Even God Would Get Fired as an Active Investor (2016) | Hacker News Reader

Even God Would Get Fired as an Active Investor (2016)

alphaarchitect.com

141 pointselemeno51 comments

This is silly. The author points out the problem of making short-term evaluations of a long-term strategy and yet he never actually says that. Also, his hypothetical omniscient "god" can see 10 years into the future but doesn't allow himself to rebalance the portfolio more often than 5 year intervals. I'm not sure what point he's trying to make or why.

I think he's just pointing out how hard it is to be a good active investor, basically exactly what Warren Buffet thinks (and won that bet over).

Warran Buffet's opinion isn't relevant here. This is presented as a study providing empirical evidence in support of it's conclusions and as such needs to stand on it's own merit.

Given the hyperbole and deity references, I thought the article was more for amusement than anything else.

I think this article is attempting to make a different point.

It essentially argues (unconvincingly IMHO) that it is difficult to _distinguish_ a (very contrived type of) "perfect" active investor from a "bad" investor whereas Buffet argues that it is difficult to _be_ a good active investor.

These are not necessarily inconsistent but they do have opposite impacts on investment decisions.

To my mind this is dishonest marketing "research" whereas Buffet is making an important point.

Oh absolutely, I thought that type of active investor was so contrived that this was just an article written more or less for entertainment (to point out how hard it is to be an active investor).

Indeed. Reading my own words, I sound rather more worked up than I am :)

I do think this article is a bit sneaky dressing itself up as entertainment since it is published by a business that profits from people deciding to invest.

It does provide a means to illustrate the agency problem though.

If I had perfect knowledge of the future and had to buy and hold a given set of positions for myself for some fixed period, I would simply choose those that maximised my return at the end. HOWEVER if instead I had to invest on behalf of others (in return for some fee) and was subject to being fired, I would be tempted to choose a different (thus suboptimal) set of positions such as those with smallest drawdown / maximised minimum rolling quarterly Sharpe / ...

His underlying point is broadly right - "past performance is no guarantee of future success" and as such historical (which includes current at the time of making an investment decision) is a fairly poor predictor.

The argument that the fund research sellers peddle is that "we recommend based on the investment methodology in use, the processes, risk management, the people, etc etc". Otherwise what would they be doing that a simple google search couldn't?

Again, arguably far better metrics to find an active fund that will on average out-perform. They're not wrong.

>how hard it is to be a good active investor, basically exactly what Warren Buffet thinks

In broad brush strokes, yes. However this blog post uses a different math scenario.

- Warren Buffett's premise for winning the bet was the hedge funds' 2% "management fee" and 20% carry. Therefore, any attempt to beat Warren's passive investing starts with a handicap of minus-2% and has to have bigger positive returns that overcome it.

- This essay is about long-term "God clairvoyance" of _eventually_ being correct is negated by short-term negative returns which make people "fire" God. E.g. the investor might have a 2-year lockup of his funds before he can redeem them. E.g. After 2 years, the investor sees that the hedge fund is losing money -- but doesn't realize that it's a temporary dip. Therefore, he redeems his money (aka "fires God") and never got see that God was ultimately correct.

(Or put another way, if the lockup period and the fund's entire lifetime were exactly the same, the blog post couldn't be written.)

I've seen (and implemented in code) some pretty brutal liquidity algorithms. I've no doubt God could use such algorithms to always ensure positive performance on those dates on which liquidation is possible.

Funds might be locked up for two years, and afterwards only redeemable on the first day of the fiscal quarter. On the first liquidation event, only 25% of the funds are redeemable. If you submit notice on the first day of the next fiscal quarter, 33% is redeemable, followed by 50% and 100%. If you miss a quarter, the sequence starts over.

That one is not even terribly complex.

You state it much more clearly than the author does.

While I would not go so far as to call it "silly". I think an interesting point is obscured for the purpose of a dramatic headline.

Having perfect foresight 5 years in future and constantly rebalancing the portfolio means that the portfolio will never match the perfect foresight. It would invalidate the point that article is making.

That kind of strategy could lose money related to index if five year prediction and business cycles cancel each other (think counterphase signal).

It would be very interesting to see simulation of perfect foresight rebalancing strategy in different timefrimes.

Yep, the 5 year window seems reasonable, but starting from a single point doesn't. It'd be very interesting to see a breakdown of when 'god' would be fired for each of the possible starting points.

Agreed. Simulating with sliding window over longer period would tell how the perfect foresight strategy performs. Now it's just single probably hand picked data point.

from the author: "1yr look ahead, rebalance each year? Yes. We also looked at monthly. If you shorten the horizon the returns get silly and the drawdowns go down. At 1yr you can still hit 20%+ dd. At monthly you become a trillionaire with essentially no risk"

They are trying to say "Please don't fire us - even God would have down years!". Although this depends quite a bit on when you start. They start in 1927, if they started in 1932, God might never be fired.

I think it says that in bold underlined text in the "Conclusions" section at the end.

It's clickbait is what it is. They've defined god with these arbitrary restrictions that correlate neither with omniscience nor with what an actual fund manager could know. Obviously even without allowing rebalancing, god could optimize for less-volatility and keep his job while still knocking it out of the park (just slightly less).

The point is fund managers often know something is going to happen in the future, but not exactly when. Of course God would know exactly when the market will tank and rebalance accordingly. However a fund manager doesn't have quite that much insight.

The fund manager who got the most out of the credit default swaps in the long run had to pull large trick in his fund to prevent his investors from firing him just months before the crash proved he was right. Inf act there was a delay between when the losers realized there was a problem and quit throwing good money after bad and when the investments actually went bad. Even if you invested on the very last date you could, you would have lost money for a short time (I'm not sure how long, probably just a couple weeks) as the losers used every trick they could come up with to prevent the crash.

This is the basic risk vs. return tradoff in investing lesson. Risk in stock market can be reduced with diversification and longer time-span.

Don't listen active investors unless they have most of their own money in the scheme and manage your money on the side (like Warren Buffet does). Figuring out winning market strategy and selling it as a service to others and living off management fees indicates that the winning strategy involves separating fools from their money.

And thanks to competition between capital suppliers (ie you) the successful active managers will raise their fees until they eat all the excess return.

The only escape from the efficient market is investing in markets that are such small pockets and difficult to reach that professionals ignore them. But that's almost impossible to square with diversification.

Active fund management competes with clever marketing.

Most active fund management companies have several actively managed funds. They close or merge low performing ones or turn them into closet index funds (index hugging). They market those funds that happen to perform just now and show 1-3 year performance numbers. Newcomers can sell simulated profits without fees or trading costs.

Yes, though the passive guys can hire some clever marketing people just the same.

> And thanks to competition between capital suppliers (ie you) the successful active managers will raise their fees until they eat all the excess return.

What actually tends to happen in practice is the really successful active managers cease accepting money, because they can effectively print it. All strategies have capacity constraints and managing investors is not so fun. If you can generate positive returns without needing to pool risk, you don't need investors anymore.

Yes. Same effect in practice: you can't give your money to people who produce above market returns after fees.

Probably not a surprise, but this also applies to value investing.

If you construct a perfect portfolio using god-like knowledge of a company's performance as measured by future free cashflow then you will also have drawdowns and down years.

Makes you think about how the market has mutated from its original purpose over time.

And obviously you would have those years: unconstrained optimization for future free cashflow differs in general from any kind of optimization that introduce the constraint that you can't have down years.

(There are only a few instances where optimizing for two different things, eg maximum flow in a network and looking for the minimum cut, produces the same answer. And those cases are mathematically significant.)

I'm making the point that these two optimizations should be fairly similar, but they are not.

Perhaps similar, but not exactly the same. So with a bit of datamining, you should see arbitrary differences.

The thing I don't get here is, if I'm God, then aren't I going to sell off right before the drawdowns and then buy back in at the local minimum?

Can you?

I can sell every stock I have tomorrow and nobody will notice I'm small potatoes that nobody on wall street cares about. When you have a significant amount of money it isn't that easy.

It is just easy for me to get 40% returns every year, just put everything into the one stock that will be up the most tomorrow morning. In just a couple years of God like foresight though I can't do that: I will be worth so much that my very act of buying/selling stock will change the price. Every time I start a trade stock for that company will stop trading while wall street tries to find buyers/sellers for my stock. The stock might be selling at $10/share, but when I actually buy the price is $20/share.

For a few years I can counter that by splitting my orders between several stocks, but each time I have to do that my returns go down just a little. Eventually even that runs out: after 60 years of 40% growth I would be worth more than all stocks in the world combined.

This is the money quote (no pun intended): "These results highlight the fickle nature of assessing relative performance over short horizons."

How many times have you told a fund manager, "Gee your fund lost 10% last year when the S&P500 was up 8%, how did you manage that?"[1] Only to move all your money into some other fund and have it get worse returns than the one you moved your funds out of? The point the article tries to illustrate is that evaluating a fund's strategy should be separate from evaluating a fund's returns. And if its strategy is sound, then even if it under performs other metrics over a shorter term, it out performs over a longer term.

And yet there aren't really any investors who are willing to 'take it on faith' and stick around for 10 years to see if their strategy is sound.

[1] Ok probably not a lot but I did get to ask one this question. They did not respond effectively.

Since you are comparing to the S&P 500 anyway, why pay the extra fees?

TL;DR: If you look at the best stocks, measured by 5-year performance, they sometimes go way down within those 5 years (mainly because of recessions). Not very surprising.

But if you also short the worst performing stocks, you still have massive drawdowns. That is somewhat more surprising. I would like to see an explanation for that.

Is it because the top decile stocks are higher beta than the bottom decile stocks?

Is it because of the selection of rebalancing periods? I.e. if the big crashes happen towards the end of a five year period then perhaps the bottom decile stocks have already dropped significantly and have nowhere to go while the top decile stocks have risen significantly and have plenty of room to fall.

This is actually a well-known issue though probably never presented to a non-trader audience like this before. As you raise your holding period, max attainable returns and max attainable risk-adjusted returns decline very sharply. This is why good quantitative hedge funds insist on short-term forecasts and active strategies that trade frequently though not necessarily HFT.

There are a number of other issues here. Even obvious things that a junior analyst on a good desk would do in about 2 minutes haven't been done in order to make God's returns look worse. For example, (I haven't tested this myself) I suspect God would do MUCH better if God rebalanced 1/1260th of his portfolio every day based on the 5-year forecast that day. (1260 trading days in 5 years), and God would likely have lower transaction costs as well.

There are also many issues with how they are constructing these God portfolios. Portfolio construction and risk management frequently have a bigger impact on investor returns than the quality of the forecasts.

Has anyone ever quantified the "self-fulfilling" component of index fund returns? ie all the money automatically going into them and that everyone knows that?

So if I get it, the God foresight would enable you to own the planet, at a CAGR of 49% but the regular losses would scare the living crap out of you.

Aan interesting exercise in learning various financial libraries this seems - and I would love to see what the end of year amounts would be for 100 dollars invested in 1927.

I guess it’s just human nature - “yes God, you may have been right about every 5 year period for the past century, but this time it’s dofferent”

The Anti-VC statement

It is also interestingly about loss aversion - a truly rational person would always take the cagr of 49% and damn the losses

On these constraints, what if you bought a basket of stocks with the highest 5 year performance, conditional on never losing more than 5%? How much of your return would be given up by needing to avoid those "risks"?

If you're going to evaluate god on something, seems only fair to allow them to optimize for it.

If I were God, I'd buy on those dips that might get me fired. Then I'd really feel like -- God.

Why would a God like that need money anyway, or care about getting fired?

This is related to why I firmly believe any amateur outside of Wall Street can beat the crap out of this deplorable industry:

- by focusing on great value, the proverbial dollar for 50 cents. This is kind of hard work, you have to put in the time, which means you have to be genuinely interested in "business" or you won't keep it up.

- by not doing anything unless you find a dollar for 50 cents. Wall Street has to play all the time, you don't.

- by holding a sufficiently long time. Of course you can't wait indefinitely; sometimes just taking a loss e.g. if you find something else which is great value.

- by stomaching volatility

- by embracing volatility and, when you can, making sweet love to it

- by keeping it simple. Your only "hedge" can be just trying to pay 50 cents for a dollar. Having some crude heuristics for timing and taking profits.

- by assuming every management team and big shareholder is constantly trying to screw you over. Tolerating no bullshit. Trying to avoid being around morally handicapped people (this applies to any part of life imho).

- most relevant to the original article: by not tolerating anybody else's opinion (which implies investing your own money and only your own money, meaning you can. not. get. fired. Ever. And if you screw up, you'll learn.)

- by avoiding tip givers and tip takers. Group think is a killer. But of course, enjoying talking about general conditions. (My honest, very personal take on current group think: "you can't go wrong with low cost index investing". Which is not a bad idea in itself, until everyone starts doing it and they flood the market with indices and dubiously structured trackers.)

- once again, by not feeling you have to be part of everything which goes up.

- by having a lack of stress (you will doubt a lot, get screwed by management and big shareholders and have plenty of losers and lumpy payoffs)

- by doing this for a very long time (as in: the 80s were much easier than the 201xs, maybe just because fewer people were watching. And back then they were sometimes hiding profits instead of faking them. Today is the hardest time ever, if I find something superficially good, probably something is wrong with it. Personally I have very little self confidence today.)

- being small, which means you can look in places the big guys who have to move around billions can't look

- staying on the "easy side" of the basic math of loss vs profit. I personally would never short or sell optionality or stuff like that. Willing to take a gentle thrashing but never ruin.

And necessarily:

- by avoiding the show stoppers (start playing bridge, take a journey around the world, get divorced, disease, death)

Everything else is bullshit. Or at least part of more complex or shorter term or "trading" strategies with which amateurs can of course never do great. Volatility is not risk. Concentration is not necessarily risky.

Edit: poured out some random thoughts and tried to clean it up later. Sorry for the mess.

Why bridge?

Any good resources you recommend? How many hours per week do you think it takes? I was thinking about playing one of those play money games for a year or so.

Bridge: just a joke

Maybe the Burry list, in order:

1. The intelligent investor, Graham & Dodd

2. Common stocks and uncommon profits, Fisher

3. Why stocks go up and down, Pike

4. Buffettology, Buffett and Clark

but of course a lot of it is trends in accounting and legal issues. Which may be hard to learn on your own.

The biggest problem IMO is being able to stomach trading a very large fraction (or more likely, multiples) of your annual salary and not being affected by very large red numbers. If you stop thinking rationally, you lose. If you experience too much stress, might as well sell your time instead and avoid the drug habits.

In other words, you have to be able to put your house on red, and restrain yourself from doing it unless you're certain the numbers are in your favour.

I don't think many people can do this.

Why do you have to use amounts of money that make you uncomfortable? A good position should end up becoming big vs. the lesser positions mostly on its own merits.

Yes the stomach and patience are hard. Basically because you will never be sure.

Because if you're comfortable, it's probably because either you aren't trading enough for it to be worth the time (it's a hobby) or you have so much in the bank that trading performance is mostly about whether you can afford a Monet.

Avoiding stress, in the context of this discussion, for me involves not giving "tips", doing your own money, accepting and sometimes loving volatility, taking losses and accept undergoing periods of not looking like part of the winning team (which I think is what the original article is about), avoiding irrelevant constraints, and obviously avoiding ruin at all cost. But you can't get around doubt.

Having proper income for daily life and budgeting for setbacks is a different discussion. That's an issue you must get in order prior to investing. It's a prerequisite.

Speaking of which, how is that long-turn bet against China going?

He died for your alpha!

Reply on news.ycombinator.com