When E.F. Hutton Talks
The concept of utility is the most fundamental concept in economics. It
gets wrapped up in impressive sounding terms like “exogenous preference
functions”, and written in all sorts of arcane runes and formulas, but all
utility means is that you like something more than something else. The
assumptions that economic theory makes about utility are really pretty simple
and mostly about consistency — if you like vanilla ice cream more than
chocolate ice cream, and chocolate more than strawberry, then economic theory
assumes you also like vanilla more than strawberry — and continuity — if you
like one scoop of vanilla ice cream, then you like two scoops even more. But as
far as what you like, what your tastesor preferences are
in ice cream or music — or health insurance plans — economic theory is
intentionally silent. Economics is all about making rational decisions given
some set of likes and dislikes. It doesn’t presume to tell you what you should
like or dislike, and it assumes that you do in fact know what you like or dislike.
Or at least that’s what economic theory used to proclaim. Today economic
theory is used as the intellectual foundation for a political stratagem that
goes something like this: you do not know what you truly like, and in
particular you do not know your economic self-interest, but luckily for you we
are here to fix that. This is the common strand between QE and Obamacare. The
former says that you are wrong to prefer
safety to risk in your investments, and so we will fix that misconception of yours
by making it extremely painful for you not to take greater investment risks
than you would otherwise prefer. The latter says that you are wrong to prefer no health insurance or a certain
type of health insurance to another type of health insurance, and so we will
make it illegal for you to do anything but purchase a policy that we are
certain you would prefer if only you were thinking more clearly about all this.
Anyone who believes that this political maneuver is inherently a phenomenon
of the Left is kidding himself. The Right — in the form of sectarian or secular
authoritarianism that imposes behavioral politics on the justification that
this is how to get into heaven or demonstrate true patriotism — is no stranger
to exactly this sort of political aggrandizement. Nor am I arguing that it’s
smart to put your money under a mattress or that it’s wise to use the local
emergency clinic as your primary care provider. What I’m saying is that the
notion that we know your interests better than you know
your interests is inherently an anti-liberal position, whether it comes from
the Left or the Right. That’s liberalism with a small-l, the liberalism of Adam
Smith and John Stuart Mill, not Walter Mondale — a political philosophy that
argues for your right to be as stupid as you want to be in your personal
economic decisions.
While there are hundreds of examples of anti-liberal policies in the annals
of Western history, QE and Obamacare stand out in two important respects.
First, they’re big. Really big. Either policy on its own would be the
largest instantiation in human history of what the French call dirigisme, at least on an absolute scale. I suppose you
could argue that the US Social Security system has evolved into something even
larger, but that took 70+ years to match what QE and Obamacare have
accomplished in a few dozen months. I’ve written at length about the manner in
which emergency policy
responses to national traumas like wars and depressions are transformed into
permanent government programs, so I won’t
repeat that here. Suffice it to say that it’s not a coincidence that Social
Security is a child of the Great Depression in the same way that both QE and
Obamacare are children of the Great Recession. The institutionalization and
expansion of centralized economic policy is what always happens after an
economic crisis, but the scale and scope of QE and Obamacare, particularly when
considered together as two sides of the same illiberal coin, are unprecedented
in US history.
Second, and this is what really distinguishes the dirigiste policies of
today from those of the past, the political and bureaucratic advocates of QE
and Obamacare have co-opted the Narrative of Science to promote these policies
to the public. If you look at the financial media’s representation of monetary
policy during, say, the Volcker years, you see a curious thing. These articles almost never mention academic papers or Fed
research. Today you can’t go a week without tripping over a prominent WSJ or FT
article trumpeting this Fed publication or that IMF working paper as the
reason behind a monetary policy rhyme. The authority vested in the Volcker Fed
was based on a Narrative of Experience, an argument for trust based on a
representation of personal leadership and experiential wisdom. Today, the
argument for trusting the Fed places zero weight on the real-world experience
or personal wisdom of the Fed Chair. Instead, both Bernanke and Yellen are
presented as Wizards who channel the transcendent magic of economic theory. For
better or worse, a popular faith in Economic Science is the source of their
authority.
As for healthcare policy — the entire edifice of Obamacare has been
presented as a self-consciously scientific, enlightened economic argument. This
allows its political adversaries to be painted as bizarrely opposed to an
objectively correct scientific position, as either know-nothing rubes who probably
don’t even believe in evolution or as greedy stooges of the criminally
rapacious insurance industry. Contrast this to the media presentation of
healthcare policy initiatives in the 1960’s, particularly the establishment of
Medicare as part of Johnson’s Great Society. As the phrase “Great Society”
implies, arguments for Medicare had nothing to do with macroeconomic theories
of efficiency and everything to do with political theories of justice. All of
Johnson’s political initiatives, from Medicare to the Civil Rights Act, were
based on a Narrative of Social Justice, an explicitly political argument that
made little pretense of marshaling social science to prove the point. Seems
like a more honest mode of politics to me, one that recognizes and embraces the
hot-blooded nature of politics for what it is rather than hiding it within a
cool armor of Science, and maybe that’s why Johnson’s policies have stood the
test of time.
Why has the Narrative of Science been co-opted in this way? Because it
works. Because Science is the dominant religion, i.e. belief system in
transcendent forces, in the West today. Because politicians have always sought
to direct or tap into these belief systems for their own ends. In exactly the
same way that French kings in the 13th century used ecclesiastical arguments
and Papal bulls to justify their conquest of what we now know as southern
France in the Albigensian Crusades, so do American Presidents in the 21st
century use macroeconomic arguments and Nobel prize winner op-eds to justify
their expansionist aims. Economists play the same role in the court of George
W. Bush or Barack Obama as clerics played in the court of Louis VIII or Louis
IX. They intentionally write and speak in a “higher” language that lay people
do not understand, they are assigned to senior positions in every bureaucratic
institution of importance, and they are treated as the conduits of a received
Truth that is — at least in terms of its relationship to politics — purely a
social construction. I’m not trying to be flippant about this, but when you
read the history of the Middle Ages I find it impossible not to be struck by
the similarity in social meaning between clerics
then and economists today.
So why does this bug me so much? What's the big deal about wrapping a
political argument in the mantle of Economics in the same way that it used to
be wrapped in the mantle of Catholicism? Isn’t this what powerful political and
commercial interests have done since the dawn of time, drawing on some outside
source of social authority to support their cause?
Part of the answer is that as a limited government, small-l liberal I’m on
the losing side of this particular political argument. I believe that it’s
crucial to allow everyone to be as stupid as they want to be in their personal
economic decisions because a) economic vitality and growth in the aggregate
requires plenty of individual mistakes and losers along the way (sorry, but it
does), and b) the alternative — allowing or requiring government to make these
decisions on our behalf — inevitably creates a terribly fragile system where a
single poor decision can lead to permanent ruin. Is it difficult and at times
inefficient to maintain limited government in a mass society? Absolutely.
Should we make small exceptions to these liberal principles to grease the
wheels of effective governance in ordinary times, and big exceptions to these
principles in a national emergency? Without a doubt. I think Lincoln saved the
United States in 1861 when he suspended habeas corpus and imposed martial law
in wide swaths of the country. I think Bernanke saved the world in 2009 when he
implemented QE 1. But like the Roman dictator Cincinnatus, a great leader goes
back to the farm after he saves the Republic. It’s the hardest thing to do in
politics — to voluntarily relinquish emergency powers used wisely for the
common good, to maintain a personal humility and trust in the system in the
aftermath of great success. George Washington did it, and that’s why he’s the
greatest President this country ever had. I understand that it’s not terribly
likely we’ll ever see Washington’s like again — different times, different
world, etc., etc. — but hope springs eternal.
The other part of the answer is that using Science for political ends
subverts its usefulness (as does using Religion for political ends — just ask
Martin Luther). We lose something very important when we associate a particular
social scientific hypothesis with a winning policy outcome or a losing policy
outcome, and that’s the recognition that social science — particularly economic
science — is never True or False, but only more or less useful depending on
whatever it is in life that you value — your utility function. Both as
individuals and as collectives, we can achieve much greater levels of utility —
we can be happier — if we maintain this agnostic view of Truth when it comes to
social science. Politicians want to sell us on the notion that they have The
Answer, that they can deliver the good life if only we keep them in power.
Social scientists — or at least honest ones — recognize that there are no
Answers in the patterns and relationships they identify, even if those patterns
can be written in the highly precise language of mathematics. There is More
Useful and Less Useful in social science — that’s all — and claims to the
contrary detract from the very real benefits and advances that social science
can provide.
Here’s a concrete example of what I mean —
Let’s say that you’re interested in wealth maximization, that this is the
utility function you are most concerned with as an investor, and you want to
know what percentage of your wealth you should allocate to the different
investment opportunities you can choose from. Paul Samuelson, the most
influential economist of the post-World War II era and the first American
winner of the Nobel prize in Economics has an answer for you: EU=Max(E(1+r)α/α). Translation:
the more confident you are in the expected return of the investment choice, the
more you should allocate to that choice, but in a more or less linearly
proportional manner. On the other hand, Edward Thorp, author of “Beat the
Dealer” and evangelist of the Kelly Criterion — an algorithm designed by
mathematician John Kelly at Bell Labs in the 1950’s and used by investors like
Warren Buffett, Bill Gross, and Jim Simons (if you’ve never read “Fortune’s
Formula”, by William Poundstone, you should) — has a different answer for you: EU=Max(E log(1+r)²). Translation, the more
confident you are in the expected return of the investment choice, the more you
should allocate to that choice, but in a logarithmically proportional manner.
The difference between investing on the basis of linear proportionality and
logarithmic proportionality is vast and incommensurable. With the Kelly
criterion, even a small expected advantage in the investment odds — say a 52%
chance of doubling your investment and a 48% chance of losing it all — requires
you to invest a significant portion of your overall wealth, in this case about
2%. With a larger expected advantage in the investment odds, the recommended
allocation gets very large, very fast. If the odds are 60/40 on doubling
up/losing the entire investment, Kelly says invest 20% of your total wealth; if
the odds are 80/20, Kelly says invest 60% of your total wealth in this single
bet! Definitely not for the faint of heart, and definitely a far riskier
strategy at any given point in time than the straightforward Samuelson expected
utility approach. But you never lose ALL of your money with the Kelly
criterion, and over a long enough period of time (maybe a very long period of
time) with infinitely divisible bet amounts and correct assessment of the
investment odds, the Kelly criterion will, by definition, maximize the growth rate of your wealth.
These are two VERY different answers to the wealth maximization question by
two world-class geniuses, each with a legion of world-class genius supporters.
Samuelson is a lot more famous and received far more public accolades; Thorp
made a lot more money from investing (Kelly died of a stroke at age 41 in 1965
and never made a dime from his theory). But they can’t BOTH be right, the politician
would say. What’s The Answer to the wealth maximization question so we can
institute the right policy? Well — they ARE both right, there is no Answer, and
the correct choice between the two depends entirely on your individual utility
function. In fact, choosing either wealth maximization algorithm and imposing
it on everyone is guaranteed to make everyone worse off in the aggregate.
How’s that? Let’s say I’m investing my life savings, and I’ve only got one
shot to get this right. Not one investment, but one shot at implementing a
coherent investment strategy for this, the only life’s savings I will ever
have. If that’s my personal situation, then I would be nuts to choose the Kelly
criterion to drive that strategy. It’s just too risky, and if I’m unlucky I’ll
be down so much that I’ll hate myself. Maybe in the long run it maximizes my
wealth growth rate, but in the long run I’m also dead. On the other hand, let’s
say I’m investing a small bonus. It’s not the only bonus I’ll ever receive, and
in and of itself it’s not life changing money. If that’s my personal situation,
then I would be nuts NOT to choose the Kelly criterion because it has the very
real possibility of transforming the small bonus into life changing money.
No one’s utility function for money is linear — $20 has more than 20 times
the utility to me than $1 — and no two people have the same utility function
for money — I’m sure there are people out there who care as little about $20 as
I do about $1. Everyone’s utility function for money changes over time, and
most are contextually dependent. It is impossible to
design a one-size-fits-all wealth maximization formula, which is why human
financial advisors have such an important job. It’s also why government efforts
to force us to converge on a utility function for investment choices,
healthcare choices, or any other sort of personal economic choice result in
such a widespread gnashing of teeth and popular dissatisfaction. At best, it’s
a myopic conception of how to generate more economic utility. At worst, it’s an
intentional subversion of useful social science to cloak politics as usual. In
either event, it’s something that deserves to be called out, and that’s what
I’ll keep doing with Epsilon Theory.
Postscript
Two quick points on portfolio management, utility functions, and the Kelly
criterion that I’ll present without elaboration and will probably only be of
interest to professional investors who are immersed in this sort of thing.
1.
In several important respects, risk parity investment allocations are to
60/40 stock/bond allocations what the Kelly criterion is to Samuelson expected
utility.
2.
The allocation of capital by an investment manager who wants to establish
multiple independent Kelly criterion strategies across traders or
sub-investment managers, each of whose individual utility functions favors a
fractional Kelly or Samuelson expected utility function, is a solvable game.
No comments:
Post a Comment