Friday, April 29, 2011

On Coordination Mechanisms and the Market

Jonathan Catalan responded extensively to "Lord Keynes"* and myself here, and then "Lord Keynes" put up a response back to Jonathan. I wanted to take issue with one thing in his response. "Lord Keynes" writes:

"...it is precisely that notion of a coordination mechanism in the market that Post Keynesians reject... Catalán might care to acknowledge that even some Austrians reject the notion of plan/pattern coordination in free markets, so even his own economic school is divided on that issue. Anyone who rejects the belief that the free market has a coordination mechanism will see that there is bound to be a waste of resources and involuntary unemployment in such a system."

Let's be very careful here, "Lord Keynes". The market does have a coordination mechanism and I don't think even Post Keynesians reject that. You have an effective demand schedule, do you not? You have a supply schedule, do you not? OK - you have market coordination. What you mean is that you don't always have a clearing market or an optimizing market. But you do have coordination. Investments aren't made and prices aren't set by throwing darts at a dart board. Prices coordinate.

This is an important distinction and it gets back to the distinction between central planning and demand management. What coordinates resources for Post Keynesians or any sort of Keynesians? Markets and the price mechanism. What are the aggregate properties of that coordination? Well that's where we Keynesians may depart from Austrians and others. But don't say there's no market coordination. Of course there is.

This bothers me about Joe Stiglitz a lot. He's very casual with phrases like "markets don't work". Well what do you mean by that? Don't work in what sense? Do they really "not work" or do they underperform in some systematic way. To say that market coordination does not guarantee full employment is very, very different from saying that there is no market coordination. The market still guarantees that positive net benefit transactions occur and negative net benefit transactions don't occur. That's coordination. No central planner could guarantee that. Whether that results in full employment is a different question entirely.

* - I don't know about anyone else, but this guy's screen name always strikes me as a little overfamiliar and presumptuous.

20 comments:

  1. "To say that market coordination does not guarantee full employment is very, very different from saying that there is no market coordination. The market still guarantees that positive net benefit transactions occur and negative net benefit transactions don't occur. That's coordination. No central planner could guarantee that. Whether that results in full employment is a different question entirely."

    Fair enough.
    But the issue then comes right back to whether Say's law works, and whether there is mechanism that causes free markets to result in full employment equilibrium and optumum use of resources.
    I think most my comments stand.

    ReplyDelete
  2. Incidently, do you regard yourself as a New Keynesian?

    ReplyDelete
  3. Oh - I definitely think your comments on Say's Law stand. I think I'm with you more often than not. But people like Steve Horwitz will say this - that there is no price coordination in the Keynesian model. He means the same thing, essentially - there's no guarantee of clearing. My concern is that when we confuse a clearing market with a coordinating market, we weaken the case for markets and free exchange. And that is very dangerous.

    On "New Keynesian"... I suppose I don't worry about it all that much and generally just call myself Keynesian. It depends on what you mean by "New Keynesian". I don't think "microfoundations" are especially important, although of course there's nothing wrong with that project. I don't rely on frictions in the way that New Keynesians do. Of course frictions are real but you don't need to assume frictions to get "full employment by accident". I don't have the concerns about the neoclassical synthesis that you guys do, but I also don't consider myself second to anyone on the recognition of the role that uncertainty plays in money demand and investment decisions.

    You know, really I'm not uncomfortable with anything the New Keynesians do, but I think in a lot of cases they're barking up the wrong tree by making so much of the story frictional. The heart of economic downturns is in the fact that the interest rate doesn't clear the system and the interest rate doesn't clear the system because the loanable funds theory of the interest rate is wrong (or incomplete), and the loanable funds theory of the interest rate is wrong because of fundamental uncertainty about the future. Frictions are icing on the cake, but not the main story.

    ReplyDelete
  4. If someone pressed me I'd say "Keynesian" - not "New Keynesian" or "Post Keynesian".

    ReplyDelete
  5. The heart of economic downturns is in the fact that the interest rate doesn't clear the system and the interest rate doesn't clear the system because the loanable funds theory of the interest rate is wrong (or incomplete), and the loanable funds theory of the interest rate is wrong because of fundamental uncertainty about the future

    We've had this argument before and you have never substantiated it. WHY does uncertainty have ANYTHING to do with the interest rate? Uncertainty is manifested in cash holdings (I'll agree with Keynes on liquidity preference here, although he wasn't the first to describe it as such). But you have yet to explain WHY liquidity preference has anything to do with the interest rate. If we agree that the interest rate is a price paid for the use of capital, then it seems to follow that the SUPPLY and DEMAND for capital are what determines its price, irrespective of how uncertain people are because uncertainty has nothing to do with saved capital; although it has everything to do with idle cash balances.

    ReplyDelete
  6. Why does that follow?

    Rent controlled rent is the price paid for many apartments in NYC. It does not "follow" that the rent controlled rate of rent is determined by the supply and demand for rental housing.

    ReplyDelete
  7. There is no necessary correspondence between the phrases "the price that is paid for X" and "the supply and demand for X determine its price". None at all.

    I would think a guy that fancies himself an a priorist and a logician wouldn't make that sort of slip.

    ReplyDelete
  8. It does not "follow" that the rent controlled rate of rent is determined by the supply and demand for rental housing.

    No, it's determined by government fiat (hence - "controlled"). In a freed economy, however, which is what we're implying when discussing theories of interest (not "controlled" by Federal Reserve fiat) the equilibrium price for capital is determined by supply and demand.

    There is no necessary correspondence between the phrases "the price that is paid for X" and "the supply and demand for X determine its price". None at all.

    It sounds like you're arguing that supply and demand are not determinants of equilibrium price. I would think a guy that fancies himself an economist wouldn't make that sort of slip, so I must be misinterpreting you.

    What are you saying, exactly?

    ReplyDelete
  9. To be precise Dan, I don't think I've said Keynes denies "price coordination" across the board, only intertemporal price coordination. If I've made the broader claim, I'd like to see the context.

    For example, check the "Contrasting Concepts of Capital" paper:

    "For Keynes, intertemporal coordination via the market is more the exception than the rule, as he says it would only be by luck that savings and investment would be equal at full employment. Where that luck does not hold, the Keynesian model shows that employment and income must adjust to ensure the equality of savings and investment. The work done in the classical and Austrian model by the interest rate in assuring that desired savings equals desired investment and thereby ensuring that changes in savings preferences do not reduce total income or employment has no counterpart in the Keynesian model. With that interest rate mechanism rendered inoperative, Keynes must rely on those changes in income (and thus employment) to adjust savings to the exogenously determined quantity of investment. Where the classical and Austrian economists saw the interest rate as ensuring that the true-by-definition ex post equality of savings and investment would match their ex ante equality, Keynes requires changes in income and employment to turn a pervasive ex ante inequality of savings and investment into an ex post equality. It is precisely on this point that economics moved from the microeconomics of the capital structure to the macroeconomic aggregates of unemployment and total income."

    You might disagree with that story, but it's not denying price coordination across the board in Keynes.

    ReplyDelete
  10. "Rent controlled rent is the price paid for many apartments in NYC. It does not "follow" that the rent controlled rate of rent is determined by the supply and demand for rental housing."

    The published rate, anyway. Sub rosa sub-leases of controlled rent apartments with under-the-table "over" payments are common. There are many variations. In some cases the landlord gets the money; in some cases the leach gets the money. Makes new apt. construction kind of hard. Some people would start spouting off about "externalities" of government planning here. Liberal heaven has particularly idiotic versions:

    http://www.thelocal.se/24910/20100210/

    But I'm not saying there are any similarities between rent fixing and fiat money price fixing.

    ReplyDelete
  11. Steve -
    Duly noted. It was that paper I had in mind, so yes, intertemporal coordination. I still disagree with you, of course - for the reasons we went over on your blog. Your passage is a good example, because again it confuses market clearance with market coordination. It is true markets don't always clear for Keynes - but the interest rate still coordinates what gets done. High net benefit projects are invested in and low ones are not because of the coordination of the price signal. But I should have been more specific on about "intertemporal" coordination.

    Mattheus -
    re: "It sounds like you're arguing that supply and demand are not determinants of equilibrium price. I would think a guy that fancies himself an economist wouldn't make that sort of slip, so I must be misinterpreting you."

    Well - that they're not *always* the *only* determinants of equilibrium prices (and prices aren't *always* equilibrium prices). What's so non-economisty about saying that? That's just a simple fact. In the real world, supply and demand rarely determine equilibrium prices: marginal cost and marginal benefit do. And that's even before we get into questions of whether markets clear.

    ReplyDelete
  12. DK,

    You can have a market clearing model of the business cycle and still have a use for an activist (“Keynesian”) government policy because of coordination failures.

    In coordination failure models, strategic complementarities can lead to multiple equilibria, some of which are bad. Here's a quick e.g.: say that there are aggregate increasing returns to scale and constant RTS at the level of the individual firm because of strategic complementarities that exist between firms. Say further that this results in an aggregate production function that is convex in the labour input (holding K fixed). Now we have an increasing marginal product of labour schedule (for the whole economy), which implies that the agg demand for labour curve is also upwards sloping. If the IRTS are large enough, the labour demand curve will be steeper than the labour supply curve. Consequently, increases in the real interest rate that shift the labour supply curve to the right will result in lower employment and lower output. That is, the output supply curve is downwards sloping in the real interest rate. Now imagine that the output demand curve intersects the output supply curve in two places. One of these points of intersection will represent an equilibrium where output and employment are low and the real interest rate is high. The other, an equilibrium where output and employment are high and the real interest rate is low.

    A couple more things: What did you mean when you wrote that microfoundations are unimportant? I read that literally and it doesn’t make sense to me.

    If the real interest rate isn’t determined by supply and demand, what does determine it?

    LK,

    Not sure that it makes sense to think of anyone as a “New Keynesian”, except in some very limited contexts. The phrase refers to a class of models that take an RBC baseline and add monopolistic competition and some nominal rigidities. It’s not a political movement.

    ReplyDelete
  13. vimothy -
    Definitely on the necessity of intervention in a market clearing model.

    I talk about some of this w.r.t. the labor market here: http://factsandotherstubbornthings.blogspot.com/2011/04/problem-of-micro-imperialism-in.html

    For some reason, we've managed to equate "labor surplus" with "unemployment", and "clearing labor market" with "no unemployment".

    Why have we done that?

    Hell if I know. Clearing markets say something about efficiency and may have relevant welfare maximization properties (if costs and benefits are sufficiently internalized), but they don't dictate whether policy is appropriate or not.

    Your examples with multiple equilibria and coordination failures are good as well.

    As for microfoundations: I simply mean that we should not get caught up in deriving aggregate properties from disaggregated behaviors. There is no reason to privelege microfoundations over macrofoundations (although if we had lots of convincing examples of both, obviously that's wonderful). The only reason why people worry about microfoundations is because of the anthropocentrism. You don't see physicists forcing relativity theorists and cosmologists to play second-fiddle to quantum theorists. Both ends of the scale are valid starting off points.

    ReplyDelete
  14. Also

    re: "If the real interest rate isn’t determined by supply and demand, what does determine it?"

    The question isn't "does supply and demand determine the interest rate?". Certainly it does. The question is "which supply and demand determines the interest rate?".

    Bond market?
    Money market?
    Both?

    That's the rub.

    ReplyDelete
  15. Every market is a money market.

    It would be just as correct/incorrect to ask "which supply and demand determines the price of widgets?"

    Widget market?
    Money market?
    Both?

    At least, I do not see a difference. Where am I going wrong?

    ReplyDelete
  16. Don’t agree at all re microfoundations. If you can’t derive aggregate behaviour from individual behaviour, you don’t understand it at all. There’s not necessarily anything wrong with this, unless you actually want to understand the behaviour of aggregates. In which case, you most certainly do need to understand how the behaviour of individuals produces the behaviour of aggregates.

    I’m not sure of what model you have in mind, but if it’s something like ISLM-ASAD or similar from PKE like an “SFC” model, I don’t know how you could find it satisfying. Or useful for much, for that matter.

    Highly aggregated macro models are basically massive cop outs, IMO.

    ReplyDelete
  17. Don’t agree at all re microfoundations. If you can’t derive aggregate behaviour from individual behaviour, you don’t understand it at all. There’s not necessarily anything wrong with this, unless you actually want to understand the behaviour of aggregates. In which case, you most certainly do need to understand how the behaviour of individuals produces the behaviour of aggregates.

    I’m not sure of what model you have in mind, but if it’s something like ISLM-ASAD or similar from PKE like an “SFC” model, I don’t know how you could find it satisfying. Or useful for much, for that matter.

    Highly aggregated macro models are basically massive cop outs, IMO.

    ReplyDelete
  18. Daniel,

    In the real world, supply and demand rarely determine equilibrium prices: marginal cost and marginal benefit do.

    You're just repeating what I said. Note: I never mentioned aggregate demand or aggregate supply. What determines prices? The demand people hold on these items, and the cost it take to produce them. That's just a longer way of saying "supply and demand."

    The question isn't "does supply and demand determine the interest rate?". Certainly it does.

    Well that was a waste of a paragraph I just wrote! Make up your mind, dude.

    The question is "which supply and demand determines the interest rate?".

    Fine, if you want to be pedantic about it. I think it is overwhelmingly obvious that we're talking about CAPITAL (the interest rate functions as a price for capital). How the hell can cash balances (and uncertainty) affect the going price of capital? Savings are not related to cash balances - or, at least, I've never heard a single reason to suppose they are related (and a very good one to show they aren't).

    ReplyDelete
  19. Um, nothing "pedantic" about it, Mattheus. What would be "unpedantic" - if I agreed with you?

    ReplyDelete
  20. It would be unpedantic if you stopped arguing over formalisms and details, and actually engaged me in debate over the meat and potatoes of our theories.

    I'm not asking you to agree with me at all.

    ReplyDelete

All anonymous comments will be deleted. Consistent pseudonyms are fine.