« What caused the financial crisis? | Main | Brokers with Hands on Their Faces »
Will the price of risk be too high or too low?
From the comments at MR:
...we had all better hope that there will be some stupid groups in the future, because if not, then our society will be poorer due to a societywide excessively high price of risk. an excessively high price of risk isn't as spectacularly catastrophic as the excessively low price of risk of the last 10 years, but compounded over time it can do just as much damage...
I hold a few beliefs:
1. For a while the price of risk had been too low.
2. Currently the price of risk is too high.
3. In response to the crisis, we will regulate to prevent the particular previous manifestations of #1. The bad news is this will be an overreaction; the good news is that because of #2 the regulatory overreaction won't matter for some while.
4. We do not know how to regulate to prevent other, future, hitherto unexperienced manifestations of an excessively low price of risk.
5. Maybe #4 is wrong, but beware of any huff-and-puff polemic discussion that is not at least considering these points.
Posted by Tyler Cowen on October 19, 2008 at 02:00 PM in Economics | Permalink
Comments
Tyler, can you speak to what you mean by "the price of risk?" I think you're using "risk" in a sense that's unfamiliar to me -- I'd say risk is the probability of losing one's investment, then risk itself is the price... Please clarify? Thanks.
Posted by: Paul Gowder at Oct 19, 2008 2:17:01 PM
Addendum: googling, it seems like "price of risk" is being used to mean "price of capital." What an odd phrase: the borrower is paying the price, and it's not the borrower that's taking a risk, it's the lender... do financial market people always use that phrase?
Posted by: Paul Gowder at Oct 19, 2008 2:20:38 PM
He's talking about risk premiums (the rate of interest required for an investment - risk free rate). Tyler says those rates were too low.
Was the problem that risk premiums were too low or that there was significant systemic risk? Or both? Has anyone come up with good ideas on how to mitigate systemic risk?
Posted by: jsalvati at Oct 19, 2008 2:22:21 PM
Tyler,
If we replaced "risk" with "DVDs of Star Wars Episode II," would you still stand behind your four points? Or would your views on the price of this DVD be more of a huff-and-puff analysis saying supply and demand should determine it?
Posted by: Bob Murphy at Oct 19, 2008 2:42:25 PM
Lending was so cheap because of the way banks calculate the risk from each loan. The Basel I and II agreements say banks must keep 8% of risk-weighted assets on hand. There are all sorts of problems with this, but the main one is that it forces most banks to use the same techniques to measure risk.
When you use these econometric techniques, risk is underpriced in the boom, and overpriced in the downturn. Thus, banks are forced to keep more capital in the downturn when there is more "risk", even though cash may be harder to come by.
If we are gonna use risk weighted assets as the basis for bank capital requirements, there has to be a simple way to make this countercyclical. Perhaps this involves taking a 5 year moving average of the risk or somehting...
Posted by: stanfo at Oct 19, 2008 2:44:49 PM
Tyler, I apologize for being rude again, but if I may, there is a basic contradiction in your thinking. Systemic risk arises because too many people see the same thing as being risky at the same time. Therefore, when too many agents use the same risk reduction method, however sound the said method might be, global risk actually increases. Yep, that's the same line of reasoning as with liquidity traps. So trying to uniformly price risk is hugely counterproductive and the main engine of the problem. Financial markets - not goods markets, mind you - can only work through diversity. Otherwise, they just sink into deepening resonance. If you attempt to universally *price* risk, as Basel II did recently, as VaR did in 1998 or Portfolio Insurance did in 1987, you only generate wild spiralling oscillations. We *have* to get rid once and for all of the idea that risk is one good or concept. It is not.
Posted by: Henri Tournyol du Clos at Oct 19, 2008 3:01:17 PM
And in the time of overreaction is exactly the wrong time to subsidize further overreaction.
Posted by: Andrew at Oct 19, 2008 3:37:14 PM
I second "Henri Tournyol du Clos"'s post enthusiastically. The idea that you can stuff market agents' heads with the true model of risk in their heads and then make them use that model exclusively is a systemic risk increasing idea. (I am not so sure that Tyler advocates this though.)
I want to make an additional point. Whether "the risk" is overpriced or underpriced is less important if market actors realize the huge possible model errors and they consequently apply less leverage to their positions.
The more skeptical they are of academic risk models, the less leverage the rational(within reason) market actors will use.
All this is modulo perverse institutional incentives.
Posted by: Alex at Oct 19, 2008 3:42:15 PM
Tyler,
All previous comments make clear that we are having problems to understand what you mean by the price of risk. First, you talk about risk as if you were talking about tomatos (each unit is a perfect substitute of any other unit) but we all know that risk is an aggregate of different types of risks and even worse that each individual type of risk does not consist of units that are perfect substitutes. You should clarify at least what are the relevant types of risks, and if you include systemic or systematic risk as a type of risk, please tell us exactly what you mean by it. Second, you about talk about the price being low and high as if you had a clear criterion to determine it, but we all know that at most what we have are time series of asset prices and interest rates that we can relate to time series of other prices. Please try to be more specific about what you mean by low and high prices of risk and how you measure them. Third, despite the two problems I've just mentioned, I believe that what you're trying to say in #4 is right but this is because you add the adverb "excessively". I believe that ex ante people cannot agree on the meaning of adjectives as excessive, unsound, systemic, etc., and therefore they cannot be used to determine policy. Fourth, your last point is right, but you have been ignoring it in your previous posts because you have relied heavily on those adjectives.
Posted by: E. Barandiaran at Oct 19, 2008 4:31:55 PM
I am reading this to mean "risk" as a catchall term to include a market sentiment as well as the cost of taking on risk. For example 4 years ago, the FFR was artificially low, people were excited about rushing into buying, it was simple to create instruments like CDOs and CDS--and there was little regulatory burden.
Posted by: pants at Oct 19, 2008 4:38:22 PM
Well in your opinion the price of risk is too high currently
but until SYSTEMIC risk is drastically reduced then there is still a chance that many assets go to zero.
Posted by: will at Oct 19, 2008 4:57:44 PM
4. is certainly correct in that you use the word 'prevent.' 'Mitigate' is more realistic.
Posted by: meter at Oct 19, 2008 5:09:52 PM
As long as politicians protect the hucksters, we will have plenty of risk. If the politicians hadn't protected Fannie and Freddie from more regulation, we wouldn't be in the situation we are today.
Posted by: jorod at Oct 19, 2008 5:16:36 PM
Prevent the low price of risk? Easy. Stop subsidizing it.
Loans with ridiculous risk were made only because the lender knew they could palm off the loan to government financed entities that were blind to risk.
No need to regulate how risk is priced. Simply let it be priced. Stop messing around with the market and lo and behold perhaps it might be able to function.
http://blog.reinventdemocracy.org/2008/10/7-destructive-ideas-at-root-of.html
Posted by: Alan Brown at Oct 19, 2008 5:55:57 PM
The price of risk is a standard term in quant finance, where many simple models assume the expected return of any asset equals the riskfree rate ( ie treasuries) plus the price of risk times some estimate of the risk of the asset. This is an oversimplification, and most people do treat the price of risk as different for different types of risk, but you get the idea. You can see the change in the price of risk in the ted spread, emerging market debt spreads, junk bond spreads etc, most of which were at extreme lows before the crisis. It is generally measured by market indices like the above -- I agree with the commenters above that a diversity of risk models helps but in the end there is a market clearing price.
Brad DeLong has posted a lot on this.
Posted by: anonymouse at Oct 19, 2008 6:07:28 PM
Anonymouse,
Thanks for your comment. Two points. First, since it is an oversimplification, it should be used very carefully, in particular to discuss policy. Second, you say that the spreads were at extreme lows before the crisis, but you mean in comparison with past values. As you may remember the price of food had been declining for many years before it suddently increased in 2007. If the spreads have been declining for a long time, maybe they were not low. After all, financial innovation was supposed to reduce the spreads. My point is that is not so simple to say that the price is low or high.
Posted by: E. Barandiaran at Oct 19, 2008 6:26:48 PM
Current A-grade bond short-term yield: 5.4% (from Vanguard's fund for that category)
Current inflation rate (lower bound using corrupt methodology): 5%
Current real interest rate on short-term A-grade bonds: 0.4%
Claim: "Risk is overpriced right now."
Silas's reaction: HAHAHAHAHAHAHAHAHA! *rolling over*
Posted by: Person at Oct 19, 2008 8:06:54 PM
Does not pricing for all risk approach infinity?
I'd argue that all actions entail some amount of incalcuable risk, and most actions involve taking a deep breath and ignoring a bit of highly-unlikely risk. Some risk is both too unlikely and too difficult/expensive to insure against.
Posted by: Jeff Garzik at Oct 19, 2008 8:26:54 PM
This discussion seems silly to me until
(1) somebody defines the price of risk, and
(2) based on that definition, shows us a graph of the price of risk over time.
Then will might have a basis for discussion.
Posted by: zbicyclist at Oct 19, 2008 10:35:10 PM
Brad DeLong on the price of risk.
He comments on the astonishingly unreasonable equity risk premium and idly wonders:
[whether] It is time for the government to seize control of the price of risk and turn it into an administered price set by centrally-planning technocrats in the interest of social welfare--just as between fifty and a hundred years ago we decided that the short-term price of liquidity, the Bank Rate or the Federal Funds Rate, was too important to be left to the market and was turned into an administered price set by centrally-planning technocrats in the interest of social welfare.
Posted by: at Oct 20, 2008 12:07:29 AM
Are we talking about risk premiums? In what context?? Either way, to say that "4. We do not know how to regulate to prevent other, future, hitherto unexperienced manifestations of an excessively low price of risk" is, i believe, looking at the problem the wrong way. When the fed came in and "sponsored" fmae and fmac it pushed risk premium on mortgages to unnaturally low rates, making market equilibrium occur with the pooled risk lower then compatible with the overall underlying assets. It´s like getting a crack addicted bum on the street who´s asking for a loan. Then you define a risk premium for loaning him money. Then warren buffet says "I know this bum, it´s for a good cause, and i´ll sponsor him if he defaults". All of the sudden people are willing to lend to this guy at rates well below the inherent risk of the loan. Why? Nobody cares if the crackhead goes bust, so long as Warren is there to pay. So then Mr. Crackhead gets an AAA rating. You then have abnormal market returns sponsored by Mr. Buffet. Add leverage and matters get considerably worse. Now Warren Buffet has inadvertently created a mini crackhead credit bubble, and ultimately will have to pay the bill. Nevertheless, this bubble will stop growing once people start to doubt Mr. Buffet´s capabilities of sponsoring every loan. Now what would happen if there were (virtually) no limit to his resources, i.e., Mr. Buffet could, say, print money?
The problem of added regulation as a method of control is, therefore, that it is merely palliative. It would do the costly job imperfectly trying to correct the evils of the fundamental problem - government intervention in the market.
Posted by: Francisco Aboim at Oct 20, 2008 12:09:00 AM
Tyler's price of risk come from finance theory.
Consider the simple case of a stock vs a risk free asset and suppose the are only two states in the future: good and bad. In the good state the stock will be work 1.2 dollars and in the bad state it will be worth 0.8 dollars. Lets say each of these states are equally likely. On the other hand suppose there is a risk free asset which is worth 1 dollar in both states of the world.
Both the stock and the risk free asset have the same expected value, $1. It turns out that most people generally prefer to have the risk free asset rather than the stock all other things being equal (exception is gamblers). For this reason the stock generally trades at a lower price than the risk-free asset even though they have the same expected value. Thus the stock provides a greater yield than the risk free asset and this additional yield is the market price of risk. One easy way to measure the market price of risk is using the sharp ratio:
(risky yield - risk free yield)/(std dev of risky asset).
This tells you how much additional yield is required to compensate for the additional volatility of a risky asset.
Posted by: assman at Oct 20, 2008 12:14:08 AM
Another article by Brad DeLong in the same vein:
The Fed and the Treasury are walking down a road that ends with making the price of risk in financial markets, along with the price of liquidity, an administered price.
Posted by: at Oct 20, 2008 12:14:39 AM
Steve Randy Waldman at Interfluidity blog had a post a while back about how investors didn't take too much risk in recent years, they took too little risk.
He points out that the people who bought "AAA-rated" CDO tranches were looking for safe assets, and Wall Street put lipstick on pigs and created "AAA" toxic waste to feed this demand. Investing is the art of taking risks, and pretending that you could buy trillions of dollars worth of "risk-free" investments turned out to be the most enormous risk of all.
Posted by: at Oct 20, 2008 12:32:35 AM
We do have some indirect evidence from asset prices that 'the price of risk' moves with business cycles. See, for example all the current asset pricing models with habits in the preferences. Empirically, the 'price of risk' increases in recessions, and drops during booms simply from the statistical properties of asset returns.
There is no really compelling structural model of the finding, though.
Posted by: a at Oct 20, 2008 9:06:17 AM
