the futility of all human striving
April 4, 2019 12:03 PM   Subscribe

The Unintended Impact of Academic Research on Asset Returns: The CAPM Alpha, Alex R. Horenstein - "This paper explores a channel whereby asset-pricing anomalies can appear as investors alter portfolios according to findings in academic research. In particular, I find that assets with low realized CAPM alphas [wiki] outperform those with high ones, but only after the CAPM’s publication in the 1960s."

some references: What is the Difference Between Alpha and Beta?
The Capital Asset Pricing Model: Theory and Evidence[PDF], Eugene F. Fama and Kenneth R. French, Journal of Economic Perspectives—Volume 18, Number 3—Summer 2004—Pages 25–46
Understanding The History Of The Modern Portfolio

Matt Levine, Money Stuff: Buying The Good Stocks Can Be Bad
The capital asset pricing model provided a new and more formal, and extremely popular, way to identify the stocks that were good. So people went and bought the stocks that it identified as good, pushing up their prices (and pushing down the prices of the bad stocks). So the stocks that were good turned out to be bad, in the sense that their subsequent performance was worse than that of the stocks that were bad
...
It is such a pleasing story about the futility of all human striving, you know? You try to figure out how to tell a good stock from a bad stock, and you do it, and now you know which stocks are good and which are bad, but so does everyone else, so the good stocks go up and the bad stocks go down until they are both properly priced, and then in fact they overshoot (because everyone wants the good ones) until the good ones become bad and the bad ones become good and you have to start all over again. No wonder indexing is popular.
posted by the man of twists and turns (40 comments total) 13 users marked this as a favorite
 
Is there a word missing in the first sentence of the abstract? My brain can't quite parse it.
posted by Drowsy Philosopher at 12:15 PM on April 4, 2019


Equally, post-earnings-announcement drift was a much more effective investment strategy before academics started writing papers about it. Then more investors starting trying to use that anomaly by responding more rapidly than the rest of the market, preempting that slow drift. Which in turn raised the speed at which the market responded to the new information.

So now, the post-earnings announcement drift happens in milliseconds.

At one level, that's a success. The market is correctly pricing based on new information. At another, it's a huge amount of human effort wasted in building systems that can take in, process, and respond to market info. The return on that effort is fleeting, being purely positional.

Anyway, I'll be over here with my low-fee index funds, free-riding on everyone else's attempts to set prices.
posted by happyinmotion at 12:20 PM on April 4, 2019 [4 favorites]


Along the same theme, Behavioral Feedback: Do Individual Choices Influence Scientific Results? "When a particular health behavior becomes more recommended, the take-up of the behavior may be larger among people with other positive health behaviors. Such changes in selection would make it even more difficult to learn about causal effects. ... I test for evidence of these patterns in the context of diet and vitamin supplementation ... I show that selection varies over time with recommendations."

In other words: 1) Someone says vitamins are healthy. 2) People who care about their health - the type who exercise and each lots of veggies - are more likely to start taking vitamins. 3) A researcher looks and finds, yes, people who take vitamins are healthy!
posted by Mr.Know-it-some at 12:40 PM on April 4, 2019 [3 favorites]


Anyway, I'll be over here with my low-fee index funds, free-riding on everyone else's attempts to set prices.

Matt Levine: Are Index Funds Communist?
The function of the capital markets is to allocate capital. Good companies' stock prices should go up, so they can raise money and expand. Bad companies should go bankrupt, so that their resources can be re-allocated to more productive purposes. Analysts should be constantly thinking about whether companies are over- or underpriced, so that they can buy the underpriced ones and sell the overpriced ones and keep capital flowing to its best possible uses.

But when those thoughtful active analysts are replaced with passive index funds, the market stops serving that function. Whatever the biggest company is today will remain the biggest company tomorrow, and capital will never be allocated from bad uses to good ones. Indexing is cheaper, yes, but that's because active management has positive externalities, and if no one will pay for it, those benefits will disappear.
posted by BungaDunga at 12:44 PM on April 4, 2019


Problem is, the bad companies have so much capital already that they can avoid going bankrupt by buying the good companies. When those good companies go bad, just buy the next good company.

Antitrust problems are a function of total scale, and with the total scale of our economy now so large, they occur long before a company achieves monopoly status. We're long past the days when corporations were limited to a single market; diversification keeps the zombies going.
posted by I-Write-Essays at 12:52 PM on April 4, 2019 [2 favorites]


I'm pretty sure there's been academic research on how index funds are best. So doesn't that mean that they'll end up being overvalued?
posted by clawsoon at 12:52 PM on April 4, 2019


clawsoon - I don't know what I'm taking about, but when I was a kid (the 90s) I remember sound financial advice being to stay out of the market and investing in bonds. These days, I hear nobody and I mean nobody talk about bonds or other vehicles besides the market.

So.... yes, maybe index funds are over valued?
posted by rebent at 1:04 PM on April 4, 2019


I love this. Now that this paper has been published, what will happen? It is wonderfully paradoxical. If its predictions are correct, its prediction will be wrong.
posted by clawsoon at 1:04 PM on April 4, 2019 [1 favorite]


> I love this. Now that this paper has been published, what will happen?

I think traders already know this, at least at the major investment institutions. This is why they guard their trading universes and models so jealously. Competition is about being the first to realize a new insight, and finding the next new insight before everyone else figures out the one you just found.

But, if there were something to happen because of this becoming common knowledge, it would be attempts to suppress new research that kills someone's golden goose.
posted by I-Write-Essays at 1:10 PM on April 4, 2019 [1 favorite]


So.... yes, maybe index funds are over valued?

The funds can't be, but the underlying stocks could be overvalued, and if so... that's sort of a baseline requirement for a bubble.
posted by BungaDunga at 1:12 PM on April 4, 2019 [2 favorites]


Alternative first link without a paywall. (Or perhaps a "you must give us your personal information" wall? Whatever it is, I can't access to the paper from within a couple of research universities, which is kind of surprising.)
posted by eotvos at 1:31 PM on April 4, 2019


I remember sound financial advice being to stay out of the market and investing in bonds. These days, I hear nobody and I mean nobody talk about bonds or other vehicles besides the market.

Of course people still invest in bonds, but the rise of the home personal computer and Microsoft Excel making individual pricing decisions easier to map, the rise of asset pricing models handled by indexed funds, the lowering of costs of personal investing (without a broker), the university research of efficient portfolio allocation, which can nearly be achieved through stocks alone, and the lower interest rates vs the '80s all make full bond portfolios less attractive.
posted by The_Vegetables at 1:41 PM on April 4, 2019


> I remember sound financial advice being to stay out of the market and investing in bonds. These days, I hear nobody and I mean nobody talk about bonds or other vehicles besides the market.

Well, the illegality of index funds[*] notwithstanding, maybe I can interest you in some bond index funds? Or the equivalent bond exchange-traded funds?

[*] BungaDunga's comment above has one of Matt Levine's ha-ha-joking-but-wait arguments for why index funds are anti-competitive, and he's been marshaling the pros and cons of this argument for a while now. Even in its maximalist form, the argument is hard to dispute.
posted by RedOrGreen at 2:56 PM on April 4, 2019 [1 favorite]


It's just a reformulation of the value factor.
posted by JPD at 4:49 PM on April 4, 2019


The point of index funds is a cheap way to capture the equity risk premium over time. They can't by definition be under or overvalued. The market can be, but on a long enough time frame (10years+) the odds of capturing the ERP is extremely high. By definition an index fund has an alpha of zero.
posted by JPD at 4:51 PM on April 4, 2019 [4 favorites]


Of course people still invest in bonds, but the rise of the home personal computer and Microsoft Excel making individual pricing decisions easier to map, the rise of asset pricing models handled by indexed funds, the lowering of costs of personal investing (without a broker), the university research of efficient portfolio allocation, which can nearly be achieved through stocks alone, and the lower interest rates vs the '80s all make full bond portfolios less attractive.

I like do this for a living and can't for the life of me make sense of 90% of this comment.
posted by JPD at 4:53 PM on April 4, 2019 [2 favorites]


My favorite factor bring competed away was the earnings quality work from the late 90s. In 18 months you could watch it get arbed away.
posted by JPD at 4:55 PM on April 4, 2019


>Matt Levine: Are Index Funds Communist?

Ha ha no.

It's true that index fund investors are free-riding on active investors' efforts to discover prices. That's just the Grossman-Stiglitz paradox as applied to share markets. Sure, if a market has too many passive investors then it won't accurately uncover prices or respond to new information. However, what does "too many" mean? Markets operated for decades with literally a million times fewer trades than at present. That implies that the vast majority of buying and selling, like 99.999x% is of vanishing relevance for price discovery.

That also implies that the vast majority of active investors are of minimal value. They are not needed to discover prices. More than that - the track record of almost all active investors is negative - literally the average return of active investment funds tends to the market return minus the fees for the active management. We'd all be better off without them.

Hence it's hilariously self-serving when active fund managers try to justify their activities by pointing out the evils of passive investment. No-one wants to admit that their lives are pointless, yet there is solid evidence that they're playing a zero-sum game and losing.
posted by happyinmotion at 7:03 PM on April 4, 2019 [4 favorites]


Well, losing for their clients. But if it puts their own kids through college...
posted by darkstar at 3:11 AM on April 5, 2019


The math of active has to be the market return -fees. How can't it be.

Even the very few active guys who outperform are mostly just taking a big factor skew either to value or momentum. The question of course is how much is it worth to pay for someone keeping you from being your own worst enemy. They don't need to allow you to beat the market by much for it to be very material to compounded returns.

That's why asset classes with real investible indexes ( which isn't actual as common as you'd think) are increasingly pricing at the institutional level as "passive + a share of the excess returns you generate w/ a highwater mark".

The issue with that of course is that it makes startups in the space impossible, and yet startups tend to deliver better performance.

But active equity is nothing like the ripoffs in the alts space.

Also there are lots of ETFs out there which are technically passive, but are economically active. If you buy a biotech etf in some greater proportion than the market, you are explicitly making an active allocation bet. Which is exactly contra to Bogle et al.
posted by JPD at 6:45 AM on April 5, 2019 [1 favorite]


BungaDunga: The funds can't be, but the underlying stocks could be overvalued

I don't understand this. If all the underlying stocks in an index are overvalued, doesn't that mean that the index itself is also overvalued relative to the universe of possible investments?

Like... sometimes index funds are outperformed by plain old cash, especially in the medium term, no?
posted by clawsoon at 6:53 AM on April 5, 2019


Yes. A market can be overvalued in totto. However even buying an overvalued market the historic data would show it doesnt take that long for stocks to correct their underperformance.

Actually there is an argument that buying an equal weighted index is a way to minimize the exposure to overvaluation, and except again, that's an active allocation decision.

You kinda just wish we could write everyone an ERP vs market return swap for the 401k. Of course I guess that's essentially what DB plans are/were doing.

Humans are incredibly shitty at dealing with volatility.
posted by JPD at 7:09 AM on April 5, 2019


However even buying an overvalued market the historic data would show

This is the half-serious point I'm thinking about: Now that three or four decades of research and investment advice has made these historical fact known to everybody and their barista, won't that result in the value factor of the stock market itself being erased?
posted by clawsoon at 7:24 AM on April 5, 2019


That's not what the value factor is.

Value as a factor hasn't worked for the last 15 years outside of a few small windows ('09 in Dev world, '12 in Europe, 16 in EM and US)
posted by JPD at 7:29 AM on April 5, 2019


Damn... I was hoping I was using the term correctly, but failed. Thanks for the correction.

Would "factor" by itself have been closer to what I'm looking for? I.e. "here's something we've discovered that'll give us excess return, but once everybody knows about it it'll stop giving us excess return"?
posted by clawsoon at 7:36 AM on April 5, 2019


The out performance of equities over time might be referred to as the "Equity Risk Premium". Essentially equity returns are more volatile, so investors demand a higher return than less volatile returns. It can't be arbed away because it's not a free lunch. You have to be able to stand large drawdowns at any given time and we aren't wired to feel good about that. We're wired to cut losses.
posted by JPD at 7:57 AM on April 5, 2019 [1 favorite]


This idea is not new to Game Theory, nor economics. Keynes was talking about it almost 100 years ago (Keynesian Beauty Contest).

Douglas Hofstadter proposed a solution which I quite like called 'superrationality', (wikipedia) where a player (investor, in this case) should reason that whatever solution she ultimately decides on, it will be the solution that everyone decides on, by virtue of all players' shared information and rationality. Therefore, the best course of action is whichever one leads to the best outcome, assuming everyone else takes the same action.

I love it for two reasons: First, it's a very mathematical feeling solution to a very human problem. Which makes it a bit silly, especially when very simple explorations of the idea make it clear that people are not superrational in practice (as in here, where the experimental results cited match very closely my results with many groups of students over many years).

But the other thing I like about superrationality is it comes awfully close to Kant's moral imperative. I'm not the only person to make this connection; A guy named Binmore wrote basically a whole book about it (Google Books link).

If only our lauded quants had a more liberal education, they might be better prepared for the situation they find themselves in.
posted by dbx at 8:05 AM on April 5, 2019


The value factor and the momentum factor represent two statistical anomalies which show that the stocks which are rated the lowest on measures of asset value to market value or earnings and cash flow as a % of market value or those stocks whose prices have been going up the fastest tend over time to outperform the broad market. A prominent quant guy jokes its always value an momentum everywhere. Meaning that this out performance occurs in nearly every asset class.

Ironically this is all predicated on markets being irrational. Trust me many many many quants and fundamental equity managers have read Hofstadter.

Don't forget George Soros is a Popper acolyte.
posted by JPD at 8:11 AM on April 5, 2019


"Good for everyone" in this case sounds like it might boil down to "good for the investor class."
posted by clawsoon at 8:13 AM on April 5, 2019


I like do this for a living and can't for the life of me make sense of 90% of this comment.

Then you should possibly study the history of your industry more? To make it simpler, more people invest in stocks now than in ye olden days because technology, cost, and gatekeepers power have changed.
posted by The_Vegetables at 8:56 AM on April 5, 2019 [1 favorite]


It can't be arbed away because it's not a free lunch. You have to be able to stand large drawdowns at any given time and we aren't wired to feel good about that. We're wired to cut losses.

So other opportunities are based on data that people don't yet know about, or minor quirks of human psychology, while the opportunity provided by the stock market itself is based on something fundamental about human nature which can't be educated away? Is that a fair restatement?
posted by clawsoon at 9:06 AM on April 5, 2019


> To make it simpler, more people invest in stocks now than in ye olden days because technology, cost, and gatekeepers power have changed.

Is there an investment class where that isn't true? If you think computers help with equities, consider how many CUISPs are trading at any given moment.
posted by pwnguin at 9:23 AM on April 5, 2019


> So other opportunities are based on data that people don't yet know about, or minor quirks of human psychology, while the opportunity provided by the stock market itself is based on something fundamental about human nature which can't be educated away? Is that a fair restatement?

Well, if I'm amateur explaining:

1. A pure total-market index fund, representing the "complete" stock market, should have returns that exceed a "safe" asset (Treasury bills) by exactly the amount of the equity risk premium. That is, the entire stock market is subject to volatility, recessions, herd mentality, whatever, and you get paid a premium for taking on the risk that your money might not be available exactly when you want it. If you don't want that risk, invest in US Treasuries or use a bank account.

2. Beyond that, even investing in a broad index fund that covers a single market segment (biotech stocks, "value stocks", dividend-paying stocks, stocks of companies with CEOs who do not play golf) is a bet that the investor has more/better information about how that sector will perform, compared to the market averaged index funds.

3. Further, investing in a single company (AAPL, TSLA, WFC) is a bet that the company is going to perform better than the market average, and that extra return isn't already priced into its stock. Maybe you know they will invent the next iPhone - but maybe their CEO will tweet materially misleading information and get fined by the SEC, or maybe they have been opening fake bank accounts.

Most market investors are doing (2) or (3), but in aggregate, they must be wrong. And if you're investing with an active fund managers, not only are you betting that they will beat the market, you're betting that they will beat the market even after their fees.
posted by RedOrGreen at 9:37 AM on April 5, 2019


wait, what does this mean???

that extra return isn't already priced into its stock.


if the extra return is priced into the stock... then why buy stocks in the first place? If the price is equal to the expected return, then wouldn't you only get back out.... an even return? Or something? Net neutral return?

Where does the increase in value come from anyway? Where is the value generated and who gets it?
posted by rebent at 9:55 AM on April 5, 2019


I think what they mean is that the future faster than market growth in cash flow/earnings/revenue isnt already understood and priced by the market.

Essentially what this paper says is that not only is it priced, it turns out it's often overpriced.
posted by JPD at 10:11 AM on April 5, 2019 [1 favorite]


> if the extra return is priced into the stock... then why buy stocks in the first place? If the price is equal to the expected return, then wouldn't you only get back out.... an even return? Or something? Net neutral return?

In aggregate, yes... In the theoretical perfect capitalist utopia, every stock is worth exactly its future discounted earnings, and the only premium comes from your willingness to take on the extra risk.

But information is never distributed perfectly evenly, and some things are transformative and cannot be priced into future expectations.

Also, I work with astronomical numbers, not financial numbers - so my terminology is all over the place, sorry. So I'm way out of my wheelhouse on this answer, and others should correct me. But:

> Where does the increase in value come from anyway? Where is the value generated and who gets it?

My brother-in-law was a big believer in Steve Jobs, and was willing to bet that when Jobs came back to Apple, he would add a lot of value to the company. The rest of the market was not so sanguine. So the market sold him Apple stock for much less than they should have, given what it was worth a decade later. And AAPL increasing in value allowed Apple to attract and retain talent with stock options, issue more stock to raise capital, etc.

Asymmetric information is the issue here. And unless you have specific insight, it's almost certainly the case that you're better served by a broad index fund.
posted by RedOrGreen at 10:17 AM on April 5, 2019


That is, the entire stock market is subject to volatility, recessions, herd mentality, whatever, and you get paid a premium for taking on the risk that your money might not be available exactly when you want it.

I'm seeing (vaguely, sorry) two kinds of premiums here:

1. Investors on average will be scared of the risks of volatility and pay less for index funds than their return is worth. Therefore, anybody who invests in an index fund earns a premium. This seems like the free lunch that would be (or has been?) eliminated by everybody knowing about the stock market premium.

2. An index fund investor who doesn't have a liquidity crisis which forces them to sell at a bad time will earn a premium if they stay invested for long enough. However, if the free lunch from point 1 has been eliminated, their excess returns will be balanced by all the people who lose money on index funds via forced selling or bad timing.

Does that make any sense? If it does, it seems like maybe (?) the type 1 premium existed in the early 1980s but has been mostly eliminated now. (?)
posted by clawsoon at 10:30 AM on April 5, 2019


The price of 1) is the volatility. There is always someone arguing the ERP has changed, and it might but so far hard to see clear evidence. But dealing with volatility isn't a free lunch. There are lots of papers on risk aversion out there. This is related to that.

The underlying broad overgeneralization of financial markets is risk and return are inextricably linked. The reason why equities are riskier is because they are the residual of the firm value. I.e. if I buy a bond my return is capped, but I get paid back first and in cash that I'm contractually obligated to. As an equity holder I have no contractual cash payment and I only get what's left of the firm.
posted by JPD at 10:38 AM on April 5, 2019


There is always someone arguing the ERP has changed, and it might but so far hard to see clear evidence.

This is exactly the question I was fumbling towards. Thanks to you and RedOrGreen for leading me there!

Are there any good standard review papers about this question out there that you know of? Summaries of the evidence and theories for and against?
posted by clawsoon at 10:49 AM on April 5, 2019


Soooosoooosooo many. Hard to pick one. Brad DeLong usually writes about new studies from time to time.
posted by JPD at 10:57 AM on April 5, 2019


« Older Delores: love story of how a man who had gas in...   |   “...but how realistic is it really?” Newer »


This thread has been archived and is closed to new comments