Monday, November 10, 2008

Battling Bubbles: Speaking for Galbraith II

Probably the most difficult proposal I shall make will be to advocate a much more vigorous role by the government in preventing bubbles from getting too large. I understand that the Fed is thinking about this, but I think this is a function that should be spread across various agencies, with perhaps the CEA being the one to try to advocate for something being done. Different bubbles may require different tools, with very precise ones that influence the relevant markets the best. Thus, for an oil bubble, use selling oil from the Strategic Petroleum Reserve. Paul Davidson has long advocated using buffer stocks of commodities to regulate excess price volatility in some markets. This can be done for some agricultural commodities as well, presumably by the USDA. In housing markets, limits on certain kinds of mortgages can be imposed, perhaps by the Fed, no interest only or other mortgages that encourage prices of housing too high for most buyers. In broader financial markets, the SEC or the Fed can use margin requirements to slow the rush to buy and push up prices, and clearly various derivatives not now under such regulations should be brought under them. So, a flexible policy not focused on one tool or approach should be used.

In this regard, let me issue a warning. While many have sneered at Alan Greenspan for his reluctance to do anything about bubbles, it should be kept in mind that, aside from margin requirements, the main tools available or the Fed are probably too blunt. Thus, no one should forget that the overly tight monetary policy of 1929-30 was substantially driven by a desire to squeeze down the stock market bubble of the late 1920s. Clearly a tool can be overused. And, while many also sneer, it is also not always that easy to know when a bubble is going on, although I think this can be done. In this regard, Greenspan himself jumped the gun with his famous remarks in 1996 cautioning against "irrational exuberance," which caused the stock market to drop the next day, but then to start climbing again, with the main market indices not ever getting down that low again (although I think the NASDAQ has, which was the bubbliest).


reason said...

Tobin tax? said...

I confess to having mixed feelings about the Tobin tax. There are studies that show that it can reduce volatility, there are others that go the other way. I am even aware of one study that says it could reduce variance but increase kurtosis, that is, fat tails or more extreme events.

Besides the theoretical issue there is the practical problem that unless it is global, one ends up pushing financial transactions to offshore places like the Cayman Islands. We already have too much of that going on. So, while I sort of like the idea of it, and it would be not a bad way to raise revenues, I also see some potentially serious problems with it, both theoretically and practically.

Anonymous said...

A Tobin Tax is exactly the wrong tool. See my article in the ECONOMIC JOURNAL,1997, vol. 107 entitled "Are Grains of Sand in the Wheels of International Finance Sufficient to do the Job When Boulders Are Often Required?" where I demonstrate that the Tobin tax does not discriminate agaist short term finance and even works aginst stabilizing arbitrage actions, etc. Jim Tobin admitted to me that my argument was correct-- and in the 2000 issue of World Development in a symposium on international finance and the payments system, his only defense of the Tobin tax was "How would you like to have a tax named after you?"

People should read my latest book JOHN MAYNARD KEYNES (Palgrave, 2007) in Palgrave's "Great Thinkers in Economics" series for a further discussion and relating to today's financial crash.

Paul Davidson said...


There is a paper floating around by Frank Westerhoff of Bamberg University that was cited in a recent New York Times column that uses a pretty believable agent-based model to show that one gets substantially reduced variance in a lot of cases for financial markets with a Tobin tax. OTOH, it was Westerhoff who brought to my attention this other paper that showed it might lead to more extreme events, which is certainly not what we want.

Anonymous said...

Well there are studies that show the opposite -- and in fact statistically they should show that a transaction tax increases the variance all other things being equal. As I point out in my paper "Volatile Finacial Markets and The Speculator" there is an excellent article by Jones and Seguin , AER, 1997 that demonstrates using NY Stock exchange , Amex and OTC data, that higher transaction costs increase volatility while decreased transactions costs lower volatility. J and S used a portfolio sample of 1872 stocks where samples were drawn for one year before and one year after a tranaction cost change occurred.

Statistically that should not be surprising since ceteris paribus (as I argue in my article) the variance is deceased as the size of sample events recorded increases. Where a transaction cost increase occurs, we should expect less transactions (a smaller sample) and therefore a larger variance.

I do not know Westhoff's study -- but I wold guss it is not a scientific ceteris paribus study as Jones and Sequin produced. Another study that has the same results as Jones and Seguin is the Umlauf study of theSwedish Stock market. in the Jour. of Financial Economics. Both are cited in my paper.

Myrtle Blackwood said...

Governments themselves have played a significant role in feeding the financial bubble. I have been collecting observations to support this assertion for a while but don't have it with me right now.

How do we limit the scope of government actions in this respect? Greater 'economic democracy' would have a role to play here. said...


Much of this depends on the power of those making money from a bubble. A major problem here is that once a bubble gets going, there are people making money from it, and they resist mightily any effort to interfere with that.


The Westerhoff model is ultimately of the species where there are types of traders, with fundamentalists (whom I know you do not think exist, or if they do, they are misguided) and some sorts of technical traders or trend chasers, with learning dynamics. Anyway, volatility arises when there are herding dynamics and the markets get dominated by trend chasers who chase each others' tails, hence the bubbles. The Tobin tax slows down that dynamic.

Regarding the studies you cite, I would agree that when markets get too thin, then volatility can increase, especially when they get very, very thin. But the sort of markets we are talking about here are extremely thick. Some of these financial markets, such as forex ones, have trillions of dollars of exchanges a day. One can slow down the volume of these operations quite considerably and still be far away from the range of volume where one starts to get those sorts of thin-market-volatility effects.

Anonymous said...

If the Westhoff study is only what you report -- and has no empirical facts to back it up -- then it is simply GIGO.

The studies I mentioned (Jones and Seguin and Umlauf studies)are empirical studies where volatility is meassured by variance. Moreoever they are set up as proper experimental studies.

And the statistical law that the variance increases with a decline in sample size is true for ALL sizes of samples.

And if the Tobin Tax does not reduce the number of trades (not necessarily the volume of dollars traded) then it is useless in stopping bubbles. And if the Tobin Tax does reduce the number of trades, then by statistical relationship law it must increase the variance -- and unless you are measuring volatility by something other than variance -- the Tobin tax must increase the volatility as measured by variance. No hypothetical Westoff model can stop that!!

Anonymous said...

By the way Barkeley, I forgot to mention that the empirical studies that I mentioned were not in thinly traded markets -- they were in the NYSE, AMEX, OTC, And Swedish stock markets -- and the size of samples was very large and number of observations over time was very large. So we are not talking about thin markets.

AS a theorist, I am somewhat surprised that I have to invoke good statistical studies and statistical theory to drive my point home--).

TheTrucker said...

Returning to the actual meat of the article as "regulation" I am inclined to cite Gramm-Leach-Blily and the Commodity Futures Modernization as the principle government failures that formed the supporting structure for the housing bubble and financial mess. And then, of course we had the cheer-leading of the Bush administration driving the bubble in order to sell trickle down stupidity as a Republican economic miracle. (And, of course, to "finance" a military adventure without taxation.)

The Repeal of Glass-Steagall (Gramm-Leach-Blily) removed transparency form relationships between banks and insurance underwriters and brokerages. It also removed the competitive relationships where each separate entity was more careful about its own well being. The resulting amalgamations were more efficient but less accountable (like a closed parachute the financial sector went very fast).

The "Commodity Futures Modernization Act" stopped the efforts to regulate credit swaps and other such devices as "futures contracts" like oil and pork bellies. The point of such an exchange would have been to assure balance and enforce margin requirements.

Now is not the time to "put Humpty Dumpty back together again", but after the salvation of nationalization (assuming there is any salvation) it may be possible to resurrect a properly regulated private financial sector.

kevin quinn said...

On the Tobin tax: isn't the idea to put some check on the growth of bubbles by checking the tendency to get on the elevator of an over-valued asset and get off before the ascent stops - to do a less extreme version of Keynes' suggestion (made with tongue partly in cheek, but still) that the asset-holder be married to the asset for life. I guess it seems to me that Paul's point about thin-ness and standard deviation may be right without under-cutting this rationale for the tax: it might still do something to prevent extreme bubbling, by focusing investors on the fundamentals. What am I missing? said...


I am with you on the Commodities Futures Modernization bill. I think we need to regulate credit-default swaps more vigorously, and probably to simply outlaw certain kinds of more destabilizing transactions, including the flakier kinds of mortgages. I am less convinced that repealing Glass-Steagall was such a big deal. The big banks currently in the best shape in the US have both kinds of operations.


I am aware that the studies you cite deal with large markets. However, they are also mostly dealing with markets not in the midst of major speculative bubbles, so they do not catch what is going on.

What is going on in the Westerhoff model, and looks quite reasonable in the real world, is that the bubbles are associated with a predominance of trend chasers in the market, with an increased volume, with that volume being by them. To model shows that in a bubbly period, those trend chasers will be disproportionately cut back, thereby stabilizing the market. During a stable period, a Tobin tax may increase volatility, as you have argued.

I remind that I was only asked about the Tobin tax here in the comments. While I am not definitely opposed to one as you are, Paul, I share enough doubts about it that it is not something I am going to be supporting, or even addressing, in my remarks on Friday. Frankly, my biggest problem with it is not your complaint, but this more practical problem regarding what happens when it is not universal and speculators can simply move the action offshore somewhere. I think other tools are superior to battle bubbles effectively than a Tobin tax.