The Religion of Rationality

Academics in the social sciences rely on theories and models of human behavior.  For academics whose entire career and livelihood depends on their theories, events that challenge their theories threatened their existence.  Thus, there are powerful incentives for those whose theories are challenged to explain away the events that undermine their credibility.

A foundation of classical economics is that people make rational decisions, or at least enough people make enough rational decisions such that their rationality nullifies the irrational.  From time to time, however, real events challenge this assumption.  In financial markets, these events are known as “bubbles,” which seem to be occurring with greater regularity. (And given the recent actions of the authorities, there is no reason to suggest this will change anytime soon.)

As one might expect, the revisionism and rationalization of the housing bubble amongst academics who say that the market was not irrational is already occurring.  At the blog Causes of the Crisis, Jeffrey Friedman argues that home buyers were not being irrational.  Why?

[I]n reality, very sober housing experts and economists (including Ben Bernanke) agreed that there was no bubble, and there was nothing “irrational” about this consensus.

First of all, there were several economists who thought otherwise, including Nouriel Roubini and Robert Shiller, as did investors such as George Soros and Jim Rogers, as well as academics such as Nassim Taleb.  Heck, even us shmucks at Running of the Bulls could see the insanity of the housing markets, and we certainly do not offer any pretense of being particularly wise and intelligent!  That we got it right and all these “experts and economists” got it spectacularly wrong is more cause for shame and embarrassment than them actually getting it wrong!

The fact that a “consensus agrees” does not make it correct.  By this logic, there was nothing irrational about the Nasdaq trading at 100x earnings when it peaked out in March 2000 at 5132 because the market “consensus” said that was what the Nasdaq was worth.  Many investors and economists at the time also argued that the Nasdaq was not a bubble.  This is not only specious reasoning, it is dangerous both to the economy and to your own pocketbook and well-being.

Friedman goes on.

After all, housing prices had risen steadily since the Depression as the U.S. population grew and its affluence increased. Why shouldn’t people keep buying bigger, more expensive homes? There had never been a significant nationwide housing bubble, and housing speculation in “hot” markets such as Miami and Las Vegas could be quite rationally dismissed as localized.

Friedman assumes that people logically processed these facts and rationally made decisions based on this information.  In fact, these were the reasons why people acted irrationally.  Because people believed such, they had a foundation for throwing caution and reason to the wind because they believed they could not lose.

I often wonder if the academics who purport this view ever actually talked to anyone who was involved in the housing market.  Yes, some were rational but many were not.   Amongst one of many examples I encountered, my friendly but uneducated neighbor who knows nothing about investing would become excited about the fortune he was going to make investing in coastal real estate in 2006, and he would cite these reasons in his excitement (not to mention other “truths” that are always bandied about at the top of the real estate market, such as “land is limited”).

The idea that this behavior is “rational” comes from the mistaken belief in financial economics that Larry Summers described as “ketchup economics,” whereby if two 16 ounce bottles of ketchup are priced the same as one 32 ounce bottle of ketchup, the price is “right” and the market is rational.  This construct excludes concepts of intrinsic value and comparisons to price. 

Using Friedman’s logic, economists and investors argued that Cisco at 100x earnings in 2000 was “rational.”  After all, people had seen tech stocks rising to astronomical heights, with tales from friends and neighborhoods of instant wealth in tech stocks.  The advances in technologies, the story went, meant astronomical profits for tech companies.  Talking Sock Puppet Companies were “rationally” valued at over $1 billion and 100x sales because people would buy endlessly buy pet food over the Internet.  It is hard to believe that anyone who lived through that period could take such logic seriously.

In the above quote, Friedman makes no mention of the intrinsic value of a house.  One would think that if one were rational, one would undertake a discounted cash flow valuation of the cash flows of owning real estate relative to renting, estimate a sale price, then make a decision based on their calculation.  The number of people I know who did this was zero.  Instead, all people usually cared about was if they could afford the monthly payment. Homebuyers rarely calculated the value of the house and whether or not the price they were paying made sense relative to the intrinsic value of a home.

To the “rational” investor, the price of the asset will approximate its intrinsic value.  When the price becomes unhinged from its intrinsic value, the asset becomes irrational.  This is what happens during bubbles, including the housing bubble.

In any given market, there are always rational and irrational agents.  Markets tend towards rationality over time but are not always rational at any given time.  The level of irrationality in the market is a function of the both the number of irrational investors as well as the intensity of the irrationality.  Irrationality manifests into momentum.  Momentum investing is not always irrational but irrational markets are dominated by momentum investors. Momentum investors care not about intrinsic value.  They only care that the market is going up.  100x earnings?  Who cares?  Losses forecasted for a decade?  Doesn’t matter.  It is the price action that matters.  In a momentum-driven market, nobody is rationally calculating the value of an asset.  It is totally irrelevant.

I have spent nearly 20 years in capital markets, and I am always amazed when people tell me that irrationality does not drive markets.  I often wonder if these people have any actual experience at all. Markets are often driven by greed and fear.  The idea that people were rationally selling stocks at the bottom in March is laughable.  They were scared out of their mind. 

……………………..

Friedman attempts to debunk three myths of the crisis, of which the irrationality of markets is supposedly one.  He sets about two other so-called myths, executive compensation and the failure of capitalism.  I am not going to comment on his other two identified myths except for one – Friedman’s comments on MBS securities purchased by bond managers. 

Perhaps the most powerful evidence against the executive-compensation thesis, however, is that 81 percent of the mortgage-backed tranches purchased by banks were rated AAA[5], and thus produced lower returns than the double-A and lower-rated tranches of the same mortgage-backed securities that were available. Bankers who were indifferent to risk because they were seeking higher return, hence higher bonuses, should have bought the lower-rated tranches universally, but they did so only 19 percent of the time. And most of those purchases were of double-A rather than A, BBB, or lower-rated, more-lucrative tranches.

Now, I have not read the piece Friedman references, so perhaps I am wrong.  However, it appears that the author may be making incorrect assumptions about portfolio construction when drawing this conclusion.  His conclusion is only true if the portfolio manager’s choices are limited to mortgage investments, and if the asset allocation of mortgage bonds was at or below policy targets.

Some institutions have a target allocation towards credit ratings.  If a bank has an allocation towards AAA-rated securities, the money manager will often view AAA-rated mortgages the same as AAA-rated Treasuries since both are rated AAA.  However, even though both may have been rated AAA before the crisis, yields for AAA-rated mortgages were 20 to 40 basis points higher than AAA-rated Treasuries.

Managers are compensated based on their returns relative to their benchmarks and to their peers.  If a manager has a pool of AAA-rated securities from which to choose, he will attempt to pick up yield by shifting his allocation away from Treasuries to mortgage bonds because if he does not, then he will lag his peers and perhaps the benchmark.  Thus, it may not be correct to look at the breakdown of holdings within the different tranches of MBS in isolation, as Friedman appears to do.  Instead, the correct analysis would be to look at the breakdown of holdings within the allocation towards each rating.  How many more mortgage bonds were held relative to Treasuries compared to the past?  I do not know the answer to this, but my guess is that given that securitization exploded this decade, and managers probably shifted fund from Treasuries towards other AAA-rated securities such as mortgages.  And as we now know, many of those AAA securities were garbage.

Likewise, money managers will make decisions based upon policy targets.  If policy targets shift towards holding more structured products, the money manager will be forced to buy more structured products.  Again, I have no empirical data to back it up, but it is not unreasonable to assume that there was a policy shift towards structured products at institutions this decade as derivatives and structured products became more prevalent within fixed income. 

Unless I am wrong and the secular shifts in structured products are accounted for in the data Friedman cites (and I will acknowledge such here if shown otherwise), the composition of holdings within the mortgage market does not tell us anything about riskiness of portfolios and thus executive compensation.

Comments

[dacian]

What is left to say about this writing? Excellent! Thanks T.

[Dark Space]

The other kicker is that about 81% of the MBS market was originally rated AAA - so if you bought the entire risk spectrum of the market, you'd have the same allocation. Next largest segment was AA, then A, then BBB - I don't think he has ever looked at a MBS deal. That's just how the deals were structure - a 20% subordinated AAA, is precisely 80% of the total deal's size.

Another point that everyone just dismisses is that a lot of bankers/traders are simply not smart enough or experienced enough. I can't tell you how many traders I've worked with two years out of B school at some shiny university that didn't know what an MBS was, or even basic bond issues like accrued interest. Sure, they picked it up enough to trade their billion dollar book or they would have lost their job, but to this day they still cannot stand up to any rigorous questioning about the very products they trade.

[domus]

One sentence only: without rationality assumptions, no theory can be proposed. Behavioral economists assume that behavior happens because....it happens. If you like it better that way, suit yourselves.

Can't lose your head when things don;t go your way.