Ben Bernanke Must Be Hoping Rational Expectations Doesn’t Hold…

23be99cb-434e-412f-a60e-a0b49bbe470d-460

In the theory of rational expectations, human predictions are not systematically wrong. This means that in a rational expectations model, people’s subjective beliefs about the probability of future events are equal to the actual probabilities of those future events.

Now, I think that rational expectations is one of the worst ideas in economic theory. It’s based on a germ of a good idea — that self-fulfilling prophesies are possible. Almost certainly, they are. But expressed probabilities are really just guesses, just expressions of a perception. Or, as it is put in Bayesian probability theory: “probability is an abstract concept, a quantity that we assign theoretically, for the purpose of representing a state of knowledge, or that we calculate from previously assigned probabilities.”

Sometimes widely-held or universally-held beliefs turn out to be entirely irrational and at-odds with reality (this is especially true in the investment industry, and particularly the stock market where going against the prevailing trend is very often the best strategy). Whether a belief will lead to a reality is something that can only be analysed on a case-by-case basis. Humans are at best semi-rational creatures, and expectations effects are nonlinear, and poorly understood from an empirical standpoint.

Mainstream economic models often assume rational expectations, however. And if rational expectations holds, we could be in for a rough ride in the near future. Because an awful lot of Americans believe that a new financial crisis is coming soon.

According to a recent YouGov/Huffington Post survey:

75 percent of respondents said that it’s either very or somewhat likely that the country could have another financial crisis in the near future. Only 12 percent said it was not very likely, and only 2 percent said it was not at all likely.

From a rational expectations perspective, that’s a pretty ugly number. From a general economic perspective it’s a pretty ugly number too — not because it is expressing a truth  (it might be — although I’d personally say a 75% estimate is rather on the low side), but because it reflects that society doesn’t have much confidence in the recovery, in the markets, or in the banks.

Why? My guess is that the still-high unemployment and underemployment numbers are a key factor here, reinforcing the idea that the economy is still very much in the doldrums. The stock market is soaring, but only a minority of people own stocks directly and unemployed and underemployed people generally can’t afford to invest in the stock market or financial markets. So a recovery based around reinflating the S&P500, Russell 3000 and DJIA indices doesn’t cut it when it comes to instilling confidence in the wider population.

Another factor is the continued and ongoing stories of scandal in the financial world — whether it’s LIBOR rigging, the London Whale, or the raiding of segregated accounts at MF Global. A corrupt and rapacious financial system run by the same people who screwed up in 2008 probably isn’t going to instill much confidence in the wider population, either.

So in the context of high unemployment, and rampant financial corruption, the possibility of a future financial crisis seems like a pretty rational expectation to me.

Advertisements

Are Markets Informationally Efficient?

A key assumption in many mainstream macroeconomic models (both formal and informal) is the Efficient Market Hypothesis. Very simply, this is the belief that markets are informationally efficient — that they reflect information with little (or no) delay, leaving few (or no) arbitrage opportunities.

So the real question here is what information do markets (and by markets, I mean free markets where market participants are free to pay and receive any negotiated value for an asset) really reflect? When we see the price of an asset or asset class fluctuating, what does this movement signify? Is it fluctuation in the fundamental value of the asset? Is it just a fluctuation in the market’s perception of an asset? Is it some combination of these two factors? Or is it just random noise? Fundamentally, markets are composed of a series of transactions, each between a bidder and a seller. Each transaction in itself reflects a discrete set of information — specifically, what the bidder and the seller are willing to pay for, and take for that specific asset. This in turn is typically (although not always!) influenced by a some or all of the following: what others are willing to pay and accept for an asset presently, the use-value of an asset, notions of fundamental value (price-over-earnings, stock-to-flow, EBITDA, cashflow, etc), notions of momentum and what others may be willing to pay and accept for an asset in the future (trendlines, gut feelings, “hot stock tips”, etc).

An intriguing addendum to this is that the automation of trading (high-frequency trading) has created bidders and sellers who are acting on the instructions of algorithms. As these instructions are programmed by humans — usually automating some form of technical analysis — the only real difference is that of (extreme) speed. The beliefs reflected in high-frequency trading reflect the underlying algorithmic instructions programmed by the humans who created the algorithm.

Ultimately, whenever we purchase an asset for the purpose of speculation or investment (and even use-value — prices can change, and the price we paid last week or last year could end up looking very expensive, or very cheap) we are taking a guess as to whether the current bid or sell value is worth it. Each agent makes their guess based on a different set of data and expectations. What the prices in markets signify is the operation of this mechanism — different agents evaluating information and making guesses about the future.

Let’s consider the example of Bitcoin, the price of which is currently soaring. Some choose to buy Bitcoins based on momentum, or their liking of the cryptography, or Bitcoin’s inherent deflationary bias or some other positive belief. This is a speculation that the price may continue to climb. Some may choose to buy bitcoins based on their use-value, as an anonymous, decentralised currency that can be used to buy a wide array of things. Holders of Bitcoins may be motivated to sell by the fact that the price has risen since they bought or mined their coins, or by the belief that bitcoin is “in a bubble”, or some other negative belief.

What the market reflects is the net weight of different opinions and resultant human actions. If those who are motivated to buy outweigh those who are motivated to sell, the price  rises and vice versa. This means that the beliefs of big players in a particular market can have strange and disproportionate effects. Consider the effect of the Hunt Brothers’ attempts to coin the silver market in 1980. The price of silver rose from $11 an ounce in September 1979 to almost $50 an ounce in January 1980, as the Hunt Brothers bought more and more. The market was very efficient at reflecting the fact that the Hunt brothers were willing to buy more than the market could supply at lower prices. And once the Hunt Brothers faced margin calls, the market quickly adjusted to reflect the fact that they were now selling instead of buying, and prices fell.

That’s what (transparent) markets are guaranteed to reflect — bidders and sellers, supply and demand. This information is still useful to firms trying to gauge what, and how much to produce.  Everything else — the information that bidders and sellers are acting upon — is not necessarily reflected in market activity. Very often, bidders and sellers are brought to the market by new information regarding a large number of things — price changes, earnings, business decisions, technologies and inventions, macroeconomic data, etc — but there is no systematic or reliable way to predict what humans will respond to, or how they will respond. Human psychology and human action in this sense is totally unstable and nonlinear — consider the recent contrast in market reaction to earnings data from Apple and Google. This instability is an alternative explanation for why consistently beating the market is indeed very difficult, as the Efficient Market Hypothesis implies.

And prices do not even reflect an aggregation of sentiment toward an asset or asset class — they only reflect the sentiment of those who are involved in the market, in proportion to their level of buying and selling activity. This means that the opinions of big players who buy or sell a lot, are reflected many times more than those of small players who buy or sell a little. And irrationality can create a feedback loop — if stock prices are rising, and macroeconomic fundamentals are weak, many market participants may initially be sceptical. Yet as more participants pile into the stock market purely for reasons of sustained upward momentum, more and more participants may begin to suspend their disbelief, if only to not miss out on a profit opportunity. This is one mechanism (of infinitely many) through which price bubbles can form.

Yet accurately reflecting supply and demand is not the same thing as informational efficiency. Empirical data show that arbitrage opportunities are widely exploitable and exploited even in modern marketsOne of the largest forms of high frequency trading is of course statistical arbitrage. This reality should probably be a final nail in the coffin of the idea that markets reflect anything more than the actions of bidders and sellers. Unfortunately, very many models rest on the assumption of informational efficiency in markets, meaning that this approach is very unlikely to die out any time soon.