Can a Computer do Your Job?

Physicist Steven Hsu has several interesting observations on the selection process of colleges and elite firms. One little aside on college admissions I found very amusing:

When I was on the faculty at Yale I knew people in admissions and it's not clear to me that they were the best able to spot potential in 18 year olds. In studies of expert performance admissions people are less good at predicting UG[undergrad] GPA than a simple algorithm. (The "algorithm" is simply a weighted sum of SAT and HS GPA!)

When I worked at a bank, I remember we had a bunch of 'underwriters', and their job was to evaluate the creditworthiness of loan applicants. Coming out of grad school, I had no experience with this, and they had this very complex, holistic methodology. Later, when I became head of capital allocations and we started quantifying their portfolios, I invariably found they could be replicated via some pretty simple rules. This got me excited about going to Moody's to help develop their new RiskCalc Private firm model, which today dominates private firm underwriting worldwide. Anyway, the simple trick was always the same:

1) identify the key indicators
2) transform them appropriately (eg, turn 'profits' into percentile for profit/assets, or some sigmoidal transformation like exp(x)/(1+exp(x))
3) weight or crosstab the indicators
4) add 'em up, and transform output into meaningful buckets used for pricing (200 bps) or risk classification (eg BBB)

Risk is invariably on a log scale, so usually the ordinal rankings correspond to exponential increases in expected default rates.

Anyway, with hindsight, I found most of the facade put forth by various departments (eg, auto lending, health care lending), was very misleading. Everyone made the simple complicated. I think deep down, no one likes to think a computer can do their job, and there are many instances where exceptions matter, so a great deal is made out of these special cases. Yet the false positives made them great anecdotes, but horrible for generalizations. Thus, simple rules dominate their much more costly, confusing, and non-quantitative product created by teams of analysts.

These jobs seem rather common. For example, In the nineties many 'traders' I knew would simply buy from a customer at the bid, and then sell to another customer at the ask, and go home 'making' $10k a day. They would then get a $500k annual bonus. A college admissions director probably interviews people all day, and writes pages of material supplementing their decisions, but all to trail a simple linear rule. I also worked in an econ department, where we created tons of forecasts with lots of commentary, all dominated by vector autoregressions. Our asset/liability committees would often have a lot of commentary about where various interest rates are going, as if these forecasts were any better than what could be lifted from forward prices.

People in these situations are rather quite pathetic. Many are really good people, and would take a big pay cut if they started over, so its not as simple as just choosing a new career.

The Office Test

Tyler Cowen supports his hypothesis that things haven't changed as much recently as they did in prior generations by looking at his kitchen.

Here is Paul Krugman, noting that innovations for the kitchen have slowed down. He cites this earlier, mid-90s piece of his on kitchens, which I would have cited had I known about it. His conclusion:

By any reasonable standard, the change in how America lived between 1918 and 1957 was immensely greater than the change between 1957 and the present.

As Krugman did in the mid-1990s, I now cook in a 1950s kitchen and it suits me fine.

My anecdote refutes his! Look at your office. From 1918 to 1957, it probably had slide rules, typewriter, desk, phone. But in 2011 we have computers, which are filled with magical programs that do things unimaginable in 1957. After all, in 1954 the head of IBM said "I think there is a world market for maybe five computers."

Cowen's Great Stagnation, von Mises's Great Solution

Tyler Cowen recently wrote a $4 e-book, The Great Stagnation, a book enjoyed by Reihan Salam on the right and Matthew Yglesias on the left. Alas, missives with such broad appeal are like David Broder or board games that are ‘fun for the whole family’: lots of praise, mainly due to ulterior motives. In this case, his even-handed criticism of the right and left allows anyone with a little nuance to claim a well-know objective economist agrees with his particular spin, always useful for pundit who has a daily word quota.

The book is motivated by our current slump, and as von Mises noted in Human Action (p441):

Nothing has harmed the cause of liberalism more than the almost regular return of feverish booms and of the lingering slumps.

That is, most observers focus on the recessions, and try to play these into broader themes. Their magnitudes pale in comparison to the secular growth rates that over generations differentiate us from Ghana and other impoverished countries. They are as relevant to our nation’s long-term growth as a flu is to the growth of a child. They breed counterproductive remedies like the National Industrial Recovery Act of 1933, or the endless extensions of unemployment insurance. The cumulative effect of our remedies is to increase the scope of government in every dimension of the economy and is probably the main reason for our productivity slowdown.

Here’s Cowen’s theme: we live in an unusual era of lower productivity, and the financial crisis was merely a symptom of this broad trend we are just now realizing. Here’s the key graph:

In a figurative sense, the American economy has enjoyed lots of low-hanging fruit since at least the seventeenth century: free land; immigrant labor; and powerful new technologies. Yet during the last forty years, that low-hanging fruit started disappearing and we started pretending it was still there. We have failed to recognize that we are at a technological plateau and the trees are barer than we would like to think. That’s it. That is what has gone wrong.

I think the housing bubble was in many ways a singular event, though a special case of a catastrophic regime collapse as outlined in my Batesian Mimicry model of business cycles. This theory posits that recessions are always different because misallocations can only occur where they are unexpected, and by definition unexpected things are somewhat unique. The housing bubble is not very relevant to the post-1975 slowdown in productivity growth.

Despite Tyler's depressing diagnosis, he notes some good news. A lot of value is being created that isn’t priced, so maybe everyone should take comfort in that. For example, in the mid 70's productivity growth started to decline, but at the same time there started an explosion of free porn created by the VCR and the internet (not his example, but serves his point). Is that a wash? Maybe measured GDP really is rising as fast as ever, we just don’t count it well anymore because so many things are often free. Perhaps, but you could say the same thing in the past, as with pasteurization or radio.

Another bit of good news according to Cowen is that there is some low hanging fruit in those diseased fields of health care and K-12 education. For example, he mentions that:

we now see a critical mass in the American electorate favoring concrete steps to bring greater quality and accountability to K-12 education...Obama has opted for an education policy that, on the whole, the teachers union strongly dislike.

I do not see any big education changes around the corner. Obama’s weak support of merit pay is going nowhere, and would not make a difference if implemented. The only real merit pay structure that works is the one used in the private sector: decentralized, unstructured reviews by a boss who has a financial stake in the employee’s success, and there's a 5%ish chance the reviewed employee actually gets fired because of it. This would be considered unfair by teachers unions so instead we have pure seniority, or merit programs such as in Minnesota where 99.94% of all teachers garnered the carrot. My kids learn more math at their $100/month Kumon program than in public schools and I live in Money Magazine’s Best City in America, so clearly our above-average school system can do better pretty easily. The key is having a curriculum where kids move through a set of exercises that build upon each other, and only if they have mastered the previous material. This maximizes their potential, but everyone moves at different rates. That would lead to massive tracking of students by ability and the demographic inequality exposed would be politically untenable.

As per health care, it is already highly regulated, and becoming more so. Highly regulated sectors are rarely set free again, rather, it’s the whack-a-mole strategy of fixing a problem created by the last fix which then creates a new problem. In Santiago Chile the bus system worked, but it was a private system that operated for profit, and there were the typical problems that seemed wasteful. The solution seemed like a simple two-fer: run it for public interest (lowering cost by getting rid of profit), and regulate it to make it safer and less polluting. The result has been a nightmare, as the average commute immediately goes from 40 minutes to an hour and 40 minutes. Yet there is no general call to go back to privatized busing, rather, rearrange management or add ‘better’ regulations. Similarly, our health care system will be not be improved by adding more mandates and regulations, but we aren’t going back to 1950 when 90% of our health care expenditure was out of pocket and decentralized consumers disciplined providers.

There are lots of little independent comments in the book. For example, there’s Cowen's call to 'raise the social status of scientists,’ presumably, because that leads to more truths, and with more truths, everyone can act in a more consistent fashion. You can’t exogenously raise a profession’s status. It’s like trying to make your kids more popular, at best ineffective. If scientists continue to produce the twaddle like Marxist sociologists, economists who can't agree on whether the minimum wage is a good idea, and our best physicists spend their time elaborating a theory that has no foreseeable testable implications, they get what they deserve.

But back to his theme, that previous higher growth rates were based on low hanging fruit. What were these in our past?

1) Free Land. Great, but, we have same productivity as Liechtenstein, Hong Kong, Japan, Singapore, Iceland, and they didn’t have much land. Free Land would explain the growth rates only if US productivity growth was unique, and it is not.

2) Technologic breakthroughs. Electric lights, the internal combustion engine, the typewriter, radio, fertilizer, transistor, were all big boosts to productivity. No more. Now all we have are Teflon and Tang. Life, he says, is not much different than in 1953. Well, reading Plato, life hasn’t changed all that much since Ancient Greece, but forget the semantics. I would say my job is incomprehensible to someone 30 years ago. Further, while I used to play with plastic army men and play new-fangled 45 records, my 3-year old uses my wife’s iPad to decorate princesses and play music. Her childhood is probably as different to me as me to my dad, and my dad to his dad, etc.

3) Smart uneducated kids. In the past, lots of smart kids did not go to college, and as they did, we generated easy returns to their education. Now, almost everyone smart enough goes to college, so there’s no more easy gain there. I’d say this one is perhaps true, but if you look at its affect on growth, I would say it has been small. Most economic historians start the Industrial Revolution as starting around 1750, for the next century we had this massive, unprecedented rise is productivity without a big, broad increase in formal schooling. Formal education mainly produces imitation and routine, not innovation and growth, when not engaging in pure blather.

Cowen’s way of thinking is rather common, that growth emanates from some limited resource, and like ‘peak oil’ goes through a natural cycle of growth and decline. The Physiocrats in the eighteenth century believed that the wealth of nations was derived solely from the value of things from soil like minerals and agriculture, all the rest—advertising, management, finance—just parasitic. Such thinking influenced Malthus’s idea that we are all doomed by a finite amount of land, which will ultimately constrain our population via starvation. Later the labor theory of value tried to make labor the source of all wealth, where capital was disembodied labor. This too is mistaken, and while no one promotes these theories anymore, their intuition lives on.

I remember working for a broker in the summer of 1989, and the principal of the firm was writing a 'big idea' piece for his clients. He noted that unlike the past, we had no engine of growth like the automobile, so times would be tough--unaware he was at the cusp of the computer revolution, highlighting that it is difficult to see major trends without the benefit of hindsight.


It is important to recognize the essential drivers of economic growth, and much of the following is based on the works of Ludwig von Mises and Frank Knight. Economics is not about mainly about technology or resources, but about people and their actions. The market economy is a social system in which individuals are constantly trying to better their situation. The most enterprising individuals are driven to earn profit by doing something new and different. Most innovation is unplanned, and people outside the industry, especially users, make the majority of significant innovations by revealing their preferences, and this is a theme of Amar Bhide's work on entrepreneurial consumers.

Before the first industrial revolution around 1750, economic life in Europe was unprogressive, and its organization collectivist. The arbitrary administration of rules was not conducive to large-scale accumulation of capital prior to the first industrial revolution. Pre-industrial revolution business was imbued with the inherited spirit of privilege and exclusive monopoly, its philosophy was restriction and the prohibition of competition both foreign and domestic. The establishment of individualism was the result of the desire for improvement, even though it was not a conscious choice of political leaders.

Growth is merely the effect of free people continually improving their situation by finding better ways to solve problems, in the process creating profits, which are then used as capital to invest in greater efficiency. When profits direct economic activity, growth naturally occurs. When we instead try to service needs without regard to profit metrics or consumer preferences--as in pre-college education today in the US--we get stagnation.

Profits appear only in an economy in which the masses standard of living improves, because it is the result of taking inputs and creating products and services more valuable than their inputs. Profits are wealth creation by definition, and they only measure the producer surplus, not the consumer surplus, so they understate value created for society. To forgo them is to reduce aggregate wealth. Aggregate growth is not a result of having an awesome endowment, but rather the incessant urge of individuals towards improving their situation. One only has to look at the oil states (Nigeria?) vs. those economies in Asia without basically any raw materials. The most enterprising entrepreneurs are driven to innovate to earn profits by readjusting again and again the arrangement of productive activities so as to fill in the best possible way the needs of the consumer.

The state does not create much wealth directly--exceptions being public goods like roads and dams--but acting as the agreed-upon social apparatus of compulsion and coercion is essential for codifying and enforcing rules. When it becomes more advantageous for businessmen to rely upon the aid of those in political office than upon the best satisfaction of the needs of consumers, profits are no longer sustainable and productive, but rather their source is the compulsive redistribution of the state. Ethanol mandates and sugar quotas create profits, but only because they coerce consumer choices, as opposed to creating products and services that people actually prefer.

A big problem with modern macroeconomics is the fallacy that underlies the ‘fallacy of composition’, that is, the idea that saving hurts the aggregate savings via the multiplier. This is the hare-brained idea that if we spend 1 billionth of our GDP on bobble-head dolls, then via equilibrium reasoning, if we create $1 dollar in bobble head dolls, this creates $1 billion. As government spending is basically an investment in our future, government equals investment (G=I) to many, and instead of bobble head dolls we use investment (which is the same as government spending, remember), so you have to keep increasing G to avoid a meltdown whenever we are below full employment (which has been ‘always’ in my adult life). Funny how things worked better, in terms of growth rates economists pine for, back in the 19th century when the tools of fiscal policy were imperceptible and monetary policy absent.

Think about the distinction between paying someone welfare, and paying them to dig holes and fill them up. In terms of creating wealth for consumer, they are identical. A Keynesian (eg, Romer-Bernstein 2009) model assumes that a 1% increase in government spending—on just about anything—leads to a permanent 1.5ish percent change in GDP, which makes sense only via the magic of the multiplier, plus the faith that marginal government spending is actually creates wealth if you measured all the spillovers. Most government work does not produce benefits above their cost, and often, in the form of preventing people from freely contracting, destroys wealth (eg, financial regulators encouraged the decline in mortgage underwriting standards via their consistency with Fannie Mae’s ubiquitous underwriting software, and the Community Reinvestment Act’s mandate of lending to historically underserved communities, etc. The whole mass of financial regulators--OTS, OFFHEO, Fed, FDIC, OCC, SEC, CFPA--could surf the internet all day and no one would notice).

The system should be optimized for the whole and for long run, not the parts and the fluctuations. Microstability should be sacrificed for macrostability. The stability of the whole depends on the instability, and resultant resiliency, of the parts. When industries are protected from recessions, the system does not grow as quickly, and the productivity of the whole is sacrificed. The path back to long-run growth rates of the past is a government that was as intrusive as back then. In 1900 it was 3%, in 1950 it was 23%; now it is around 40%. The number of pages added to the Federal Register, which lists all new regulations, reached an all-time high of 82,590 in 2010, up from 9,562 in 1950. These rules create hidden costs often far beyond their direct costs, in the way that medical malpractice lawsuits are insignificant by themselves (1% of health care expenditures), though they significantly affect medical protocols (eg, my wife had Caesarians for each of our three children).

High unemployment is the result of a misallocation exposed, which happens often quite suddenly as in the 2008 housing debacle. It is not a paradox that wisdom of crowds, like the wisdom of individuals, is subject to error, as you can find people with very different purposes pre-2007 thinking that no-money-down housing is just and profitable--even Robert Shiller only could muster the prediction in 2005 that housing in some areas might decline, consistent with the erroneous assumption that a broadly diversified portfolio of mortgages had infinitesimal risk. We invested too much in housing, so that misallocation must be rectified, meaning, lots of people need to get different jobs, and the first step in this process is losing their old jobs. They need a market shove. It's regrettable, but recessions are inevitable given human fallibility (this is a deviation from von Mises).

The economy that grows best has the most freedom, and is built as a by-product of profits. As freedom and wealth are all good things, there must be a catch, and it is a big one: inequality. People hate inequality more than they love wealth, which is why people would prefer to be above average in a poorer world rather than below average in a richer world. Profits accrue to a minority of individuals via their ownership of capital, and the distribution of these profits among demographic groups is highly auto-correlated year after year. Most people are not in this fortunate minority, do not like it, and in a democracy their preferences determine policy. Sure, the German, Irish, Italian, and Jewish immigrants were once relatively poor in the US, and now not, but that took a long time. People want equality now, which means redistribution and patronage jobs (as Shirley Sherrod of the USDA said, ‘Have you heard of anybody in the federal government losing their job?’).

Most politicians see profits as a vice, not a virtue. The idea that profits primarily reflect an unfair system, as opposed to growth, leads us to hinder the source of our growth. Profits are often seen as symptoms of either thievery, or an irrelevance to what really matters, things like global warming or social justice. Recessions and poor growth are correlated with low profits--the 30's and 70's were bad for profits and asset prices--yet solutions often center on having the 'rich sacrifice', as if taxing the source of our productivity improves things. This only makes sense if you see profits not as the source of productivity, but the result of powerful people taking from an exogenous source of income.

Freedom and equal outcomes are incompatible, yet as the socialist economies showed, restricting freedom does not eliminate or even decrease inequality, it merely makes it less conspicuous and more onerous. That John Paulson made billions of dollars last year only bothers someone if 1) they are feeling petty envy or 2) they think this was merely taken from others. Most do not admit to the former, and rationalize using latter. People I do not know who truly annoy me use the force of law, as when they tell me where I can send my kid to school, tell me I am unqualified to invest in hedge funds, determine who can cut my hair, and what my medical insurance must provide.

The economy is like an ecosystem, a complex adaptive system, and the key to sustained growth is letting people with the requisite information and incentives engage in activities that are appreciated so broadly, his service or product will be needed again and again, a sustainable pattern of specialization and trade (see Kling). It is impossible to predict what fields in an economy, or species in an ecosystem, will prosper, but it is a sure bet that it will find its most sustainable, consistent equilibrium if one merely leaves it alone. No one thinks you can improve a rain forest via top-down Federal programs, but those same people find it obvious that more Federal regulations and purchases will improve a market economy, though it is just as complex and natural.

When people interact freely, they are directed by profits, and this creates something out of nothing, productivity. It does not take any specific technological or material endowment, merely liberty. There is a lot of low hanging fruit because leaving people alone is eminently feasible, but it is highly unlikely because our democracy does not trust the market to produce trickle down results sufficiently fast or fairly distributed. 'Trickle down' is referred to derisively, as if advocates are disingenuous about this patently absurd theory. This very nonintuitive idea--the invisible hand, that individual freedom maximizes growth--hatched the field of economics, which unfortunately has lost its way, lured by the hope of becoming a dentist to the economic patient with a cavity; instead, macroeconomists, by focusing on federal stimulus plans, are like modern-day bloodletters, counterproductive. Hopefully, such delusional expert folly won’t last as long.

What Top Level Sociology Looks Like

I read the devastating review of Erik Olin's latest book, Envisaging Real Utopias. That motivated me to watch his video on his latest book, and it contains anecdotes supporting his vision of people working in the Marxist ideal: "from each according to his ability, to each according to his need". To highlight he's in the mainstream, here's Lane Kenworthy, a sociologist at the University of Arizona, arguing we need more welfare of all sorts. There's a simple, egalitarian theme to their findings. They give a good example of what top-level sociologists are doing

Olin notes Wikipedia is based on his ideals and works, ergo...I guess, why not have General Motors and Google run that way! He's a fan of neighborhood participatory budget assemblies, as local communities serve the commonweal without the inefficiency and duplicity of Washington. He didn't mention it, but as he's a good liberal (Marxist, actually) I'm sure he's against states' rights, or cutting state taxes to replace with local property taxes, and for all sorts of Federal regulations. See the PowerPoint slides here. I generally find the highly abstract, and parochial proofs in economics to generally be a waste of time, yet they are superior to theories presented as a flow chart.

He's the president of the American Sociological Society. I'm sure he's a smart, thoughtful guy. He's just worked himself into an intellectual cocoon where his silly talk seems respectable. As a 'scientist', he presumes his simple biases are actually fact-based, but he is so selective in his view of the data, it highlights that education mainly allows us to articulate our prejudices better. He highlights that one can be an intellectual, respected by one's peers, have an esteemed affiliation, and be totally clueless.

Adverse Timing Kills Returns

The average return of an asset is merely looks at the price. One can argue whether geometric or arithmetic averages are best, but let's ignore that here. Look at the firm ESLRD, Evergreen Solar. They make solar power cells, so it's a sexy product, one very amenable to sales pitches of the next black swan that rubes do not fully appreciate. It came out in November 2000, with $112MM of insider money, and $42MM from outsiders via their IPO. The price (split adjusted) was $84.

The price fell over the next couple years, then rebounded in 2006 and 2007, and lately collapsed, and is now trading at $2.50. They have never made money. In 2006 revenue peaked at $103MM, up sharply from 2005, which instigated the bounce in stock around that time. Alas, it did not continue, and revenue has fallen, and it appears they will never be viable (current market cap is only $87MM).

Unfortunately for investors, their total of $446MM worth of injections were not random, but rather at relative stock highs, such as in 2007 and 2008. Look below and the red lines are investor contributions, and the black line at the end is their total value of those red lines today. If you look at the Internal Rate of Return, you get a very different picture than the top-line average return. The internal rate of return on an investment or project is the "annualized effective compounded return rate" or discount rate that makes the net present value (NPV) of all cash flows (both positive and negative) from a particular investment equal to zero.


It's especially useful when analyzing funds, because often they have incubation periods with little money, and they are presented only if successful, a clear selection bias. Then, they muddle along. But if they made 50% their first year with $10MM (their equally prescient yet less fortunate peers lost money and never went live), and then made 5% the following 6 years when they had $250MM under management, allowing them to look like all-stars based on top-line returns.

In contrast, most stock data merely looks at the return of stocks independent of cash flows, the annualized return is around -10%, consistent with a stock going to zero over 10 years. But the Internal Rate of Return was -47%, much worse. This is the average return to your average dollar that came in, weighted by how long it was there. It is a much more accurate picture at what an average investor endured.

Clearly in cases of failed companies, they generally will have much lower IRRs than 'average rates of return'. Ilia Dichev (2005) looked at a variety of indices, and equity net inflows, and found this adjustment reduces returns by 1.3% for NYSE/AMEX for the period 1926-2002, 5.3% for Nasdaq from 1973-2002, and 1.5% for 19 major international stock exchanges from 1973-2004. Applying this correction to the entire market is something that is never done in all those 'stocks for the long run' or Ibbotson analyses.

Why don't these experts recognize this bias? Well, economists hate to do anything that decreases their best example of the risk premium, the elusive thing that explains everything and nothing; it is implied by their basic conception of utility used everywhere, and so they can't simply turn it off. As to index providers, every one of them has a vested interest in indices they catalog, they are not disinterested providers of information. Just as Moody's used to selectively present their default data (note that munis and asset backed securities are separated from corporate defaults?), the major index providers have a strong vested stake in having their product look good.

Socrates's Lament

Nicholas Carr wrote The Shallows: What the Internet Is Doing to Our Brains, which talks about how the internet is turning our minds into mush. It is too easy to seem to understand what you are talking about, to get distracted, etc.


This seems to be a constant refrain, and I'm sure it came up when calculators were invented. But I was surprised that Socrates came up with it first. In Maryann Wolfe's book Proust and the Squid, she notes the Greek alphabet was a real innovation in human culture. An alphabet uses the minimum of notations necessary to express a spoken language unambiguously to its native speakers. It has a mere 26ish letters, as opposed to 900 cuneiform characters in Sumerian, and thousands of hieroplyphs in Egyptian.

Yet Socrates complained that with written texts, it would be less necessary to memorize passages, because you could always just pick up a tablet when needed. When you actually have to memorize the Iliad you understood it better than if you relied on cheat sheets. He thought easy information acquisition lead to an illusion of knowledge, and thus curtailed the more difficult, critical thought processes that lead to knowledge. The semblance of knowledge is easier, and can lead to a lack of knowledge if someone stops as soon as they merely appear knowledgable by using their new fangled written text.

I must say there's some truth here, which is why it keeps recurring. Surely if you memorized the Iliad you understood it moderately well. Similarly, if you have only used statistical software with prepacked routines, rather than coding it up yourself, you probably do not understand it as well as you could. Yet, once you do understand it, having the prepackaged routines makes you a lot more efficient, as anyone who uses Excel can understand. The key is to spend time really learning tools that give you a competitive edge. Ultimately, anything really important you bring to the table takes some barriers to entry, which involves the time needed to develop the technique or wisdom of some special insight you have. You have to simply not fool yourself when you use things you do not really understand; it takes simple discipline.

Plus, there's the matter of memory limits. Our brain is finite. Wolf's a specialist on language processing, and seems especially motivated by the dyslexia in her family. These are people otherwise normal who have difficulty reading. On problem with dyslexics are the visual reversals, as when "b" is confused with "d" or "p" with "q", Dysexic children can write these different letters correctly, but will often say the wrong names. They have trouble associating the letter to a phoneme, and this hinders their understanding of the writing because phonemes are intrinsic to how we actually process language (explains why I often misspell homonyms).

Wolf notes that dyslexics tend to use more of their right hemisphere processing speech than regular readers. However, she argues this frees up the left hemisphere for other tasks, and so many dyslexics are phenomenally productive using this extra space for visual-spacial competence. She argues that these deficits are directly related to excellence in the examples of Leonardo Da Vinci, Einstein, and Thomas Edison.

This is an interesting hypothesis, that 'freed space' makes one better at something. It reminds me of that scene in Sherlock Holmes' Study in Scarlet, where Watson was amazed that Holmes was ignorant of famous writers like Thomas Carlyle, and did not understand Copernican Theory of the solar system. Holmes interjected:

I consider that a man's brain originally is like a little empty attic, and you have to stock it with such furniture as you choose. A fool takes in all the lumber of every sort that he comes across, so that the knowledge which might be useful to him gets crowded out, or at best is jumbled up with a lot of other things, so that he has a difficulty in laying his hands upon it.

The brain can hold a lot of concepts, but not an infinite number. The idea that the size of my brain is a constraint is interesting, and there is rather strong evidence of a positive correlation between brain size and IQ (>0.4). As a materialist, this should come as no surprise.

If you have good discipline, access to easy information is a good thing. It's may be easier to cheat and appear knowledgeable now more than ever, and if that's your goal, you live in good times, but it's as hard as ever to find something new, true and important. You simply can't afford to not use the short-cuts offered by prepackaged software and internet searches that enable one to kludge together data and ideas that make us productive relative to others.

Tiger Mom Priorities

David Brooks wrote an op-ed criticizing Amy Chua's hypercritical, education-status oriented child-rearing strategy (see her WSJ piece here). Chua clearly presents an unsympathetic portrait: no play dates, 3 hour mandated violin practice, etc. I think if kids want to practice for 3 hours, great. If, like me, you hated practicing instruments, it just teaches you to hate it even more. By clearly going too far on an otherwise good idea, it makes fun fodder for commenters.

Brooks' main point is that preventing kids from sleep-overs and other childhood games deprived them of essential, nay, dominant skills in managing status rivalries, negotiating group dynamics, understanding social norms, navigating the distinction between self and group. These are the really important skills learned in childhood.

Clearly, these are essential skills, but the problem with prioritizing this kind of competence is you can't really monitor them, and what you can't monitor, you can't instruct, drill, or correct, other than simple things like teaching your kids to say please and thank you, and to see 'getting mad' as basically an internal failure. So, while letting your kid play with other kids is essential, I don't see it as a more important parenting strategy because kids will naturally work these skills without prompting. Math, reading, and writing, are not natural activities, and there's a brief window in childhood where one can put these concepts into their brain efficiently. As they were discovered a mere 5 millenia ago, our monkey brains need external stimuli from outsiders old enough to know they are actually important to force us to focus on these skills. On the other hand, our brains were made to master speech and social networks, so that doesn't need so much guidance.

As the Serenity prayer notes, the issue of what is important yet unmanageable, and what is important and manageable, is a very important distinction. It is far more important your kids have common sense than know calculus, but while you do you best every day in teaching common sense, it isn't really something to address directly. You can and should, however, make them learn their math tables.

Tolerance can be Insulting

Over at Bloggingheads.tv, psychologist Joshua Knobe brings up the point that if someone is tolerant of your differences on something important, it highlights they basically don't care about you. This highlights that virtues, like vitamins and oxygen, can be too much or too little.

Examples of the Consensus

Robin Hanson reminds us that the scientific consensus is often wrong. Ron Bailey did a Nexis search of the phrase 'scientific consensus over the past 25 years, and found the following:

• saccharin causes cancer in humans
• dietary fiber appeared to reduced the incidence of colon cancer.
• agents found to cause cancer in animals should be considered suspect human carcinogens
• fusion energy reactors would produce more energy than it consumed within five years
• acid rain is destroying lakes and forests

These are no longer consensus findings. He did find the phrase 'scientific consensus' in regards to uncertainty about when life starts, which probably still stands. Yet in all, that's a pretty weak record for the consensus.

A Different 'No Free Lunch' Theorem

Milton Friedman popularized the phrase 'there's no such thing as a free lunch' by using it as the title of a 1975 book, and it often appears in economics textbooks. It is a core idea in economics, but I was surprised to learn it comes up in AI.

My experience with neural nets is decidedly negative. They overfit and provide no intuition for modification. They are consistent estimators, that is, with enough observations they will find patterns, but financial data are notoriously correlated in such a way that we rarely have enough data to make these asymptotic properties relevant. That is, look at cross sectional equity returns. There are officially around 5000 stocks in the USA, so after a decade, you have about 12 million daily observations, which seems like a lot. Yet the really small 2000 stocks are still measured poorly in standard databases, as in the CRSP tapes a strong 'low price' effect still exists even after they supposedly rid the data of the delisting bias discovered by Shumway (delisted firms would often have 'N/A' for their final return, when in fact they would lose 40%). But then even with these 3000 stocks left, if you simply find something correlated with the internet bubble or the mortgage crisis, you will discover a high t-stat in explaining cross-sectional returns, and these patterns are sui generis, not fundamental characteristics of the data.

Jeff Hawkins is a successful computer programmer who is really interested in aping the brain to come up with more efficient algorithms. In his book On Intelligence, he argues our neocortext is the essence of our intelligence. It is the wrinkly surface of mammalian brains that is about the size of a napkin folded up in your head. 1k square centimeters, 2-3mm thick, about 30 billion neurons. He calls this Heirarchical Temporal Memory because data is stored in heirarchies: we understand phonemes, words, phrases, ideas, each one a higher level understanding of the concepts below it.

We experience the world through a sequence of patterns, and when we recall them we do so to predict. It's an internal metric, not one based on behaviorism. Most patterns are spatial or temporal, in that a melody has has a a chord (set of frequencies), and the temporal relation of of those chords. When we comprehend a piece of music we are making predictions about what comes next, anticipating the next note. We recognize the song when our predictions are correct, and this generates an 'aha!' moment, as when you see the dalmation in the picture of black and white spots. Hawkins argues that intelligence is essentially this, the prediction of these patterns: when we predict something well our neocortex strengthens these neural connections, when it doesn't predict well the unsuccessful pattern disappears. The brain makes predictions down to what we perceive, and up to what we interpret. Intelligence is prediction.

It's an intriguing concept, and a rather stark contrast to two ideas from Alan Turing. The first, that the brain is just a big computer, with a bunch of AND-gates and OR-gates, just like logical circuits. Turing proved mathematically that with an infinitely long program, it can perform any definable set of operations in the universe. His second idea was that you could judge whether a computer is intelligence if you can't tell it's not human by asking it questions, and see if you could tell it was a computer.

As mentioned, Hawkins thinks intelligence is not emulating the behavior of humans, but rather, predicting well at every level, mainly for the many little consistencies we see in how the things in the world relate spatially and temporally. Secondly, there is the problem that Turing's program is very inefficient. There is a 'no free lunch' theorem in optimization programs: no learning algorithm can be better than all other learning algorithms for all problems. To be best for one problem you have to tune it to the problem at hand. A Turing machine makes no assumptions, but you pay a cost because it is not necessarily organized hierarchically or temporally, which turns out to be a good way of describing our world.

A general algorithm must be tweaked to incorporate the specific problem at hand. This mainly concerns the selection of inputs (not too many), and their transformation (sigmoidal transformation, putting raw data into stationary time series, etc.), but also the functional form. This 'outside the box' framing problem uses up a lot of degrees of freedom, but doesn't really show up anywhere because once done, any algorithm ignores them in their standard errors, as if this was the first and only way of apprehending the data. Specific algorithms contain a lot of specialized intuition or common sense, which is why really smart people are often clueless at modeling the real world.

Hawkins' book is a good read, and he has several interesting videos on YouTube. I like the idea of the no free lunch, and adding structure in the way he mentions. Yet, I still think his ideas are too unstructured. He has had access to a lot of funding and has been working in this dimension for almost 10 years. He seemed to think his track was on the cusp (couple years) of finding some really better artificial intelligence algorithm back in 2005 or so. I haven't see those. His signature achievement, the creation of the Graffiti software to recognize letters in things like the Palm Pilot, seems like your typical useful, highly specialized algorithm. I suppose some neural net modified in this 'temporal hierarchical' structure is better than one with merely layers of hidden nodes, but that still leave a lot undefined in the step where you choose what to look and what functional forms it should constrain itself to. Robot maids, machines with general intelligence, are still science fiction. Instead, we have great machines for applying specific logic, which invariably is specialized in such a way to make any big optimization idea rather uninteresting. The 'No Free Lunch' theorem Hawkins mentions seems to counter his big idea.

Asset Pricing Theory Video

I did this using Xtranormal avatars, which perhaps will be the new PowerPoint. It's a substitute for Cochrane's Asset Pricing text.

Why Do People Gamble?

I was in Vegas last weekend, doing field research ;) on utility functions. People clearly like to gamble, and are willing to pay a premium for the exposure to random payoffs with large skew. This is contrary to the usual assumption that people are risk averse, and necessitate a positive expected return to take a gamble. Early on in utility theory, Friedman and Savage (1948) siezed on this anomaly to argue that the curvature of an individual's utility function differs based upon the amount of wealth the individual has. This curving utility function would thereby explain why an individual is risk-loving when he has less wealth (e.g., by playing the lottery) and risk-averse when he is wealthier (e.g., by buying insurance). Harry Markowitz, a former student of Friedman's, argued the implications of the Friedman–Savage utility function were paradoxical. Specifically, its implication that those at the highest level of income would never take risks. His solution was to relate the curvature of an individual's utility function to increases in wealth.

Thus, from the very begining, utility theory was problematic. Solutions had a whack-a-mole tendency to rid one anomaly and create another. The basic problem is that

1) People pay to buy insurance.
2) People pay to gamble.

It's still a puzzle, though every so often a new paper shows up to solve it (eg, Kahneman and Tversky's prospect theory). I think people's paying for risky investments that are a function of skill makes total sense in a search for alpha, discovering one's comparative advantage. Yet, the pure randomness of much gambling--slots, roulette, craps--makes no sense unless you assume people have some belief in luck.

Kanye Likes Himself

It's actually an ugly watch:

The rap giant's latest self-appreciative purchase is an $180,000 watch emblazoned with his own portrait. The Tiret watch took more than five months to make, and features yellow, black, brown and white diamonds placed to resemble His Kanyeness.

Plus, one nice thing about a really nice watch, is the thought of giving it to your grandkids, and having your picture in it makes that more difficult.

Beta1.0 2010 Results

I have a website betaarbitrage.com, which I update monthly, and has the daily and monthly historical returns for minimum variance portfolios (MVPs), and targeted beta portfolios. MVPs are described here. Targeted beta portfolios are my little project, and I describe them here. The basic idea for these is, instead of investing in 'high' and 'low' beta stocks, target a direct beta of say 1.0, or 1.5. This avoids the variability in the beta deciles, which as any good investor knows, goes up during expansions, and compresses as correlations rise during bear markets.

The basic idea for the Beta1.0 portfolio, is to outperform the S&P500 while minimizing benchmark risk. By using only stocks with market caps above $500MM, and then choosing 100 stocks with historical betas closest to 1.0, the prospective beta is near one. That is, you grap this middle section. This gives you the benefit sf trivial benchmark risk, but by avoiding the super high beta stocks that are part sf the index, you add a couple percent to you return:



Now, I have been doing this historically for a while, choosing portfolio constituents once every 6 months, and applying them forward. Data at betaarbitrage.com are free and available for download without any pesky login stuff. If you look at the actual beta of this in 2010 using daily data, you see a near 1.0 beta, pretty much as expected:



How did it do last year? Fabulously! Check it out:


The beta1.0 portfolio was up 30%, vs 14% for the S&P500. How did this happen? Mainly because the beta1.0 portfolio drew from all stocks over $500MM, and so reaped the benefit of the small cap boom in 2010.



Here the RTY is theh Russell 2000, or the largest stocks outside the top 1000 stocks. It trounced the S&P500 and Russell 1000 last year (25% for RTY vs. 13ish% for SPY and RIY), primarily because of the rise in the second half of the year. Using data since 1962, the beta1.0 portfolio outperforms about about 2% annually, slightly higher volatility (~3% annualized), and a beta around 1.0 (hence the name).

I actually have a patent pending on the Beta1.0 idea ... being on the wrong end of the law, I thought I might try using it to my advantage.

Forecasting Charlatans

An article in the Boston Globe by Joe Keohane discussing economic forecasting:

Economists who had a better record at calling extreme events had a worse record in general. “The analyst with the largest number as well as the highest proportion of accurate and extreme forecasts,” they wrote, “had, by far, the worst forecasting record."
...
And success, as Denrell revealed in an earlier study, is an especially bad teacher. In 2003 he published a paper arguing that when people study success stories exclusively — as many avid devourers of business self-help books do — they come away with a vastly oversimplified idea of what it takes to succeed.

When I worked for an economics department, I quickly learned what a lame business we were in. Our stated purpose--to forecast the economy to allow people to make better decisions--was different than our actual purpose--to provide rationales for decisions already made, to serve as an excuse to have a get together. The sad thing is that a Big Lie needs many little lies, as the stated goal of forecasting accuracy could not be discussed openly and honestly, because if one did the stated purpose becomes untenable, and then the unstated purpose becomes unworkable. It's one of those phony little kabuki dances that seems so quaint in primitive cultures, but just as common in our own.

A problem in this field is that accuracy spells extinction because no one wants to listen to an honest forecaster, they don't purport to know enough. Rather, listen to someone who can make you rich! In selling forecasts to the masses, honesty is a strictly dominated strategy.

When I first got to KeyCorp after grad school I was most excited by an opportunity to work with their asset management group, which was the main reason I went there--the chance of getting into being a pm for equities. I remember pitching the idea of beating the S&P by a couple percent and having lower volatility by investing in low volatility stocks. The guy in charge said, 'I have people here who can outperform the S&P by 10%!' Now, clearly he didn't, and to this day I don't know if he was so stupid to think he really believed it, but it sure sold better than the truth, and selling funds is more important than returns because due to the magic of survivorship bias the current set of funds will always be better than average. Economic forecasts are even less constrained, with extreme statements as out-of-the-money options either paying off big or disappearing down the memory hole.

The article quotes Philip Tetlock, an expert in experts and their predictions. In a lecture he outlines his extensive survey of experts and their predictions, and categorizes them as either foxes (detail experts) or hedgehogs (big picture experts). He finds that foxes are usually right, but hedgehogs are "occasionally" right on the farthest out predictions. It's a ambiguous finding, especially considering he criticizes experts for not being honest and diligent in their self-evaluation. When asked if he is a fox or a hedgehog, Tetlock said he didn't know, but perhaps in general a fox with a little hedgehog in him--isn't everyone? He notes this is the general pattern: foxes qualify their statements a lot, but are more correct--when you ignore the qualifications. Notice the recursive property here?

Gay homophobes and racist anti-racists are as common as dogmatic atheists and heuristic-bound behavioral economists. I think those who study biases are disproportionately guilty of what they study. This isn't a necessary paradox, just a tendency.

Grameen Bank Hits Wall

For decades The Grameen Bank in Bangledesh was held out as sort of an anarcho-capitalist dream: small groups lending and monitoring borrowers, and making money too boot. Banks are usually big, controlled by wealthy parties, but here was a case where the decision makers were little people, and historically loss rates were pretty low. Most loans were for little amounts like $100, and made to women involved in little home businesses.

Although each borrower must belong to a five-member group, there is no form of joint liability. However, in practice the group members would contribute the defaulted amount with an intention of collecting the money from the defaulted member at a later time. There was a great deal of invasive monitoring, cajoling, and often physical intimidation, but when done informally by friends, was not considered oppressive. This clearly would not scale.

The bank has grown a lot in the past 10 years, and the founder, Muhammad Yunus, won a Nobel Peace Prize in 2006 for this idea. The problem with banking is that when borrowers can't pay, there's a political incentive to paint the lender as the bad guy and repudiate the debt. The bigger Grameen got, the greater the political incentive treat them like a regular bank, as the Bangledeshi Prime Minister recently denounced the bank as “sucking blood from the poor in the name of poverty alleviation.”

Elsewhere, imitators have seem similar ends. Nicaragua’s president, Daniel Ortega, for example, supported the no-pay movement, which was started in 2008 by farmers after some borrowers could not pay their debts. Partly as a result of that campaign, a judge recently ordered the liquidation of one of the country’s leading microlenders, Banco del Exito.

Banks should not be thought of as cookie jars full of money, but rather things that manage the savings-investment nexus. They evaluate borrower's creditworthiness, and repossess pledged collateral if payments are not made. If banks simply wrote-off every bad loan as opposed to seizing collateral per the original contract, recovery rates would be so low the whole industry would have to change, and interest rates would rise several fold in anticipation of the higher expected loss rates. Every so often the great idea comes up: why not just stop paying the banks? Borrowers are invariably more numerous and have less net wealth than the savers. Thus, we had the expropriation of the Knights Templar in 1307, several Jewish expulsions in the middle ages (Jews specialized as money-lenders), and various communist takeovers such when the Khmer Rouge took over Cambodia. In every case, this had disastrous effects on the economy.

Surely, if you can't pay back a debt there should be a way to avoid slavery, and we have that now, it is called bankruptcy. To simply repudiate loans, as with the whole 'robo-signer' pretext, is a naive, populist remedy that has always been regretted. The basic problem is, by getting rid of debt you get rid of savings, and without savings, you have no investment.

Penny Stock Risk Premium Has Wrong Sign

In my book Finding Alpha I document 20 different asset classes where the risk premium is either missing or goes the wrong way. But I did not include penny stocks, so it's a good 'out of sample' test of my hypothesis, that he risk premium rises as one goes from super safe to safe, stays flat, then goes the wrong way for lottery ticket investments like sports betting, distressed debt, high volatility stocks, and of course lottery tickets.

Bjorn Eraker and Mark Ready just put out a paper entitled, Do Investors Overpay for Stocks with Lottery-Like Payoffs? An Examination of the Returns on OTC Stocks. They find:

In our nine year sample, approximately $820 Billion traded in the OTC markets. This represents less than a month of typical NYSE volume. Still, the OTC markets are not insignificant: the average stock in our sample lose about 15M dollars from the first observation to the last. In total, stocks in the OTC market lost approximately $180 Billion of market value over our sample period. These numbers themselves are conservative because we did not make any assumption about the value depreciation taking place for stocks that no longer trade.

In case you didn't know, penny stocks are lousy investments; high risk, negative return. They suggest either the Miller explanation, that differential expectations imply a winner's curse, or the Barberis and Huang explanation, that people really love positive skew. They leave with the note that they are looking at a model in which investors display risk aversion over losses and risk-loving behavior over large gains would conceivably explain the results.

That's the great thing about the 'risk premium'. It's not a theory, but rather a framework, so that any anomaly merely rejects an incorrect theory within this framework, and not the framework itself. Nonetheless, it is supposedly the foundation of modern finance, based on Markowitz's brilliant, and rigorous, insights. It explains everything, because when it doesn't, you just say people are not risk averse, but rather risk-loving over various parts of the return distribution for specific asset classes. Kahneman and Tversky's famous prospect theory argued that things like lotteries could be explained by 'risk-seeking behavior in losses involving moderate probabilities and of small probability gains', which of course is different. The Black Swan idea is that people prefer high probability bets with small gains but occasion high losses. The point is clear: the utility function explains residual returns, and places no restrictions on them.

My argument is much less plastic. First, people are benchmarking against their peers and colleagues, and this leads to a zero risk premium as a first order approximation. There is no 'risk premium' in financial markets, just as there is no courage premium to the aristocracy. But people also have a strong preference towards lottery tickets, as explained in my post 'Why are High Risk Stocks so Crappy.'

Benchmarking, or a 'relative status' or 'envy' utility function, are all the same thing. It explains why risk premiums don't exist in general. It also explains the Easterlin Paradox, where greater wealth over time has neither reduced the risk free rate, nor increased general happiness. It explains too why people's asset allocations generally ignore strict mean-variance optimization, because they consider what the 'standard allocation' to be risk neutral, because keeping-up-with-the-Joneses is risk neutral. It's rather straightforward to increase one's Sharpe ratio (buy low volatility stocks), or one's covariance with their income (invest in other developed country indices), but this is considered risky, because then you have benchmark risk, which is what investors really hate more than volatility itself.

A Theory of Growth

Business cycles and the essence of long-run economic growth are distinct issues. Preventing recessions is not the key to growth, as these are regrettable but unavoidable companions to an economy directed by a capital allocation process that is susceptible to systematic failure. Preventing the last failure is pretty irrelevant, because the next systematic failure will be different. Last I checked, only the US government is offering low-down payment loans, and no one offers no-documentation loans, so our government is not really helping here. As for creating growth via something new, if centralized governments could do that, the Soviet Union would still be around.

That decentralized, self-interested, people can collectively make such large errors seems irrational or corrupt to many, but they should remember that growing economies require people to be making things better, which means, new ways of doing things. New ideas are often wrong. Economics has gone onto intellectual cul-de-sacs many times (socialism, Keynesian macro models, input-output models, Hilbert spaces in finance, Arbitrage Pricing Theory, Kalman-filter macroeconomic models, etc.). Other scientific disciplines have their own mistakes, and political mistakes--stupid wars--are also common. These are rarely conspiracies, but rather, smart people making mistakes because the ideas that are true, important, and new, are really hard to discern, and tempting ones are alluring when lots of other seemingly successful people are doing them.

My Batesian Mimicry Theory posits that recessions happen because certain activities become full of mimics, entrepreneurs without any real alpha who got money from investors looking in their rear-view window of what worked and focusing on correlated but insufficient statistics. For example, people assumed a nationally diversified housing prices would not fall significantly in nominal terms, because they had not for generations; people assumed anything related to the internet would make them rich in the internet bubble, conglomerates would be robust to recession in 1970, that the 'nifty fifty' top US companies had Galbraithian power to withstand recessions in 1973, that cotton prices would not fall in 1837, etc.

As in ecological niches, there is no stable equilibrium when mimics arise to gain the advantages of those with a real, unique and costly, comparative advantage. Every so often there are too many mimic Viceroy butterflies, not enough real poisonous Monarch ones, and a massive cataclysm occurs as predators ignore the unpleasant after-effects and start chomping on all of them. The Viceroy population grows until this devastating event occurs, a species recession. Next time, it won't happen in butterflies, but rather, among frogs or snakes. They key is, some ecological niche is always heading towards its own Mayan collapse (distinct from the 2012 Mayan apocolypse).

The key to wealth creation is doing more with less--destroying jobs at the micro level and creating jobs at the macro level by reallocating capital and labor to more valuable pursuits. The computer got rid of things from typesetters, secretaries, to engineers working with slide-rules, but these people didn't stay unemployed, they did something else, making the economic pie bigger. This is antithetical to government and unions who think creating a permanent 'job' creates productivity--stability at the micro level and stagnation at the macro level. Wealth is created by having decentralized decision-makers focused on simple goal of making money, which means, they oversee transactions where revenues collected are greater than expenses paid. If externalities are properly priced (I know, most liberal think this never happens), this implies value is created. The continual improvements in method (ie, productivity, wealth creation) merely maintain profits in a competitive environment; to do nothing would see their profits eaten away by competitors would could easily copy what they did and just undercut their prices.

The key to this is having managers who keep their workers focused. A good example is a story I heard second-hand about a football player for Minnesota Vikings in the 1970s. Coach Bud Grant called this marginal player into a meeting, and said, 'Here's what I need you to do...'. The player, an articulate fellow quite confident in himself, interrupted with an explanation of why he wasn't doing better and suggestions about how to correct it, mainly focused what others were doing wrong. Grant cut him off: 'You don't understand. This isn't a negotiation. Do what I'm telling you, and you have a role here. Otherwise, you don't.' Hierarchies only work well when people have clearly defined goals, and managers who manage their direct reports singlemindedly.

Private firms can do this much more quickly and often than government, and are rewarded with investment and retained earnings to the degree they do it well. When the government wants to do something, like build a light-rail system, it instead satisfies all its stakeholders who have no financial downside, only veto power, and so the cost/benefit calculus is almost irrelevant. The probability that benefits will outweigh costs when not prioritized is negligible, as highlighted by the fact that companies have to work very hard to make this positive when all those other considerations are ignored.

Thus, Minneapolis's light rail, at the cost of $1.1B for 12 miles of track, takes me longer to go downtown than a car because it stops 19 times at places no one wants to go because these 'hubs' were then sold as development opportunities, and an unusual number of ex-city councilmen are part owners of coffee shops and stores near these stops. Ridership does not even cover their marginal costs. It could have worked if they had an express train that went non-stop from end to end, but doesn't because it was not designed with the goal of making money, only the hope.

Good companies like Facebook, Apple and Google, have this sense of really understanding their users. Lots of simple things that making going to their sites and getting what you want. Their inferior competitors are relatively ugly, cluttered, and clunky. These generally weren't genius ideas like the ideas needed to create the first transistor, or Cantor's diagonal argument, in that their competitors had similar raw competence in these field, but it did take people looking to do things better than others, and decisive people who could empathize with their customers created really great things.

Robin Hanson had a neat article about the Myth of Creativity, where he criticizes Richard Florida's vision of bohemian lead productivity:

This is a Star Wars vision of innovation: "Feel the force, Luke; let go of your conscious self and act on instinct." And it is just as much a fantasy as that celluloid serial. Innovation is no more about releasing your inner bohemian than it is about holding hands, singing Kumbaya, and believing in innovation.

In truth, we don't need more suggestion boxes or more street mimes to fill people with a spirit of creativity. We instead need to better manage the flood of ideas we already have and to reward managers for actually executing them.

Sure, it's good to punish fraudsters, and be wary of the stupid ideas that were passed off as brilliant in the prior cycle (eg, Angelo Mozilo winning the American Banker's Lifetime Achievement Award in 2006, celebrated by politicians on the right and left, prized by Fannie Mae, and Harvard, is now an example of the 'unregulated predatory private sector'). But this is like learning not to put one's hand on a hot stove--good to know, but old news to most. Our priority at the top level should be to get out of the way, and so government should focus on its essential but limited perennial tasks as opposed to creating some new engine of growth. Leave that for the millions of people making sure millions of small changes are constantly made to daily procedures. Such changes do not require vision from politicians, subsidies, or tax breaks, but are rather the natural by product of people trying to make a buck. It's the standard Hayek/Friedman view of macroeconomics, and it's still the best description of how the complex adaptive system of our economy works.

The Next Train Wreck

From Joshua Rua's Are State Public Pensions Sustainable?

Based on September 2009 asset values, if state pension fund asset returns have an average return of 8% going forward (the states’ typical assumption), states in aggregate will run out of funds in 2028. If average returns are 10% through 2045, the funds in aggregate will be roughly sufficient to cover liabilities to existing workers under the states’ actuarial assumptions. If average returns are only 6%, state funds in aggr gate will run out in 2024. This analysis assumes that state inflation forecasts, which average 3%, are met. If inflation is greater holding the investment outcomes fixed, then even under the higher asset returns the funds will run out sooner, as many state systems provide inflation-linked cost of living adjustments (COLAs) to beneficiaries.

Illinois, Connecticut, New Jersey, and Indiana are worst, whereas New York(!), Florida, North Carolina, Nevada and Alaska are best. He estimates the present value of the shortfall at $3T.

This is merely state pensions, mind you. The US social security system has about $13-17T in unfunded obligations, and $30T or so for Medicare (old folk free medical). The Federal deficit is about $1-2T, on GDP of around $14T.

I think our only hope is the Mayan apocalypse. Or inflation.

The Education 'Risk' Premium

A big puzzle in education is that there appears to be a huge wage premium to college--about 80%--and yet only 40% of high school graduates go to college. Above is a graph from a paper by Athreya and Eberly to be presented at this weekend's AEA meetings on the subject 'Risk and Expectations in Higher Education' [click to enlarge].

Looking at various types of post-secondary education and completion rates, there is a perverse lack of response to this carrot, typically described as strangely “anemic” (see Altonji, Bharadwaj, and Lange (2008)).

There are several ways to interpret this, such as constraints to kids going to college who can and want to, constraints facing households considering investments in human capital, or that this reflects compensation for, and responses to, an investment opportunity that is lumpy, irreversible, and most crucially, risky. That is, it's a risk premium. [The simple idea that this is from increased globalization and migration making unskilled labor worth less is not a prominent explanation, probably because academics like to assume that in aggregate skill is a function of education, a simple choice variable].

The risk premium story works like this: the average college-drop-out rate is 50%. Taking into account dropout risk, a simple calculation of risk premium accounts for about half of the excess return to college education. Thus, Gonzalo Castex proposes that the risk-premium of college participation accounts for about 29% of the excess returns to college education. Risk averse agents are willing to forgo the higher return to college in order to avoid the dropout risk.

Ah, the omnipresent risk premium. In theory, it explains so much of the variation in wealth and income around us. But when you examine it further, it simply acts like another free parameter that 'explains' the data via the magic of overfitting. Note that schools with the lowest completion rates are generally lousy schools like Southern University at New Orleans (5% graduation rate!). These schools don't generate a big wage premium, though their risk is huge. Or, if you look across majors, and account for the fact that more people 'switch' from hard majors like engineering to easy majors like sociology rather than vice versa, this drop-out risk also fails to explain the differential major wage premia. Lastly, there's PhD programs, where people seem to spend a lot of time for virtually no measurable increase in earnings compared to getting a Master's degree (see Economist article here).

A good theory doesn't have to be be consistent with all the data, but it does have to explain at least 'most' of the conspicous data points it was not designed to explain. In this case, adding a risk premium can explain the 'puzzling' college wage premium, but does not generalize across colleges, within colleges, or across degrees (BA, MA, PhD). The 'risk premium' explanation is always a red herring.

Federal Regulators At Work

From Slashdot:

As a result of the US Government's complete failure to investigate credible warnings about 'Underwear Bomber' Umar Farouk Abdulmutallab from none other than Abdulmutallab's father, senior American counterterrorism officials say they have altered their criteria so that a single-source tip can lead to a name being placed on the watch list.

The government's ability to discriminate is basically zero, so you have to choose to allow everything or nothing, because that's all it can decide.

NYT Pixel Chart

The Sunday NYT has a neat pixel chart, where the blocks are years from 1920 to 2010. Each cross-tab represents the return from the vertical axis to the horizontal, so dots near the diagonal are short-term investments, those towards the right, buy-and-hold.

What it says to me is that the long run is easier to predict than the short run in this sample. Jeremy Siegel suggests these are invariant parameters, yet as Lubos Pastor and Robert Stambaugh point out in their Are Stocks Really Less Volatile in the Long Run?, we have to remember there are only a handful of '20-year' independent samples here, hardly enough to say these are population means. Further, data from the 19th century was often filled with selection biases (see here ), which makes sense when you consider who has the incentive to derive these indices (those selling equities).

Ezra Klein's Rise, Kaus's Stagnation

Ezra Klein recently create a lot of snickering when he stated that no one understands the US Constitution because it is over 100 years old. As many of the US Constitution's Amendments are a couple sentences long, I think is is mainly because many don't like what it says. It is important to remember, as Fareed Zakaria noted in Illiberal Democracy, the Western model of government is best symbolized not by the mass plebiscite but the impartial judge reading from a rule-book (eg, the US Constitution). "Congress shall pass no law" regarding X is only ambiguous wording to do-gooders who find this law inconvenient.

In general, Klein's a typical wonk: a smart person who can pepper his nostrums with academic studies. That selectively pulling research, even disinterested research, is biased is one of those meta-effects most people don't notice, and in general is unprovable. The truth is complex enough that you can cherry pick data, snippets from legislation, quotes from academics, that supports/critiques any big issue of the day, and this is all the easier if it doesn't bother you to be a partisan shill.

Mickey Kaus, a 50-something who has studied these issues for decades, criticized Klein's dismissal of union problems on Slate. A couple years ago as Ezra Klein's star was rising, Klein wanted to debate Kaus about health care on Bloggingheads.tv, but Kaus would have none of it because he thought Klein was an unworthy neophyte (5 years out of college with a poli-sci degree). Now, Klein's one of the top lefty pundits, a columnist for The Washington Post, Newsweek, a contributor to MSNBC, and often on the Sunday talk shows. Klein is now too popular for Kaus and the tables are turned: Kaus wants to debate Klein but Klein won't stoop to Kaus's lower level of popularity.

It's kind of a sad arc, the elevation of Klein, the stagnation of Kaus. The problem with Kaus is that he's accumulated principles: he's generally a Liberal for greater equality, but thinks many standard solutions--unions, welfare, untrammeled immigration--are counterproductive. This leaves Kaus without a base because conservatives don't like his Liberal likes, whereas good Democrats don't criticize their base on things like unions and immigration.

Klein's ascent highlights that such accumulated wisdom isn't helpful to a career as a pundit. Paul Krugman's perpetually peevish posts highlight that the dominant strategy to be popular is to write-off your opposition as either stupid or evil and just document the latest data and theory that supports your Weltanschauung that brings back like-minded readers for bi-weekly confirmation of their biases. In contrast, someone like Kaus who actually thinks for himself will invariably present idiosyncratic ideas which by definition are seen by most people as unconventional and wrong. This rubs the rubes the wrong way, they see you as an undependable intellectual ally, not part of the team. In a sense, they are right.

To see how lame tendentious ratiocination is, listen to this 20 minute discussion of the economy by two macroeconomists who obviously are trying to support their quaint macro forecasts. All the experts, even these guys, know their forecasts are inferior to inarticulate Vector-Autoregressions or consensus forecasts, so it's rather pointless to hear the 'why' behind their inefficient forecasts. In any case, it's what pundits do, but without the obvious partisanship such blather is exposed for what it is: boring and pointless.