Wednesday, August 31, 2011

Are Recessions All About Money?

There is a view that seems to be hegemonic among liberal economists, that recessions are fundamentally about money or finance. Not just causally, not just in general, but always, by definition. In this view, the only sense in which one can speak about aggregate demand as a constraint on output, is if we can identify excess demand for some non-produced financial asset.

In the simplest case, people want to hold a stock of money in some proportion to their total income. Money is produced only by the government. Now suppose people's demand for money rises, and the government fails to increase supply accordingly. You might expect the price of money to rise -- that is, deflation. But deflation doesn't restore equilibrium, either because prices are sticky (i.e., deflation can't happen, or not fast enough), or because deflation itself further raises the demand for money. It might do this by raising precautionary demand, since falling prices make it likely that businesses and households won't be able to meet obligations fixed in money terms and will face bankruptcy (Irving Fisher's debt-deflation cycle). Or deflation might increase demand for money by because it redistributes income from net borrowers to net savers, and the latter have a higher marginal demand for money holdings. Or there could be other reasons. In any case, the price of money doesn't adjust, so government has to keep its quantity growing at the appropriate rate instead. From this perspective,  if we ever see an economy operating bellow full capacity, it is true by definition that there is excess demand for some money-like asset.

This sounds like Milton Friedman. It is Milton Friedman! But it also seems to be most of the liberal macroeconomists who are usually called Keynesians. Here's DeLong:
there was indeed a "general glut" of newly-produced commodities for sale and of workers to hire. But it was also the case that the excess supply of goods, services, and labor was balanced by an excess demand elsewhere in the economy. The excess demand was an excess demand not for any newly-produced commodity, but instead an excess demand for financial assets, for "money"...

How, exactly, should economists characterize the excess demand in financial markets? Where was it, exactly? That became a subject of running dispute, and the dispute has been running for more than 150 years, with different economists placing the cause of the "general glut" that was excess supply of newly-produced goods and of labor at the door of different parts of the financial system.

The contestants are:

Fisher-Friedman: monetarism: a depression is the result of an excess demand for money--for those liquid assets generally accepted as means of payment that people hold in their portfolios to grease their market transactions. You fix a depression by having the central bank boost the money stock...

Wicksell-Keynes (Keynes of the Treatise on Money, that is): a depression happens when there is an excess demand for bonds... You fix a depression by either reducing the market rate of interest (via expansionary monetary policy) or raising the natural rate of interest (via expansionary fiscal policy) in order to bring them back into equality....

Bagehot-Minsky-Kindleberger: a depression happens because of a panic and a flight to quality, as everybody tries to sell their risky assets and cuts back on their spending in order to try to shift their portfolio in the direction of safe, high-quality assets... You fix a depression by restoring market confidence and so shrinking demand for AAA assets and by increasing the supply of AAA assets....

From the perspective of this Malthus-Say-Mill framework Keynes's General Theory is a not entirely consistent mixture of (1), (2), and (3)...Note that these financial market excess demands can have any of a wide variety of causes: episodes of irrational panic, the restoration of realistic expectations after a period of irrational exuberance, bad news about future profits and technology, bad news about the solvency of government or of private corporations, bad government policy that inappropriately shrinks asset stocks, et cetera. Nevertheless, in this Malthus-Say-Mill framework it seems as if there is always or almost always something that the government can do to affect asset supplies and demands that promises a welfare improvement
That's an admirably clear statement. But is it right? I mean, first, is it right that demand constraints can always and only be usefully characterized as excess demand for some financial asset? And second, is that really what the General Theory says?

The first answer is No. Or rather, it's true but misleading. It is hard to talk sensibly about a "general glut" of currently produced goods except in terms of an excess demand for some money-like financial asset. But recessions and depressions are not mainly characterized by a glut of currently produced goods. They are characterized by an excess of productive capacity. Markets for all currently-produced goods may clear. But there is still a demand constraint, in the sense that if desired expenditure were higher, aggregate output would be higher. The simple Keynesian cross we teach in the second week of undergrad macro is a model of just such an economy, which makes sense without money or any other financial asset. (And is probably more useful than most of what gets taught in graduate courses.) Arguably, this is the normal state of modern capitalist economies.

I'll come back to this in a future post, hopefully. But it's important to stress that the notion of aggregate demand limiting output, does not imply that any currently-produced good is in excess supply. [1]

Meanwhile, how about the second question -- in the General Theory, did Keynes see demand constraints as being fundamentally about excess demand for money or some other financial asset, with the solution being to change the relative price of currently produced goods, and that asset? Again, the answer is No.

In his explanation of the instability of capitalist economies, Keynes always emphasizes the fluctuations in investment demand (or in his terms, the marginal efficiency of capital schedule). Investment demand is based on the expected returns of new capital goods over their lifetime. But the distribution of future states of the world relevant to those returns is not just stochastic but fundamentally unknown, so expectations about profits on long-lived fixed capital are essentially conventional and unanchored. It is these fluctuations in expectations, and not the demand for financial assets as expressed in liquidity preference, that drives booms and slumps. Keynes:
The fact that a collapse in the marginal efficiency of capital tends to be associated with a rise in the rate of interest may seriously aggravate the decline in investment. But the essence of the situation is to be found, nevertheless, in the collapse of the marginal efficiency of capital... Liquidity preference, except those manifestations which are associated with increasing trade and speculation, does not increase until after the collapse in the marginal efficiency of capital.  
It is this, indeed, which renders the slump so intractable. Later on, a decline in the rate of interest will be a great aid to recovery and probably a necessary condition of it. But for the moment, the collapse in the marginal efficiency of capital may be so complete that no practicable reduction in the rate of interest will be enough [to offset it]. If the reduction in the rate of interest was capable of proving an effective remedy by itself, it might be possible to achieve a recovery without the elapse of any considerable interval of time and by means more or less directly under the control of the monetary authority. But in fact, this is not usually the case.
In this sense, Keynes agrees with the Real Business Cycle theorists that the cause of a decline in output is not fundamentally located in the financial system, but a fall in the expected profitability of new investment. The difference is that RBC thinks a decline in expected profitability must be due to genuine new information about the true value of future profits. Keynes on the other hand thinks there is no true expected value in that sense, and that our belief about the future are basically irrational. ("Enterprise only pretends to itself to be actuated by the statements in its prospectus ... only a little more than an expedition to the South Pole, is it based on an exact calculation of benefits to come.") This is an important difference. But the key point here is the bolded sentences. Keynes considers DeLong's view that the fundamental cause of a downturn is an autonomous increase in demand for safe or liquid assets, and explicitly rejects it.

The other thing to recognize is that Keynes never mentions the zero lower bound. He describes the liquidity trap as theoretical floor of the interest rate, which is above zero, but nothing in his argument depends on it. Rather, he says,
The most stable, and least easily shifted, element in our contemporary economy has been hitherto, and may prove to be in the future, the minimum rate of interest acceptable to the generality of wealthowners. (Cf. the nineteenth-century saying quoted by Bagehot, that "John Bull can stand many things, but he cannot stand 2 percent.") If a tolerable level of employment requires a rate of interest much below the average rates which ruled in the nineteenth century, it is most doubtful whether it can be achieved merely by manipulating the quantity of money.
This is an important part of the argument, but it tends to get ignored by mainstream Keynesians, who assume that monetary authority can reliably set "the" interest rate. But as we see clearly today, this is not a good assumption to make. Well before the policy rate reached zero, it had become effectively disconnected from the rates facing business borrowers. And of course the hurdle rate from the point of view of the decisionmakers at a firm considering new investment isn't just the market interest rate, but that rate plus some additional premium reflecting what Keynes (and later Minsky) calls borrower's risk.

So, Keynes thought that investment demand was subject to wide, unpredictable fluctuations, and probably also a secular downward trend. He doubted that very large movements in the interest rate could be achieved by monetary policy. And he didn't think that the moderate movements that could be achieved, would have much effect on investment. [2] Where did that leave him? "Somewhat skeptical of the success of a merely monetary policy directed toward influencing the rate of interest" at stabilizing output and employment; instead, the government must "take an ever greater responsibility for directly organizing investment."

Of course, DeLong could be misrepresenting Keynes and still be right about economic reality. But we need to at least recognize that aggregate demand is logically separate from the idea of a general glut; that the former, unlike the latter, does not necessarily involve excess demand for any financial asset; and that in practice supply and demand conditions in financial markets are not always the most important or reliable influences on aggregate demand. Keynes, at least, didn't think so. And he was a smart guy.

[1] The other point, to anticipate a possible objection, is that the investment decision does not involve allocation of a fixed stock of savings between capital goods and financial assets.

[2] The undoubted effectiveness of monetary policy in the postwar decades might seem to argue against this point. But it's important to recognize -- though Keynes himself didn't anticipate this -- that in practice monetary policy has operated largely though its effect on the housing market, not on investment.


  1. As you know, Schumpeter didn't think so either.

  2. I think this view of Keynesianism settled in pretty early. The view is already recognizable in Modigliani's 1948 paper. Patinkin's 1965 book codifies the view that Keynesianism requires sticky prices.

  3. 5371-

    Right. In that first footnote, I was thinking specifically of Schumpeter's point (echoed by Minsky) that investment is normally financed by the creation of new money through the banking system. This new money either mobilizes idle resources (in a Keynesian story) or bids away resources from other uses via inflationary forced savings (in Schumpeter's story.) But either way, an increase in investment doesn't require anyone to affirmatively decide to reduce current consumption, or to affirmatively decide to reduce their holdings of safe financial assets. Keynes is very clear on the first point but less clear on the second, IMO. Now it is true that banks' willingness to hold new liabilities (and also probably of entrepreneurs to issue them) will depend to some extent on the availability of safe assets to serve as reserves. So to this extent the supply of these assets acts as a constraint on investment, and this is one of the margins monetary policy has traditionally operated on. (Most directly in the old days when the statutory reserve requirement on bank liabilities was effective.) But the strength of this link is going to depend on a lot of specific institutional factors, it doesn't justify equating a fall in investment demand with an increase in demand for safe assets.


    Sure, or even earlier, with Hicks' "Mr. Keynes and the Classics."

  4. Keynes in the General Theory is trying to explain a depression, more than a recession. There might be a difference. On that I would tread carefully.

    The defining core of an economic depression is frustration. People want stuff, they want to work producing stuff for other people, in order to get a share of stuff for themselves, but they cannot. The decentralized market economy ought to allow people to independently supply services and goods, and from the proceeds of that supply, make effective demand.

    If you imagine the circular flow of the economy as money moving in one direction, and goods and services moving in the other, there are distinct questions possible. One is, how is it possible that some people or resources might simply fall out of the circular flow, and have no way to get back in? A separate question might be framed as what governs whether the flow spirals upward or downward, creating an aggregate cycle of ebb and flow? The first question is about a depression, and the second, about recessions in the business cycle.

  5. There are a lot of moving parts in the system under examination, and, by needs, everyone focuses on a few connections at a time, and abstracts the rest. The standards for abstracting the rest are often an embrace of the worst ignorance, which tends to confuse everyone.

    The contemporary, conventional understanding of aggregate demand driving economic fluctuation, for example, rests not just on money, and "sticky prices", but on an often half-forgotten recognition of a particular kind of pervasive industry structure: one in which many firms would like to sell more at the current market price.

    Nick Rowe did a soldierly review here.

    Brad DeLong's summary categorizations are always clear, but seldom altogether accurate. His invention of a Malthus-Say-Mill framework is clever, but leaves out the classical obsession with rent. Rent was also an obsession for Schumpeter. The significance of rent is the risk-bearing capacity it represents.

    In our world of genuine uncertainty, the economy has structure, because some resources, in some uses, earn substantial rents, meaning that even with fluctuation, those resources remain committed to their best uses. Business firms are organized to control production processes, to achieve technical efficiency in reducing error and waste, and firms are organized around their ownership of rent-earning resources. Financial leverage and capital structure are simply "money" mirrors of the "real" structure of ownership of rent-earning resources.

    For Malthus and company, rents were a source of discretion and degrees of freedom in the macroeconomy. It was the landowners, who, producing nothing, but consuming much, with a great capacity to borrow and ride out the storm, held the key to resolving what we would call macro disequilbrium. For Schumpeter, rents were the goal(s) in the contest of rentier and entrepreneur: ways and means and ultimate object, altogether.

    When I think about the U.S. and global economy of the 1920s, increasingly de-stabilized by the aftermath of war, and the culmination of the Second Industrial Revolution, I think I dimly see in the problems of eroding rents, fluctuating output and prices, and the implications of increasing returns for price (p>ac>mc), root causes of the Great Depression, which are not entirely "financial".

    Electrical generation is one example, where increasing returns are enormous, and that poses serious problems for financing, privately, electric utilities. The story of the stock market boom and crash is, in large part, a problem of unrealistic expectations for electric utilities and their profitability.

    The farm economy is another obvious problem. Increasing volatility of farm product prices, increasing potential productivity (from hybrid seed, fertilizers, electricity) were immiserating a large part of the population.

    It is against concrete cases like that that I struggle (in vain) to make sense of Mr. Keynes' abstract "marginal efficiency of capital".

    (When I woke up this morning, this comment promised in imagination to be much better than it turned out, but I'm posting it anyway, with apologies for its raggedness)

  6. I keep coming back to this post, puzzling over it, and writing long, discursive comments (as is my unfortunate wont) and then discarding them.

    My problem is that I want to lecture on "economic rent", the key term (imho) missing from your post. But, I will stifle myself, mostly.

    The basic idea that recessions are about money and finance -- as you describe -- rests on the Walrasian argument for general equilibrium.

    Money, per se, is an explanation for why AD can depart from AS, but not why AD is the governing constraint on output. Nick Rowe did a soldierly review here of that latter issue.

    Brad DeLong is projecting -- anachronistically -- Walras onto his "Malthus-Say-Mill" framework. I don't have texts in front of me, so I will probably be even more egregious in inventing the past than DeLong, but, as I recall, the classical economists were obsessed with rent, eh? Says' Law -- that supply creates its own demand -- was posing the same puzzle: why should AS depart from AD? Malthus, at least, as I dimly recall, thought that the landlords, producing nothing (in Says' sense) and consuming much, were the source of the independence of expenditure necessary to recovering from depressions.

    Investment as a component of AD might fluctuate wildly, but, by itself, that doesn't explain why there should more than transitional and ephemeral unemployment of resources. AD and AS should still converge, even in a society, which lives only for the present moment.

    In a world of genuine uncertainty, all firms are rent-seekers, not profit-maximizers, and money and finance is all about hedging and arbitrage, to make use of the partial knowledge one does have, while minimizing the risk of what one does not know and cannot control. Incentive structures and the distribution of income is all about the distribution of risk and the availability of insurance.

    To make any sense, as an explanation for persistent unemployment, Keynes' musings on the marginal efficiency of capital has to be translated into something more akin to Schumpeter's contest of rentiers and subversive entrepreneurs.

    Too woolly a comment, I fear, but there it is.

  7. Responding to your first comment:

    There are a lot of moving parts in the system under examination, and, by needs, everyone focuses on a few connections at a time, and abstracts the rest.

    Keynes discusses this problem in chapter 21 of The General Theory.

    You're right that mainstream macro grounds sticky prices in imperfect competition. I have to admit, that's one bit of the mainstream that I don't have a problem with. One thing I like about Tom Michl's textbook is that, unlike pretty much any other undergrad macro text I've looked at, it makes this grounding explicit.

    On the rest of the comment, I basically agree, but: The point of this post is to start clarifying what it means to talk about effective demand, or about demand constraints on output. It's mainly intended to push back against the idea, which DeLong insists on, that to say that the economy is demand-constrained just is to say that there is excess demand for some kind of financial asset. I'm not denying that a sudden shift in preferences in favor of safe or liquid assets could cause a shortfall of demand. But I am denying that they are the same thing.

    The key insight we get from Keynes (well, one of them, at least) is that capitalist economies involve a positive feedback loop where incomes determine expenditure, expenditure determines output, and output determines income. Which means that many different levels of activity can be equilibria, in the short run or indefinitely, given the same stock of productive resources and the same individual preferences.

    In that sense, recessions always involve demand shortfalls, since that's what amplifies the initial shock and makes the new, lower-activity position stable. As for the initial shock, it seems to me there are three possibilities. One is a pure demand shock, like a fall in government spending or export demand, or a fall in investment that can't be explained by economic variables. Two is a financial crisis, the preferred New Keynesian story, but also Minsky's. And third is a fall in profits, leading to a decline in investment. That in turn can be explained either by some change in technology or government policy, as the New Classicals do; or by a fall in the profit share due to a stronger bargaining position of labor, as in various Marxist stories; or because intensified competition results in a lower markup (or, I suppose, more precisely, a lower markup relative to the fixee share of costs) -- the story you're suggesting for the 20s and maybe for today. I like Michael Perelman on that stuff.

    Formally a falling markup and rising profit share are equivalent, but of course the sociology and the dynamics are both quite different.

    The m.e.c. looks a little better if you remember that Keynes is very careful to always define it as the yield the entrepreneur expects from incrementally increasing the capital stock, and not the yield the increment will in fact produce. So you could think of it as incorporating the degree of monopoly power the representative producer expects to enjoy.

  8. "a fall in the profit share due to a stronger bargaining position of labor, as in various Marxist stories"

    There is also another possibility, that I think is what Marx himself meant: that the aggregate profits do not fall, but the total of capital rises faster than profits, so that the profit RATE falls.

    For example, think of an economy where all capital goods are financed by credit, so that all profits go into financial interest.
    In this world, all capitalists are financial capitalists.
    Financial capitalist A has a principal of 100, and has a rate of interest of 10%, so that at the end of the year he has 110.
    He immediately reinvests 100 (it would be foolish to bite into principal), and he has also 10 more to spend. But he is thrifty, so he only "eats" 5, and reinvests the other 5 with principal, for a new principal of 105.
    But all the interest come from some material capital good, so if the rate of profit of that good doesn't increase, the total of interest paid to financial capitalists cannot increase either.
    So if the total principal increases by 5, the total rate of interest has to fall. If A is the only capitalist who is increasing his capital, he is basically stealing the share of other capitalists; if all capitalists act as A the rate of interest will fall sharply.
    Even if the increased quantity of capital spurs new investiments, it spurs those investiments in those fields where the rate of profit was lower, and thus were previously ignored, or either increases the competition in those fields where the rate of profit was higer, thus lowering the rate of profit (the situation where all firms in a given field would love to produce more at current prices, but can't find buyers).

  9. RL-

    This gets into some very tricky territory, and I am far from an expert guide.

    But, two things:

    First you need to be careful with

    if the rate of profit of that good doesn't increase

    Why won't the incremental return on new capital be the same as return on existing capital? Marx's main story seems to be that as technology progresses, competition favors more capital-intensive methods of production, while the labor share remains constant. In other words, if we write s/(c+v) = (s/v) / (c/(c+v)), then we think c/(c+v) (the organic composition of capital) rises while s/v (the rate of surplus value) is constant, or at least rises more slowly. It's not an unreasonable set of assumptions, and probably does apply to some periods, but it implicitly assumes a more or less constant wage share, which sits a little uneasily with Marx's assumption elsewhere of a constant real wage, fixed by the (historically determined) value of labor-power. (This raises another question, how we think the wage/profit share is determined. If real wages are fixed, or at least if money wages are more flexible than prices, then we would expect the wage share to be determined by the balance of power in the labor market, as Marxists generally believe. But if output prices are more flexible than money wages, then we would expect the wage share to be set by competitive conditions in the product market, as you find in Kalecki and in many Keynesian models, both mainstream and heterodox.) And the assumption that production becomes more capital-intensive over time isn't as safe as it looks, either: It's generally true within sectors, but historically this has largely been offset by a compositional shift from more to less capital-intensive sectors. So it's not clear how general this tendency is. Anyway, -

    Second, this is very much a long-term story, implying a secular decline in profits. I don't think anyone, Marx, Marxists or anyone else, argues that short-term downturns (recessions, etc.) are due to sudden, temporary increases in the organic composition of capital. There is an interesting literature on how to talk about cycles within a Marxist framework; I like the approach developed by Tom Weisskopf that decomposes changes in the profit rate into changes in the potential output-capital ratio, the utilization rate, the real wage, and the relative price of wage goods. There's a good overview of that approachhere.

  10. Bruce W.-

    Woolly comments are always welcome here. Will have a reply later today.

  11. JM
    (sorry for the late reply)

    "Why won't the incremental return on new capital be the same as return on existing capital?"

    I'm trying to "translate" Marx's logic in modern terms. I believe that the point is that the demand for consumption goods is somehow fixed or limited. So when new capital comes in (I'm speaking of material capital, like new factories etc.) it will either:
    a) produce more of the same stuff, thus causing a fall in price of this stuff and lowering the rate of profit in this branch of the market, or
    b) produce a different, new kind of stuff, trying to satisfy new "demands" that weren't previusly perceived (that is, creating new markets). But, if we believe that cappitalists are more or less rational, I think that it is a fair assumption that most "good" markets are already filled, and thus most new markets will have low possible profits.

    Referring to the "secular" fall of the rate of profit, I believe that Marx in fact was speaking of different time horizons:

    1) on the short term prices fluctuate because of supply and demand

    2a) but on the medium term, supply and demand are basically driven by technological factors. This means that on the medium term equilibrium the cost of stuff is proportional to the amount of labor needed to produce it.

    2b) in this medium term, the rate of profit has to fall, because of the reasons that I said above. This would be a permanent depression/deflation economy (but Marx never really explains the role of credit).

    2c) but this permanent state of depression can be relieved by various kinds of exogenous or semi exogenous shocks (the "countervailing factors" to the tendency of the rate of profit to fall, listed on the Capital). One of those countervailing factors is technologic improvement.

    3) but on the long, secular term the tecnologic improvement is supposed to cause an increasing centralization of the capitalist system, which causes the phenomenon of the decrease of the "variable" part of capital.