Tuesday, December 09, 2008

Urgent message for Brad

I have found that posting things here is more effective than commenting on his blog which is more effective than e-mail.

Brad is live at Cato.

He makes Weiner Schnitzel of a Neo-Austrian. In fact the analysis is brilliant. It is based entirely on dollar quantities.

However, there are typos.

First in three, Brad switched "present" and "future".

3)

Liquidity Discount: The cash flowing to capital arrives in the *present* rather than the *future*, and people prefer—to varying degrees at different times—the bird in the hand to the one in the bush that will arrive in hand next year. Fluctuations in this liquidity discount are a third source of fluctuations in global capital wealth.


Second Brad changed the numbering of 3 and 4, but in the text refers to time preference as 4) and default risk as 3. (too many cases to list).

Below non urgent comments.

Also why do you call time preference liquidity preference ? They aren't the same. Active traders want liquid assets so they don't lose on bid/ask spreads. That has nothing to do with their time preference. This is important as the liquidity of assets can change quickly. For an asset to be liquid it must be either cash or there must be high volumes of trade for reasons other than rationally expected returns (you proved this in high school unless you were in junior high). Since we don't have a good model* for noise traders, we can't predict changes in market thickness. This is what has changed. Liquidity preference which is different from time preference.

You have two anomalies and one idiotic assumption. You assert that the intertemporal elasticity of substitution of consumption is between 0.5 and 1. Based on what evidence ??? Empirical estimates are from 0.1 to 0.5. It is just a rule among mainstream economists that one must assume it is between 0.5 and 1 no matter what the data say. Then there is the equity premium puzzle. To be mainstream you must assume that the coefficient of relative risk aversion is between 1 and 2 (corresponds to intertemporal elasticity of consumption between 0.5 and1). Therefore the equity premium is a big mystery. If you accept a reasonable intertemporal elasticity of substitution such as 0.25 both mysteries become less mysterious. Where the hell did you get your "estimate" of the intertemporal elasticity of substitution in consumption ?

Obviously when highly leveraged banks take a hit, there is an increase in default risk.

* We do, however, have a bad 18 year old model which will work fine. We need to modify DeLong et al 1990 to add investors who have information on next periods noise. Their problem is that they don't know next periods mu (share of people who are noise traders.

My capsule summary is that investment banks and hedge funds were making a killing fleecing the sheep -- tricking uninformed investors into betting against them. The housing bubble bursting was an event so big that noise traders realized that they were better off holding only safe assets or holding the market (I don't know which). The financial operators took losses (not huge but painful) and wanted to de-leverage a bit. They discovered that they couldn't find suckers to buy their risky assets and sell them assets to cover their risky short positions. Thus the whole system collapsed. The Freezing up and melting down and other metaphorical phase transitions can be understood as a decline in the fraction of money traded by uninformed investors.

Update: Nick Rowe (who has learned how to log in to blogger comments see comment below) was on this before I was.


update 2: pulled back from comments.

Comments:
What I want to know is why he can write a persuasive comment on the current crisis without talking about balance sheets, bankrupcy, agency or international finance. That tells me where the holes in macro-economic theory are.
# posted by Blogger reason : 12:02 PM

Yes!

Check my post along similar lines:
http://worthwhile.typepad.com/worthwhile_canadian_initi/2008/12/liquidity-time-preference-and-brad-delong.html
You do a good job of explaining WHY liquidity declined (I didn't try, except to offer evidence that it had in fact declined.

Nick Rowe
# posted by Anonymous Anonymous : 2:38 PM

Dear Nick

I saw your post more or less copied over at economists view. I seriously considered commenting "I said that too". Thoma found this post without my help.

I think Brad is channeling Keynes. I'm sure Keynes conflated time preference and liquidity preference. Also it made more sense to conflate them when talking about finance back before Black and Scholes (most things about finance made more sense before Black and Scholes).

Reason, I too found it odd that he didn't talk about balance sheets etc. They are left as part of increased perceived risk. This doesn't work so well, since he distinguishes risk aversion and fear of default and the new risk is that financial counterparties will default.

I would also mention that, for the editor of evisceration weekly, he is very very polite. I mean that Fannie Freddie argument is blatant nonsense, yet he manages to critique it gently.
# posted by Blogger Robert : 2:47 PM

I agree with what you say about Keynes. A lot of Cambridge (UK) seem to have been very shaky on time preference; I wonder if any of them had ever seen an Irving Fisher diagram, because the preference side is missing from all of the UK end of the Capital Controversy. I don't see how Black-Scholes changes this though?

My next task for myself is to try to understand better what determines liquidity. First stab: if someone is trying to sell me a used car cheap, is it because it's a lemon, or is he really leaving the country next week? Liquidity is determined by the ratio of ex-pats to lemons.

Nick Rowe (I must learn how to log in properly, or maybe I've succeeded?)
# posted by Blogger Nick Rowe : 4:07 PM

The logic, such as it is, of my reference to Black and Scholes is as follows.

I think that this is mostly a verbal tick (like calling Ukraine "the Ukraine". That is, the old UK English definition of liquidity preference = time preference was, OK in its time until people began to value liquidity for its own sake.

I don't know about Keynes but he seems to have assumed that the only reason one would want to liquidate assets would be to consume with the resulting cash. Roughly, his model of investing seems to be that one saves, buys an asset and holds it unless and until one wants to consume (or for a firm invest in fixed capital or pay workers or suppliers).

Today's huge volume of trading is new.

Black and Scholes presented a model in which one can make money without risk if an option is miss-priced. They assume 0 bid ask spreads and the hedging strategy requires constant trading. A huge number of people make (or made) excellent incomes applying their insight to cases in which their assumptions didn't hold. For such traders, liquidity is very important even though they are not buying goods or services.

Now ask how much *should* decreased liquidity reduce the value of assets. I would say very little. Let's assume that everyone is rational and that all assets are traded (no risky labor income, no accidents). In such a model everyone should buy and hold the market. If they are saving, they are buying and if they are dis-saving they are selling, but there is not huge rush.

With current bid ask spreads, they either have to forgo a few days or maybe weeks of interest or pay an expected loss and bear the risk on the order of two bid ask spreads for the period between saving and dis-saving. That just doesn't amount to 3.3% per year. It shouldn't add up to 0.3% per year. Maybe 0.03% per year but I doubt it.

Now there are untraded assets. People might want to liquidate assets, because they lost their jobs, or might want to buy more because they won the lottery. Roughly gross flows of saving and dis-saving are greater than they would be in a simple life cycle model. This gets us up to maybe 0.1% per year.

In theory they might want to hedge their labor income risk (but I mean, really, obviously no one does that).

So if all investors are rational, liquidity shouldn't matter much at all.

Now the active traders who suddenly don't want assets because they are illiquid might be rational. However, that would imply miss-pricing *and* people willing to take the other sides of their trades. That requires irrational investors (or absurd assumptions about non traded assets).

My claim is that liquidity shouldn't matter much. If people care a lot about it, either they are irrational or they are taking advantage of irrational people or, most likely, both.

This is of some interest because defenders of financial engineering (innovative assets) say they are good because they increase liquidity which is good, because it is useful for rational active traders who drive prices towards fundamentals.

I don't think financial operators tricking less sophisticated people out of their money is a good thing. I also think that said operators are clearly irrational themselves. And the idea that higher trading volumes are associated with prices closer to fundamentals requires the idea that the fundamental value of the S&P 500 changes by a few percent a day.
# posted by Blogger Robert : 9:03 PM

8 comments:

  1. What I want to know is why he can write a persuasive comment on the current crisis without talking about balance sheets, bankrupcy, agency or international finance. That tells me where the holes in macro-economic theory are.

    ReplyDelete
  2. Anonymous2:38 PM

    Yes!

    Check my post along similar lines:
    http://worthwhile.typepad.com/worthwhile_canadian_initi/2008/12/liquidity-time-preference-and-brad-delong.html
    You do a good job of explaining WHY liquidity declined (I didn't try, except to offer evidence that it had in fact declined.

    Nick Rowe

    ReplyDelete
  3. Dear Nick

    I saw your post more or less copied over at economists view. I seriously considered commenting "I said that too". Thoma found this post without my help.

    I think Brad is channeling Keynes. I'm sure Keynes conflated time preference and liquidity preference. Also it made more sense to conflate them when talking about finance back before Black and Scholes (most things about finance made more sense before Black and Scholes).

    Reason, I too found it odd that he didn't talk about balance sheets etc. They are left as part of increased perceived risk. This doesn't work so well, since he distinguishes risk aversion and fear of default and the new risk is that financial counterparties will default.

    I would also mention that, for the editor of evisceration weekly, he is very very polite. I mean that Fannie Freddie argument is blatant nonsense, yet he manages to critique it gently.

    ReplyDelete
  4. I agree with what you say about Keynes. A lot of Cambridge (UK) seem to have been very shaky on time preference; I wonder if any of them had ever seen an Irving Fisher diagram, because the preference side is missing from all of the UK end of the Capital Controversy. I don't see how Black-Scholes changes this though?

    My next task for myself is to try to understand better what determines liquidity. First stab: if someone is trying to sell me a used car cheap, is it because it's a lemon, or is he really leaving the country next week? Liquidity is determined by the ratio of ex-pats to lemons.

    Nick Rowe (I must learn how to log in properly, or maybe I've succeeded?)

    ReplyDelete
  5. The logic, such as it is, of my reference to Black and Scholes is as follows.

    I think that this is mostly a verbal tick (like calling Ukraine "the Ukraine". That is, the old UK English definition of liquidity preference = time preference was, OK in its time until people began to value liquidity for its own sake.

    I don't know about Keynes but he seems to have assumed that the only reason one would want to liquidate assets would be to consume with the resulting cash. Roughly, his model of investing seems to be that one saves, buys an asset and holds it unless and until one wants to consume (or for a firm invest in fixed capital or pay workers or suppliers).

    Today's huge volume of trading is new.

    Black and Scholes presented a model in which one can make money without risk if an option is miss-priced. They assume 0 bid ask spreads and the hedging strategy requires constant trading. A huge number of people make (or made) excellent incomes applying their insight to cases in which their assumptions didn't hold. For such traders, liquidity is very important even though they are not buying goods or services.

    Now ask how much *should* decreased liquidity reduce the value of assets. I would say very little. Let's assume that everyone is rational and that all assets are traded (no risky labor income, no accidents). In such a model everyone should buy and hold the market. If they are saving, they are buying and if they are dis-saving they are selling, but there is not huge rush.

    With current bid ask spreads, they either have to forgo a few days or maybe weeks of interest or pay an expected loss and bear the risk on the order of two bid ask spreads for the period between saving and dis-saving. That just doesn't amount to 3.3% per year. It shouldn't add up to 0.3% per year. Maybe 0.03% per year but I doubt it.

    Now there are untraded assets. People might want to liquidate assets, because they lost their jobs, or might want to buy more because they won the lottery. Roughly gross flows of saving and dis-saving are greater than they would be in a simple life cycle model. This gets us up to maybe 0.1% per year.

    In theory they might want to hedge their labor income risk (but I mean, really, obviously no one does that).

    So if all investors are rational, liquidity shouldn't matter much at all.

    Now the active traders who suddenly don't want assets because they are illiquid might be rational. However, that would imply miss-pricing *and* people willing to take the other sides of their trades. That requires irrational investors (or absurd assumptions about non traded assets).

    My claim is that liquidity shouldn't matter much. If people care a lot about it, either they are irrational or they are taking advantage of irrational people or, most likely, both.

    This is of some interest because defenders of financial engineering (innovative assets) say they are good because they increase liquidity which is good, because it is useful for rational active traders who drive prices towards fundamentals.

    I don't think financial operators tricking less sophisticated people out of their money is a good thing. I also think that said operators are clearly irrational themselves. And the idea that higher trading volumes are associated with prices closer to fundamentals requires the idea that the fundamental value of the S&P 500 changes by a few percent a day.

    ReplyDelete
  6. Thanks Robert. Now I understand your Black Scholes reference.

    I'm still trying to decide how much ordinary people (i.e. non professional traders) need liquidity. Buying big consumer durables? Sudden emergencies? I will have to think about that one.

    I had another go at the topic over at Worthwhile Canadian Initiative, building a toy model based on Akerlof's Lemons model. The key is that only the "lemon" assets come to market, rather than a bid-ask spread in the market. There is no bid-ask spread, but you can't sell assets for what they are worth, so they are illiquid.

    ReplyDelete
  7. Anonymous2:29 PM

    nice pictures

    ReplyDelete
  8. Anonymous4:43 PM

    Thanks for this nice article.

    ReplyDelete