Reviewing Gregory Zuckerman’s book on the trading firm Renaissance Technologies, I write
Much of the book consists of tales of the gifted mathematicians who ran the firm, along with their foibles and conflicts. These are entertaining enough, but I want to focus instead on two deeper economic issues that arise from the story.
1) Does the success of Renaissance show that financial markets are inefficient?
2) Are the social benefits of the trading conducted by firms like Renaissance commensurate with the profits that they earned?
Pricing is a proxy.
They really bet congestion in the market when considered as a flow. Chart analysis does this, it looks for a congestion buildup in a time series, as if traders need to wait in line to get their trade.
Consider, if instead of betting security prices, we had a currency banker that just optimally allocated the liquidity among the risk equalized deposits and loans. In hat case the numerator in price would be arbitrage free, the only bet was the denominator In this model, for theoretical analysis, there is no stock market. The arbitrage is gone in money so investors can effectively loan money to corporations with a additional debt insurance fee. In that case, all the arbitrage moments are stuffed into the debt risk category. One step in the funds flow is removed, half the short term arbitrage is gone.
This theoretical model tells me tat have the short term bets are about central banking which deliberately adds arbitrage moments to the flow of fund. In fact, a purely competitive monetary system would take out most of the hedges normally introduced by central bankers.
Kling’s asks:
Yes, at least for any given timeframe which is why low latency trading exists. From Kling’s description, it seems to me that Renaissance is exploiting inefficiencies in the minds of competing analysts. Once the knowledge of what patterns signal mental/process/algorithm inefficiencies, the opportunity will disappear just like the Moneyball signals/patterns did in baseball.
Speaking of algorithms, the autocorrect algorithm in Fire OS is brain dead. I make enough writing errors in my own. I don’t need software adding errors to my correctly written words.
“in” my own… Sheesh. I have to rewrite my sentences to see if it’s my fingers or Google/Amazon (not me again).
Arnold, your “we just don’t know” conclusion about the EMH is too cautious. I would not say you are notable for an excess of modesty, so I encourage you not to discover it here. There are statistical ways to say something about whether luck could explain the magnitude and persistence of Renaissance’s returns. The fact that they are something like 70% on average over a period of 20-30 years is much more significant than an array of “but Buffett and some other hot managers’ alpha has decreased or vanished during a relatively short period of 5-15 years”.
Of course, the latter may indeed point to fundamentals-driven alpha being incorporated into the market, i.e., the market has become more efficient, perhaps because of entities like Renaissance.
Let me say it in familiar terms.
Look at PSST. Is there anything in the definition about price?
No, it is all about sustainable patterns, patterns of flow that do not bunch up in congestion.
What is price? dollars/item, a ratio. Why does that ratio work? Because the numerator works, the denominator is already just fine.
But arbitrage points exist in money, folks. For example, when the central banker says, ‘We are introducing an accommodative arbitrage point’ They actually announce it, and when announced the marginal traders look for portfolio congestion that result as folks change their portfolio suddenly. Read Arnold’s description, he is describing just that.
Just three day ago Bernanke claimed the Fed had more tools. What he means is the Fed has more methods to introduce arbitrage points. Who removes the arbitrage points, the super wealthy via the primary dealers, in fact the super wealthy just earned 25% last year doing just that. They do it by lkaw, the primary dealers system is a legal, required framework.
Central banks insert arbitrage with their current structure. We actually have numbskull economists who think this is a good idea.
An excellent essay, and perhaps an excellent illustration of Merle Kling’s first law: “1. Sometimes it’s this way, and sometimes it’s that way.” The review convinces that it is difficult to draw reliable inferences from particular case studies. Buying and selling between willing, uncoerced parties need not mean very much at all. The federal Thrift Savings Program buys huge numbers of shares on a regular basis for its federal employee clients and their index fund accounts and one wonders how much of the expectation that it will continue to do so is baked into the price of stocks.
The ECMH has also been used to infer that firm management faces capital market discipline but that case seems tenuous at best for certain segments of large publicly traded firms commonly held in index funds. This suggests that there is a non-zero Ponzi scheme element to stock prices. Straussian capitalists like Tyler Cowen who want to allow firm managers to subordinate shareholders interests to firm management’s favorite social justice causes may pull the card that causes the house to tumble.
This may simply be confirmation bias on my part. Lately I have been under the spell of Gary Saul Morton, who wrote of Tolstoy’s War and Peace at The New Criterion last March (sorry no link, yes, I am using my cell phone) that it illustrates how the regularities in the world are overwhelmed by “sheer contingencies.” One wonders how much Tolstoy Hayek read: “The curious task of economics is to demonstrate to men how little they really know about what they imagine they can design.” And sometimes it’s this way. But mostly it is difficult to draw reliable inferences.
Regarding market efficiency, it’s insane to classify the results of Renaissance Technologies as participation in a coin flipping tournament. They don’t just flip one coin per year. They flip thousands of coins per day.
I think the better framework to apply to RT, Buffet, and Gross is from machine learning:
– There is a hidden true “target concept” (a function of how the market actually works over some period of time)
– There are lots of investors trying different “hypotheses” (models of how the market works, which dictate trading strategies) trying to match the target concept
These hypotheses are not created uniformly at random like the results of coin flips; they are informed by experience, intuition, analysis, etc. In other words, some (varying) amount of domain expertise is applied to the problem in generating hypotheses.
Over time, some of these hypotheses prove to be profitable, and some don’t. But there’s another idea from ML called “concept drift,” which is that the true underlying function may change over time. This almost certainly applies to the market, especially since it’s a dynamic, competitive market with feedback loops. So some hypotheses that fit the true concept well in some circumstances and time periods don’t work well in others. When the market concept changes (sometimes called market regime), which hypotheses have success will also change. There is certainly some element of luck with timing and other aspects of the process, but that doesn’t make winning strategies equivalent to coin flips.
It’s also the case that a trader can have a hypothesis that is “wrong” (in the sense that it doesn’t match the true concept well), but still leads to profitable actions for some period of time by luck. And then there are lots of hypotheses that are wrong and are never very profitable. These “wrong” hypotheses are more like coin flips. But sustaining them over decades is improbable, no matter how many participants there are.
So are markets efficient? Depends on how literally and broadly you are trying to apply it. It seems that some people are able to reliably make profits over long periods of time over thousands of trades, because they have hypotheses about how the market works that are close to reality for some period of time. But these people either have information that is unavailable to the market, or use information in a better way than the rest of the market. It sounds like RT was doing the latter.
The way I tend to think of EMH is that it’s hard for almost everyone to beat the market in the long run. That doesn’t mean it’s impossible over stretches for some people. I also like the description from the book “The Quants” about the “piranhas” that make the market efficient:
https://books.google.com/books?id=-ydNYWGIussC&pg=PA85&lpg=PA85&dq=fama+piranhas&source=bl&ots=yn7_D1-HVb&sig=ACfU3U3dikQilol0eaLjuXiCnSLueW5Bhg&hl=en&ppis=_c&sa=X&ved=2ahUKEwi8r-nB1vTmAhWQQc0KHciTCywQ6AEwCHoECAoQAQ#v=onepage&q=fama%20piranhas&f=false
The basic gist is that markets become efficient by people capitalizing on inefficiencies. It’s not unreasonable that a single person or group capitalizes on the same inefficiencies until the rest of the market becomes wise to it.
Straight from the Fed mouth, via Mish:
https://moneymaven.io/mishtalk/economics/foreign-banks-benefit-more-from-fed-interest-on-excess-reserves-weR09OWfBEmeBGK-ETlwzQ
In the post-crisis environment of abundant reserve balances, most fed funds borrowing transactions are motivated by “IOER arbitrage”—borrowing overnight at a rate below the IOER rate, leaving the funds at the Federal Reserve overnight to earn the IOER rate, and earning a positive spread on the transaction. Foreign banking organizations (FBOs) currently have a large presence in the fed funds market because of their borrowing advantage over domestic banking institutions*. Since FBOs do not offer retail deposits, they are exempt from paying Federal Deposit Insurance Corporation (FDIC) insurance assessments, which all domestic banks must pay*.
—-
The Fed inserts arbitrage, on purpose, and it is mandated by law in both the primary dealer system and in the twin directives. Our macro economists with their hydraulic macro have completely fouled central banking theory. Their is nothing hydraulic about currency banking because currency banking is all about quantization, finding the appropriate ‘chunks’ which minimize transaction costs, the Baumol process, I have named it. In theory it is all about congestion management, price is a secondary proxy. This description of excess reserves is mostly a match to what Arnold was saying in general terms.
So, our central banking is mostly about legally requiring the rich to get richer on the backs of taxpayers. Fools who believe in hydraulic macro do not belong in policy circles.
If Renaissance didn’t exist, is there any reason to believe that there wouldn’t be enough liquidity for markets to function in their primary task (raise capital for investment and signal the relative value of different investments via price signals).
I mean I have to think that someone buying their S&P 500 fund is going to be able to get that order filled immediately at the going price whether Renaissance is there or not.
Liquidity as a justification has always fallen on flat ears for me. For it to be a real reason you’d have to prove to me that markets wouldn’t clear (obviously bad) or would only clear at huge price differences to fundamentals (breaking the price signal). I’m not sure that is the case here.
Seems these PhDs could have been inventing things that increased the pie, rather then grabbed a bigger slice.
Talk to some quant managers about the trades they could not execute because just their being in the market drove prices to where there was no profit in the trade.
To a firm like Renaissance, liquidity refers to how large a block you can trade without moving the price.
This implies that they make a very large number of small trades and very few large trades.
There is no “market”, it’s just an abbreviated way of saying “millions of buying and selling decision makers deciding on agreed upon prices for millions of transactions, agreeing based on their own current states of information and desires”.
For each transaction, there are two decision makers, a buyer and seller – the price is their agreement point. But the states of info change every second as there is always “news” with varying relevance. For any stock to increase its price, there must be more decision makers who decide it’s better to buy than to sell. Whoever buys the most of that stock first will have the highest returns from that stock when they sell.
Ren Tech’s business model was being the first buyer, for stocks going up; and the first seller (of owned stock) for those going down. Literally, “market leaders”; first in line towards making the “market” more efficient. The first of those decision makers whose aggregate decision making is resolved in a price change for that stock (or any traded financial product).
Note the key about how an “efficient market” is defined:
future returns cannot be predicted. There is no absolute “real” price that is being moved towards, there is only a series of price changes based on decision makers. So for most people, most of the time, the market is pretty efficient. Partly because of RT and other traders trying to be first in line for trades after new info. In a zero-sum sense, they can only gain by a purchase when the seller loses by selling too early.
All the big managed funds and traders are implicitly promising to have better than random returns, tho studies show most managed funds get less return than a no-load mutual index fund. Yet it is because of RT and other successful traders that so many fund managers believe they can get “first in line” at those 51% likelihood going up buying lines. It is the total of such traders which makes the market efficient, making it so that average investors cannot expect to earn excess returns.
But they can expect average returns, while hoping for excess returns, paying the price of such hope with smaller total returns more often.
(2) After tax Profits in the Financial Sector should probably be more like 1% of total US after tax profits, rather than 4%, so they are undertaxed. They are just one of the more overpaid actors, but critics should have some targets (even if arbitrary, like 1%) before just saying too much profit.
Interest on (Excess) Reserves is also terrible — the banks should be pushing to loan that money out, with required non-interest bearing Reserves as part of the price to be in the business.
Missing the entrepreneur’s trial and error discovery.
Upon closer reading of Arnold’s critique of Keynesianism it seems there is a similar criticism of what an “efficient market” is.
the fundamental flaw in Keynesian economics is that it relies on aggregation and thereby ignores the need for discovery and adjustment. Treating the economy as if it were a single GDP factory is a defect in both popular Keynesianism and rigor- seeking Keynesianism.
The flaw in the efficient markets definition is the implicit idea that there exists some “efficient price” of a stock, and thus an aggregation of all such prices would be the efficient market price. This ignores the discovery/ adjustment needs. What RT is doing is betting/ investing in a direction of change of the price of a stock.
This financial investment benefit shows investor sentiment rather than any other short term benefit to the company whose stock is being invested in. It’s necessary but far from sufficient to help employment or even new business investment.
Profits from such trading are far less socially useful than profits from Venture Capital and even IPO investment.
Capital gains taxes on Trading are less harmful to society than income taxes, so I’d happily trade less income taxes for more cap gains taxes.