normally moved in opposite directions. He later retracted this view, the only major error he conceded in the GT (which makes it a bit unfortunate that it's also the book's first substantive claim.) Schumpeter made a similar argument in Business Cycles, suggesting that the most rapid "progress in the standard of life of the working classes" came in periods of deflation, like 1873-1897. Marx on the other hand generally assumed that the wage was set in real terms, so as a first approximation we should expect higher productivity in wage-good industries to lead to lower money wages, and leave workers' real standard of living unchanged. Productivity in this framework (and in post-GT Keynes) does set a ceiling on wages, but actual wages are almost well below this, with their level set by social norms and the relative power of workers and employers.

But back to Schumpeter and the earlier Keynes. It's worth taking a moment to think through why they thought there would be a negative relationship between nominal and real wages, to get a better understanding of when we might expect such a relationship.

For Keynes, the logic is simple. Wages are equal to marginal product. Output is produced in conditions of declining marginal returns. (Both of assumptions are wrong, as he conceded in the 1939 article.) So when employment is high, the real wage must be low. Nominal wages and prices generally move proportionately, however, rising in booms and falling in slumps. (This part is right.) So we should expect a move toward higher employment to be associated with rising nominal wages, even though real wages must fall. You still hear this exact argument from people like David Glasner.

Schumpeter's argument is more interesting. His starting point is that new investment is not generally financed out of savings, but by purchasing power newly created by banks. Innovations are almost never carried out by incumbent producers simply adopting the new process in place of the existing one, but rather by some new entrant -- the famous entrepreneur-- operating with borrowed funds. This means that the entrepreneur must bid away labor and other inputs from their current uses (importantly, Schumpeter assumes full employment) pushing up costs and prices. Furthermore, there will be some extended period of demand from the new entrants for labor and intermediate goods while the incumbents have not yet reduced theirs -- the initial period of investment in the new process (and various ancillary processes -- Schumpeter is thinking especially of major innovations like railroads, which will increase demand in a whole range of related industries), and later periods where the new entrants are producing but don't yet have a decisive cost advantage, and a further period where the incumbents are operating at a loss before they finally exit. So major innovations tend to involve extended periods of rising prices. It's only once the new producers have thoroughly displaced the old ones that demand and prices fall back to their old level. But it's also only then that the gains from the innovation are fully realized. As he puts it (page 148):
Times of innovation are times of effort and sacrifice, of work for the future, while the harvest comes after... ; and that the harvest is gathered under recessive symptoms and with more anxiety than rejoicing ... does not alter the principle. Recession [is] a time of harvesting the results of preceding innovation...
I don't think Schumpeter was wrong when he wrote. There is probably some truth to idea that falling prices and real wages went together in 19th century. (Maybe by 1939, he was wrong.)

I'm interested in Schumpeter's story, though, as more than just intellectual history, fascinating tho that is. Todays consensus says that technology determines the long-term path of the economy, aggregate demand determines cyclical deviations from that path, and never the twain shall meet. But that's not the only possibility. We talked the other day about demand dynamics not as -- as in conventional theory -- deviations from the growth path in response to exogenous shocks, but as an endogenous process that may, or may not, occasionally converge to a long-term growth trajectory, which it also affects.

In those Harrod-type models, investment is simply required for higher output -- there's no innovation or autonomous investment booms. Those are where Schumpeter comes in. What I like about his vision is it makes it clear that periods of major innovation, major shifts from one production process to another, are associated with higher demand -- the major new plant and equipment they require, the reorganization of the spatial and social organization of production they entail ("new plant, new firms, new men," as he says) make large additional claims on society's resources. This is the opposite of the "great recalculation" claim we were hearing a couple years ago, about how high unemployment was a necessary accompaniment to major geographic or sectoral shifts in output; and also of the more sophisticated version of the recalculation argument that Joe Stiglitz has been developing. [2] Schumpeter is right, I think, when he explicitly says that if we really were dealing with "recalculation" by a socialist planner, then yes, we might see labor and resources withdrawn from the old industries first, and only then deployed to the new ones. But under capitalism things don't work like that  (page 110-111):
Since the central authority of the socialist state controls all existing means of production, all it has to do in case it decides to set up new production functions is simply to issue orders to those in charge of the productive functions to withdraw part of them from the employments in which they are engaged, and to apply the quantities so withdrawn to the new purposes envisaged. We may think of a kind of Gosplan as an illustration. In capitalist society the means of production required must also be ... [redirected] but, being privately owned, they must be bought in their respective markets. The issue to the entrepreneurs of new means of payments created ad hoc [by banks] is ... what corresponds in capitalist society to the order issued by the central bureau in the socialist state. 
In both cases, the carrying into effect of an innovation involves, not primarily an increase in existing factors of production, but the shifting of existing factors from old to new uses. There is, however, this difference between the two methods of shifting the factors : in the case of the socialist community the new order to those in charge of the factors cancels the old one. If innovation were financed by savings, the capitalist method would be analogous... But if innovation is financed by credit creation, the shifting of the factors is effected not by the withdrawal of funds—"canceling the old order"—from the old firms, but by ... newly created funds put at the disposal of entrepreneurs : the new "order to the factors" comes, as it were, on top of the old one, which is not thereby canceled.
This vision of banks as capitalist Gosplan, but with the limitation that they can only give orders for new production on top of existing production, seems right to me. It might have been written precisely as a rebuttal to the "recalculation" arguments, which explicitly imagined capitalist investment as being guided by a central planner. It's also a corrective to the story implied in the Slate piece, where one day there are people driving taxis and the next day there's a fleet of automated cars. [3] Before that can happen, there's a long period of research, investment, development -- engineers are getting paid, the technology is getting designed and tested and marketed, plants are being built and equipment installed -- before the first taxi driver loses a dollar of income. And even once the driverless cars come on line, many of the new companies will fail, and many of the old drivers will hold on for as long as their credit lasts. Both sets of loss-making enterprises have high expenditure relative to their income, which by definition boosts aggregate demand. In short, a period of major innovations must be a period of rising nominal incomes -- as we most recently saw, on a moderate scale, in the late 1990s.

Now for Schumpeter, this was symmetrical: High demand and rising prices in the boom were balanced by falling prices in the recession -- the "harvest" of the fruits of innovation. And it was in the recession that real wages rose. This was related to his assumption that the excess demand from the entrepreneurs mainly bid up the price of the fixed stock of factors of production, rather than activating un- or underused factors. Today, of course, deflation is extremely rare (and catastrophic), and output and employment vary more over the cycle than wages and prices do. And there is a basic asymmetry between the boom and the bust. New capital can be added very rapidly in growing enterprises, in principle; but gross investment in the declining incumbents cannot fall below zero. So aggregate investment will always be highest when there are large shifts taking place between sectors or processes. Add to this that new industries will take time to develop the corespective market structure that protects firms in capital-intensive industries from cutthroat competition, so they are more likely to see "excess" investment. And in a Keynesian world, the incomes from innovation-driven investment will also boost consumption, and investment in other sectors. So major innovations are likely to be associated with booms, with rising prices and real wages.

So, but: Why do we care what Schumpeter thought 75 years ago, especially if we think half of it no longer holds? Well, it's always interesting to see how much today's debates rehash the classics. More importantly, Schumpeter is one of the few economists to have focused on the relation of innovation, finance and aggregate demand (even if, like a good Wicksellian, he thought the latter was important only for the price level); so working through his arguments is a useful exercise if we want to think more systematically about this stuff ourselves. I realize that as a response to Matt Y.'s silly piece, this post is both too much and poorly aimed. But as I say, Seth and Noah have done what's needed there. I'm more interested in what relationship we think does hold, between innovation and growth, the price level, and wages.

As an economist, my objection to the Yglesias column (and to stuff like the Stiglitz paper, which it's a kind of bowdlerization of) is that the intuition that connects rising real incomes to falling nominal incomes is just wrong, for the reasons sketched out above. But shouldn't we also say something on behalf of "the left" about the substantive issue? OK, then: It's about distribution. You might say that the functional distribution is more or less stable in practice. But if that's true at all, it's only over the very long run, it certainly isn't in the short or medium run -- as Seth points out, the share of wages in the US is distinctly lower than it was 25 years ago. And even to the extent it is true, it's only because workers (and, yes, the left) insist that nominal wages rise. Yglesias here sounds a bit like the anti-environmentalists who argue that the fact that rivers are cleaner now than when the Clean Water Act was passed, shows it was never needed. More fundamentally, as a leftist, I don't agree with Yglesias that the only important thing about income is the basket of stuff it procures. There's overwhelming reason -- both first-principle and empirical -- to believe that in advanced countries, relative income, and the power, status and security it conveys, is vastly more important than absolute income. "Don't worry about conflicting interests or who wins or loses, just let the experts make things better for everyone": It's an uncharitable reading of the spirit behind this post, but is it an entirely wrong one?

UPDATE: On the other hand. In his essay on machine-breaking, Eric Hobsbawm observes that in 18th century England,
miners' riots were still directed against high food prices, and the profiteers believed to be responsible for them.
And of course more generally, there have been plenty of working-class protests and left political programs aimed at reducing the cost of living, as well as raising wages. Food riots are a major form of popular protest historically, subsidies for food and other necessities are a staple policy of newly independent states in the third world (and, I suspect, also disapproved by the gentlemen of Slate), and food prices are a preoccupation of plenty of smart people on the left. (Not to mention people like this guy.) So Yglesias's notion that "the left" ignores this stuff is stupid. But if we get past the polemics, there is an interesting question here, which is why mass politics based around people's common interests as workers is so much more widespread and effective than this kind of politics around the cost of living. Or, maybe better, why one kind of conflict is salient in some times and places and the other kind of conflict in others; and of course in some, both.

[1] Seth's piece in particular is a really masterful bit of polemic. He apologizes for responding to trollery, which, yeah, the Yglesias post arguably is. But if you must feed trolls, this is how it's done. I'm not sure if the metaphor requires filet mignon and black caviar, or dogshit garnished with cigarette butts, or fresh human babies, but whatever it is you should ideally feed a troll, Seth serves it up.

[2] It appears that Stiglitz's coauthor Bruce Greenwald came up with this first, and it was adopted by right-wing libertarians like Arnold Kling afterwards.

[3] I admit I'm rather skeptical about the prospects for driverless cars. Partly it's that the point is they can operate with much smaller error tolerances than existing cars -- "bumper to bumper at highway speeds" is the line you always hear -- but no matter how inherently reliable the technology, these things are going to be owned maintained by millions of individual nonprofessionals. Imagine a train where the passengers in each car were responsible for making sure it was securely coupled to the next one. Yeah, no. But I think there's an even more profound reason, connected to the kinds of risk we will and will not tolerate. I was talking to my friend E. about this a while back, and she said something interesting: "People will never accept it, because no one is responsible for an accident. Right now, if  there's a bad accident you can deal with it by figuring out who's at fault. But if there were no one you could blame, no one you could punish, if it were just something that happened -- no one would put up with that." I think that's right. I think that's one reason we're much more tolerant of car accidents than plane accidents, there's a sense that in a car accident at least one of the people involved must be morally responsible. Totally unrelated to this post, but it's a topic I'd like to return to at some point -- that what moral agency really means, is a social convention that we treat causal chains as being broken at certain points -- that in some contexts we treat people's actions as absolutely indeterminate. That there are some kinds of in principle predictable actions by other people that we act -- that we are morally obliged to act -- as if we cannot predict.


  1. PHENOMENAL blog post! One question, though. You wrote:

    "a period of major innovations must be a period of rising nominal incomes -- as we most recently saw, on a moderate scale, in the late 1990s."

    Imagine a situation where we had national work sharing. As a result, rapid NGDP growth is not required in order to sustain low unemployment.

    That scenario could also result in major innovations, right? So isn't it more appropriate to say that "a period of major innovations must be a period of low unemployment"?

    1. Perhaps I'm missing something, but I don't see why that matters. Schumpeter's point is just that Ina decentralized capitalist economy, resources get allocated to new forms of production without being withdrawn from old ones.

    2. Also, my one request to commenters here is that they not post as "Anonymous." Would you mind using a handle of some kind? Thanks!

  2. Good post.

    But I think driverless cars are coming. Safety-related maintenance won’t be much of an issue. Right now we trust slapdash individual owners to keep their brakes in working order, and those are just as crucial to preventing crashes as driving-computers would be (and more prone to breakdown because they suffer intense mechanical stresses.) Bumper-to-bumper at 75 mph is maybe a little extreme, but we will just legislate speed vs. spacing rules and program those in.

    I hate the idea of driverless cars too—just another step in the machine-made atrophy of human competence and initiative. (Not to mention the mass unemployment for blue-collar professional drivers.) But E. is wrong to argue that people will reject it because of the implied abdication of individual moral agency to the impersonal traffic system. Driving is already a responsibility-free zone where individual fault is very rarely an issue. 35,000 people a year die in car crashes and few drivers ever go to jail for those vehicular homicides, unless they violate certain procedural regulations—mainly, the “no alcohol, no foul” rule. A driver who is simply inattentive and kills a child who darts into the street will usually face no criminal or even civil penalty. Society recognizes that a huge body count is a systemic feature of high-speed motoring and tacitly absolves individuals of blame for the resulting carnage.

  3. Interesting post; of course, anything that involves Schumpter interests me.

    Just as a matter of intellectual history, Schumpter's concept of the business cycle, as cycles within cycles (of varying lengths and drivers), is quite different from more modern and pedestrian ones. I'm not sure how his ideas correspond with those of W.C. Mitchell and the NBER, and, of course, he did not live to see the post-WWII pattern of recessions triggered by deliberate monetary policy inverting the yield curve. Schumpeter, as I recall, seemed to think that his longer cycles, driven by the big innovations like railroads or electricity, could impart lengthy seasons to business conditions, as the innovation played out during a long S-curve of diffusion, during which the price/cost pressure on established allocations of resources varied in character and intensity.

    The big, dramatic crises, booms and busts, in the Schumpeterian vision were the product of shorter and longer cycles coinciding, but against a background effect imparted by the "seasons" of long, technological cycles.

    The 19th century was not, generally speaking, an entirely happy period for wage-earners, and the emergence of boom and bust business cycles to rival and eclipse the misery of harvest failures, something of a mystery. I don't know that there are good statistics for early and mid-century, but, qualitatively, it often looks as if wages are being driven by immiserating trade with the capitalists and their machines. How much of that immiseration is the elimination of economic rents for skills and how much is the side-effect of taking income out of consumption and dedicating it to capital accumulation, I couldn't say.

    I do know that the huge expansion of U.S. output during the Civil War did not set off the kind of runaway boom seen in the 20th century's world wars. Real wages did not seem to rise appreciably, despite the diversion of a significant part of the labor force into the military, and the vast purchase of agricultural and manufactured output. It was definitely good times in the North, as manufactured goods displaced homespun and increasingly mechanized farms exported grain to offset harvest failures in central Europe, but wages showed little inflationary potential.

    The Long Depression, 1871-1896, may well have influenced Schumpeter's idea of long cycle seasons. The common driver of the Long Depression across the industrializing countries, is the deflationary effect of rapid expansion of production output against gold standard currencies, at a time when gold production lagged. For Great Britain, the Long Depression was the period during which the economic rents of the Workshop of the World in the First Industrial Revolution of iron, steam, coal and textiles, were being eroded by the Second Industrial Revolution in the U.S., France, Germany and central Europe, in steel, petroleum, electricity and chemicals.

    In the U.S., (real) wages were driven down during the depression that followed the Panic of 1873 -- a record five-years of business contraction in the NBER cycle record -- and recovered only partially in the booms of the 1880s, but lost again in the depression of the 1890s. (The union won at Homestead in 1882, but lost in 1892.) This was still a period in which agriculture in the U.S. was expanding, drawing in more land and more labor, and a farmer could hope to earn a median or better living, but many farmers, especially in places reacheable only by railroad, found themselves under intense financial pressure.

    The deflations and consequent financial panics seemed to have the effect of concentrating the ownership of capital, exacerbating the extreme concentration of wealth and income in the Gilded Age. In the two-recession Depression, which followed the Panic of 1893, every railroad west of the Mississippi, save one, went bankrupt and changed ownership, and something like 10,000 banks failed.

  4. Some analytical points.

    Rising productivity from innovation tends to diffuse outward from the innovating industry or sector. This is not a case of virtue rewarded. A sector in which input productivity is rising rapidly is also likely a sector in which (output prices are falling, so marginal revenue product of labor is very likely to fall as a result of increasing (sector) labor productivity.

    The sectors, which will see an increase in wages from rising productivity, may very well be the sectors in which there are less than average increases in productivity.

    A general rise in factor productivity implies a general rise in factor cost. A sectoral rise in factor productivity implies . . . well, hard to say.

    Substitution between labor and capital implies that increases in the invested capital stock may boost marginal productivity of labor and wages.

    An increase in sectoral factor input prices relative to general factor prices, implies that the sector needs to increase its share of that input; higher (relative) wages in a sector implies that that sector needs more labor.

    It is certainly possible to imagine rising productivity in a sector resulting in a need to shed factors, as well as to attract factors, depending upon the elasticity of output demand. In the Great Depression in the U.S., productivity was rising very rapidly in both manufacturing sectors, where continuous process methods (assembly lines) could be developed, such as auto manufacturing or electrical products or food processing, and in farming.

    The rise in productivity in farming required that output prices and revenues be stabilized in such a way that large capital investments could be made, and the sector would still shed labor. Shedding labor was actually a requirement for stabilizing farm wages and incomes.

    The rise in productivity in manufacturing seems to have required an increase in wages, both relative to the median wage, to attract additional labor, but also, relative to capital.

    Continuous process manufacturing was increasing total factor productivity substantially. Normalized against unit of output, an auto assembly plant uses less labor and less capital per unit of output, but the capital is amplifying the output of labor. In common parlance, mass wages had to rise, if output from mass production was to be sold.

    1. A lot here to think about, as always, Bruce.

      Rising productivity from innovation tends to diffuse outward from the innovating industry or sector. This is not a case of virtue rewarded. A sector in which input productivity is rising rapidly is also likely a sector in which (output prices are falling

      This is an important point, but I think the most important thing about it is that it sometimes holds and sometimes does not. Some important historical cases where it did hold include proto-industrialization outside Europe, agriculture in the US, industry in the early USSR (the famous "scissors" problem of rising relative agriculture prices), and primary product exports through much of modern history. But I think there are other cases where the gains from rising productivity are substantially captured in the sector where they happen.

      Substitution between labor and capital implies that increases in the invested capital stock may boost marginal productivity of labor and wages.

      Yes. Careful with that last "and" tho!

      It is certainly possible to imagine rising productivity in a sector resulting in a need to shed factors, as well as to attract factors, depending upon the elasticity of output demand.

      Yes. This is one of the main points of Noah Smith's post that I linked to.

      The rise in productivity in manufacturing seems to have required an increase in wages, both relative to the median wage, to attract additional labor, but also, relative to capital.

      Yes, I maybe should have highlighted this point more. MY kind of gives the game away when he moves from the claim that somebody's income has to fall with innovation, to the claim that that somebody is labor. He's got the model in his head that labor is highly specialized to complement certain technologies, but of course that isn't the case. Logically, the punchline of his post would better have been that since technology and investment boost the capital stock and labor is relatively fixed, the natural result of higher growth is a higher labor share. The problem is that owners of aren't willing to accept the natural fall in their share that goes with faster growth. So the important thing -- if we want all that great new stuff that technology promises -- is not to stop complaining about nominal wages *falling*, but rather to stop complaining about them *rising* -- which of course is not an error of some dudes at Occupy but of the most powerful economic policymakers. The natural logic of MY's argument should lead him to say that wage growth should not be treated as necessarily inflationary, and to the extent that it is we should target firms' markups rather than demand, or just accept higher inflation as the price of progress.

      To be fair, I think he has made some steps in this direction in the past.

  5. Another great post, Josh. Re: the footnote on driverless cars - As a determined non-driver I hope they do arrive. But I do think of Kim Stanley Robinson's 'The Gold Coast', set in an LA with fully automated traffic. Well, almost fully-automated - one of the main characters is a paramedic whose job it is to pull bodies from the wreckage of daily malfunctions.

  6. Where are you going to get the gas for cars, with or without drivers?

    Take the train, fool.

  7. I had to stop reading streetsblog because it was just too depressing. Every other day somebody in NYC is murdered with a car and the cops generally blame the victim. Near Seattle a teenager got a $42 ticket for killing a cyclist on the side of the road. It's hard to imagine how driverless cars could be worse than human-controlled cars. Most people just aren't suited for piloting a two-ton killing machine. The same people who are too lazy to get up of the couch to change the channel burst into rage if they have to lift their foot two inches to avoid killing somebody. I could go on, but I'm already ranting. The big advantage of driverless cars will be their inhumanity, their ability to drive an SUV at 20mph instead of 40mph.

    I think the way it'll play out is old folks and teenagers will be given the choice of driverless cars or carlessness, and choose the former.

  8. Most people just aren't suited for piloting a two-ton killing machine

    Yeah, tell me about it. Just Monday, a van jumped the curb and ran a bunch of people down at LaGuardia Community College, where my kid's mother teaches.

  9. "New capital can be added very rapidly in growing enterprises, in principle; but gross investment in the declining incumbents cannot fall below zero. So aggregate investment will always be highest when there are large shifts taking place between sectors or processes."

    I don't think this is correct: it depends wether you are speaking of capital as "cost of production" or as "market value".

    For example, suppose that automaker A builds a car factory, for a cost of 100; this factory produces enough cars that each year A gets a net profit of 5.
    But some time later, automaker B builds a factory with a new technology; this new factory still costs 100, but produces more cars so that B's net profits are 10/year.
    A's investiment looked good when A did it, but now is underperforming. However, A cannot immediately disinvest his 100.
    But the market value of A's factory is no more 100: no one would be so stupid to pay 100 for a factory that provides the same profits that one could have investing just 50 in the new technology; as a consequence, the market value of A's factory is now 50 (much below it's production cost).

    So the question is, when you speak of aggregate investiment, do you mean production cost (that would be A+B = 100 + 100 = 200) or market value (A+B = 50 +100 = 150)?

    Since the lower market value of A's capital good limits the amount of money that A can borrow (and hence the quantity of money that is created by A's non-disinvestiment) I think that the aggregate market value of capital goods in a time of big tecnological shift will be lower than usual (since most capital is already existing and will be devalued).

    1. But it's cost that matters here. The original question is, do we expect innovation to be associated with rising or falling wages (and other nominal incomes)?

      Take your case: Prior to B's factory coming on line, there are substantial costs, both for the factory itself and -- especially in technologically leading industries the R&D and related expenses that allow the new plant to have a cost advantage in the first place. All during the time when those costs are being incurred, A's plant is operating as normal. So during that period, the total demand for labor and other inputs is higher than in "normal" times. And even after B's plant comes online, A will not exit immediately, so there will be some period of "overcapacity" in which total costs incurred are excessive relative to final demand for the industry's output, which again means higher than normal demand for labor. This is the process I'm talking about.

      Now it is true, as you say, that A's capital loses value in money terms before it is withdrawn from production. And it may well be true (tho it may not be) that A's capital loss is initially greater than B's capital gain. But I think it would be a big mistake to assume that this translates into a fall in demand.

      First, because the degree to which losses bind current output depends entirely on the flexibility of owners with respect to dividends and other payouts, and the willingness of the firm's lenders to continue extending credit; and both of these factors vary a lot with the specific capitalism we are talking about. The overvalued dollar in the first of the 1980s and the overvalued yen in the second half of the 1980s mean very large losses for US and Japanese exporters; but loss-making American manufacturers shut down much more readily than loss-making Japanese manufacturers did.

      Second, because in situations like this, losses at firms like A do not necessarily reduce desired investment. In the same Hobsbawm essay I quote in the update, he says: "Most capitalists took the new machine in the first instance not as an offensive weapon, to win bigger profits, but as a defensive one, to protect themselves against the bankruptcy which threatened the laggard competitor." This idea of "coerced investment, originally of course from Marx, has been thoroughly developed by my teacher Jim Crotty. In general, I think it's quite plausible that innovation and high investment will be associated with low profits and large losses for many firms -- I think it's almost inevitable -- while regimes of stable, corespective competition will see high profits and low investment.

    2. So just to be clear -- this is an argument for why innovation and technological change is normally (almost always, at least in existing capitalist economies) associated with rising nominal wages and output. Which I think is really important to get straight if we want to think clearly about booms and slumps.

    3. But, wouldn't it be simpler to say that, when an innovation happens, some capitalists have higer than normal profits, and this lead to irrationally high expectations, that in turn lead to an high demand for labor, that in the end results in higer wages both in relative and nominal terms? Inflation happens because formerly unemployed workers start buying goods whose supply didn't increase IMHO, not because there is a shortage of workers (eg. Inflation in food in China today .

    4. No, I don't think that profits are necessarily higher in the aggregate, and I definitely don't think that irrational expectations by capitalists are required. Given that new technologies are embodied in new capital, if the new process is going to replace the old one at a faster rate than the depreciation rate of the old capital, you will see a higher overall level of investment demand.

      The inflation question is separate. You are right that in China and other labor-surplus countries, high demand does not necessarily translate into higher real wages. This is probably one of the conditions for sustained very high growth rates in China today, and in Japan and other Asian countries in earlier decades. This was also true of the US in the 19th century. But in countries like the contemporary US, without a deep reserve of rural labor to draw on, I don't think there's any question that booms do see labor shortages and wages to at least some extent pushing up prices.

  10. "There's overwhelming reason -- both first-principle and empirical -- to believe that in advanced countries, relative income, and the power, status and security it conveys, is vastly more important than absolute income."

    I tried to make this point the other day at Crooked Timber, though I think it was lost in the comments. There is a quote from Marx to this effect that I like to trot out*. It's an overwhelmingly obvious point, whether you look at international comparisons or anecdotal evidence about the status rivalries of the very wealthy or whatever. But it doesn't stop there from being an endless number of arguments to the effect that "poverty is not a problem because Louis XIV didn't have an iPad and I saw a poor person with an iPad." What Yglesias said was not quite that glib, but amounted to a futilistic expression of the same idea ("you can't change relative wealth; so change absolute wealth!".

    * "A house may be large or small; as long as the neighboring houses are likewise small, it satisfies all social requirement for a residence. But let there arise next to the little house a palace, and the little house shrinks to a hut. The little house now makes it clear that its inmate has no social position at all to maintain, or but a very insignificant one; and however high it may shoot up in the course of civilization, if the neighboring palace rises in equal or even in greater measure, the occupant of the relatively little house will always find himself more uncomfortable, more dissatisfied, more cramped within his four walls.

    "An appreciable rise in wages presupposes a rapid growth of productive capital. Rapid growth of productive capital calls forth just as rapid a growth of wealth, of luxury, of social needs and social pleasures. Therefore, although the pleasures of the labourer have increased, the social gratification which they afford has fallen in comparison with the increased pleasures of the capitalist, which are inaccessible to the worker, in comparison with the stage of development of society in general. Our wants and pleasures have their origin in society; we therefore measure them in relation to society; we do not measure them in relation to the objects which serve for their gratification. Since they are of a social nature, they are of a relative nature." (Wage-labor and Capital, Ch. 6)

    1. Yes exactly. I'd completely forgotten that passage.

      Capital is such an amazingly rich book, isn't it? It's almost always worth going back to -- so many of our current debates you can find already addressed there.

    2. Yeah. The only issue with it is that it presupposes a certain familiarity with Aristotle, Hegel, and the rhetoric and arguments of the classical English political economists, without saying as much up front. The humorous intent of some of the language can be unclear to the student who is not fortunate enough to have covered this ground.

  11. Very interesting. Though I think I could have anticipated the "substantive" take-away.

    At the risk of sounding unappreciative of the post/blog, I do have to say I find the econ-jargon heavy going (e.g. "marginal product" and that's as someone who has a lot of math and social science in my background but only a few semester of economics). Despite that, I do think these discussions are important both within the Left and across social science disciplines. (Maybe for my benefit and those of others I will work on a glossary...)

    1. Fair point. This blog has always been mainly a virtual notebook for me to write down stuff I happen to be thinking about, but as it's gradually acquired more readers, it's worth thinking about ways to make it more accessible and less jargonny.