Even God Would Get Fired as an Active Investor

//, Key Research, Tactical Asset Allocation Research/Even God Would Get Fired as an Active Investor

Even God Would Get Fired as an Active Investor

Empirical asset pricing research can sometimes get monotonous because you end up circling back relentlessly to the same conclusions: value, momentum, trend-following are all interesting, and yet, markets are remarkably competitive (perhaps not efficient). But, sometimes, research uncovers absolutely stunning and counter-intuitive results–and this is where things get truly exciting. The study below is what we consider “exciting” research because the results are so profound (at least to us).

Our bottom line result is that perfect foresight has great returns, but gut-wrenching drawdowns. In other words, an active manager who was clairvoyant (i.e. “God”),(1) and knew ahead of time exactly which stocks were going to be long-term winners and long-term losers, would likely get fired many times over if they were managing other people’s money.

Question: If God is omnipotent, could he create a long-term active investment strategy fund that was so good that he could never get fired?

The answer is striking: God would get fired.

Let that settle in a bit.

The Design of Our “God” Study

Starting on 1/1/1927 we compute the 5-year “look ahead” return for all common stocks for the 500 largest NYSE/NASDAQ/AMEX firms. For simplicity, we eliminate any firms that do not have returns for a full 60 months.(2) We look at gross returns and all returns are total returns including dividends. Next, we create decile portfolios based on the forward five-year compound annual growth rate (CAGR).

We rebalance the names in the portfolio on January 1st of every fifth year. The first portfolio formation is January 1, 1927 and is held until December 31, 1931. The second portfolio is formed on January 1, 1932 and held until December 31, 1936. This pattern repeats every fifth year. To be clear, this is a non-investable portfolio that would require one to know with 100% certainty the performance of the top 500 stocks over the next 5 years.

We are explicitly engaging in look-ahead bias.

Returns are analyzed from 1/1/1927 to 12/31/2016. Portfolios are value-weighted returns for month t are weighted using the market capitalization at the end of month t-1. All returns are gross of transaction costs, taxes, and fees.

Performance of the Decile Portfolios

We first look at the decile portfolios rebalanced every 5 years. These portfolios highlight what perfect foresight can achieve. The Decile 10 portfolios represent value-weighted portfolios sorted on future top 5-year performers and the Decile 1 represent value-weighted portfolios sorted on future bottom 5-year performers. The compound annual growth rates for the 10 decile look-ahead portfolios are mapped below:

The results are hypothetical results and are NOT an indicator of future results and do NOT represent returns that any investor actually attained. In fact, these returns are EXPLICITLY IMPOSSIBLE TO ACHIEVE. Indexes are unmanaged, do not reflect management or trading fees, and one cannot invest directly in an index. Additional information regarding the construction of these results is available upon request. Note: these results were updated on 6/14/2017. 

As expected, a portfolio formed on the names that have the best 5 -year performance, have the best 5-year performance. Duh. God would compound at nearly 29% a year, in theory. In practice, he would run into capacity constraints and own the entire market (see here for details).

We know God would knock it out of the park, but the details are interesting…

Summary Statistics

Here we investigate some statistics and charts on the performance of the 5-year look ahead portfolio.

  • God_Best = Top decile 5-year winner portfolio
  • God_Worst = Bottom decile 5-year loser portfolio
  • SP500 = S&P 500 Total Return Index

First, the raw summary statistics:

Summary Statistics God_Best God_worst SP500
CAGR 29.37% -15.32% 9.87%
Standard Deviation 22.41% 29.13% 18.96%
Sharpe Ratio (RF=T-Bills) 1.12 -0.53 0.42
Worst Drawdown -75.94% -99.99 -84.59%

The results are hypothetical results and are NOT an indicator of future results and do NOT represent returns that any investor actually attained. In fact, these returns are EXPLICITLY IMPOSSIBLE TO ACHIEVE. Indexes are unmanaged, do not reflect management or trading fees, and one cannot invest directly in an index. Additional information regarding the construction of these results is available upon request. Note: these results were updated on 6/14/2017. 

The 29% CAGR is obviously awesome for the look-ahead portfolio. Expected.

The volatility is high on the God_Best portfolio — higher than the market. Interesting.

The Sharpe Ratio is above 1, but not by much. A far cry from the 2+ Sharpe Ratios touted by some hedge funds. Interesting.

But how about them drawdowns! The perfect foresight portfolio eats a devastating 76% drawdown (Aug 1929 to May 1932). But the pain doesn’t end there, here is a chart of the drawdowns on the portfolio over time:

The results are hypothetical results and are NOT an indicator of future results and do NOT represent returns that any investor actually attained. In fact, these returns are EXPLICITLY IMPOSSIBLE TO ACHIEVE. Indexes are unmanaged, do not reflect management or trading fees, and one cannot invest directly in an index. Additional information regarding the construction of these results is available upon request. Note: these results were updated on 6/14/2017.

And here are some details on the drawdowns:

Drawdown Rank Drawdown Date of Prior Peak Date of Low Date of Recovery Peak to Low (days) Low to Recovery (days) Peak to Peak (days)
1 -75.94% 8/30/1929 5/31/1932 6/30/1933 1005 395 1400
2 -40.75% 5/31/2008 2/28/2009 3/31/2010 273 396 669
3 -39.51% 8/31/2000 9/30/2001 9/30/2003 395 730 1125
4 -38.54% 2/27/1937 3/31/1938 12/31/1938 397 275 672
5 -30.81% 12/31/1973 9/30/1974 4/30/1975 273 212 485
6 -27.69% 8/31/1987 11/30/1987 1/31/1989 91 428 519
7 -26.94% 5/31/1946 11/30/1946 4/30/1948 183 517 700
8 -24.61% 11/30/1980 9/30/1981 8/31/1982 304 335 639
9 -21.53% 2/28/1962 6/30/1962 1/31/1963 122 215 337
10 -20.13% 3/31/1934 7/31/1934 4/30/1935 122 273 395

The results are hypothetical results and are NOT an indicator of future results and do NOT represent returns that any investor actually attained. In fact, these returns are EXPLICITLY IMPOSSIBLE TO ACHIEVE. Indexes are unmanaged, do not reflect management or trading fees, and one cannot invest directly in an index. Additional information regarding the construction of these results is available upon request. Note: these results were updated on 6/14/2017. 

Clearly, even a “perfect” long portfolio can bring a long-only investor a ton of pain.

How About We Create A Hedge Fund Managed by God?

In the analysis above we highlight that God’s long portfolio can endure enormous drawdowns and enhanced volatility. But perhaps we can leverage God’s perfect foresight and go long the known winners and short the known losers. Slam dunk, right?

Let’s investigate…

God’s long/short portfolio is constructed as follows:

  • Long God_Best and short God_Worst, rebalanced monthly.

The following portfolios are examined:

  • God L/S = Long 5-year decile winners; short 5-year decile losers
  • SP500 = S&P 500 Total Return index

Summary Statistics

Here are the high-level stats:

Summary Statistics* God L/S SP500
CAGR 46.23% 9.87%
Standard Deviation 20.08% 18.96%
Sharpe Ratio 1.86 0.42
Worst Drawdown -47.28% -84.59

The results are hypothetical results and are NOT an indicator of future results and do NOT represent returns that any investor actually attained. In fact, these returns are EXPLICITLY IMPOSSIBLE TO ACHIEVE. Indexes are unmanaged, do not reflect management or trading fees, and one cannot invest directly in an index. Additional information regarding the construction of these results is available upon request. Note: these results were updated on 6/14/2017. 

Yowza!

Clearly, the ultimate hedge fund does amazingly well — 46% CAGRs would have you owning the world’s stock market in short order. Obviously, this sort of return is not possible over a long period — even if someone had perfect “Biff-like” foresight.

Yet, check out the worst drawdown on the PERFECT hedge fund — 47%+. Incredible. And it gets better…

Here is the time series of drawdowns over time for the God L/S portfolio. Certainly not a cake walk!

The results are hypothetical results and are NOT an indicator of future results and do NOT represent returns that any investor actually attained. In fact, these returns are EXPLICITLY IMPOSSIBLE TO ACHIEVE. Indexes are unmanaged, do not reflect management or trading fees, and one cannot invest directly in an index. Additional information regarding the construction of these results is available upon request. Note: these results were updated on 6/14/2017. 

Let that chart sink in a bit. Multiple opportunities to lose 20%+ over time. Clearly not riskless. But it gets even better…

As many investment pros painfully recognize, managing money is often not about absolute performance, but relative short-term performance. Another truism is that the S&P 500 ends up being everyone’s benchmark, regardless of the strategy — especially during a long-term bull market!

Let’s look at the 1-year relative CAGR over time between God L/S and the S&P 500.

The results are hypothetical results and are NOT an indicator of future results and do NOT represent returns that any investor actually attained. In fact, these returns are EXPLICITLY IMPOSSIBLE TO ACHIEVE. Indexes are unmanaged, do not reflect management or trading fees, and one cannot invest directly in an index. Additional information regarding the construction of these results is available upon request. Note: these results were updated on 6/14/2017. 

What the chart highlights is that even GOD HIMSELF would get fired multiple times over. The relative performance of God’s hedge fund is often abysmal and he’d surely make the cover of Barron’s or the WSJ on multiple occasions throughout his career. The passive index would eat his lunch on multiple occasions — often getting beaten by 50 percentage points — or more — on multiple occasions!

These results highlight the fickle nature of assessing relative performance over short horizons. We’ve shown this quantitatively, but Ben Carlson talks about the challenge of short-horizon thinking here, and Meb Faber recently highlighted that investors are terrible at timing active investments.

Conclusions

The famous quote attributed (wrongly?) to Keynes is spot-on:

Markets can remain irrational longer than you can remain solvent!

This study also highlights a truism for all active investors:

Active investors MUST have a long-horizon!

Good luck out there…(3)


  • The views and opinions expressed herein are those of the author and do not necessarily reflect the views of Alpha Architect, its affiliates or its employees. Our full disclosures are available here. Definitions of common statistics used in our analysis are available here (towards the bottom).
  • Join thousands of other readers and subscribe to our blog.
  • This site provides NO information on our value ETFs or our momentum ETFs. Please refer to this site.

Print Friendly, PDF & Email

References   [ + ]

1. We mean no offense by the use of the term and this could be construed as a single deity, multiple deities, or whatever fulfills your definition of an entity or concept that is all powerful and all knowing.
2. Results are similar with or without this assumption.
3. h.t., Arturo B. . An old Chicago PhD (1980) we met at the Nantucket Project, who suggested we explore this research question…

About the Author:

After serving as a Captain in the United States Marine Corps, Dr. Gray earned a PhD, and worked as a finance professor at Drexel University. Dr. Gray’s interest in bridging the research gap between academia and industry led him to found Alpha Architect, an asset management that delivers affordable active exposures for tax-sensitive investors. Dr. Gray has published four books and a number of academic articles. Wes is a regular contributor to multiple industry outlets, to include the following: Wall Street Journal, Forbes, ETF.com, and the CFA Institute. Dr. Gray earned an MBA and a PhD in finance from the University of Chicago and graduated magna cum laude with a BS from The Wharton School of the University of Pennsylvania.

30 Comments

  1. ptuomov February 2, 2016 at 2:17 pm

    If God would use a sensible risk model, an optimizer, and some common sense portfolio constraints, God would not get fired in this experiment.

    Keeping the portfolio small/mid/large cap balanced and industry balanced (relative to whatever benchmark chosen) would probably bring in the drawdowns a lot.

    If the point is that poor risk control can squander away even great insights, I agree. The same goes for trading costs, by the way.

    • Wesley R. Gray, PhD February 2, 2016 at 3:10 pm

      That is a hypothesis that can be tested. Please explain the exact rules you believe would prevent God from getting fired and we can empirically test this proposition.

      In general, owning the stocks that will have the best 5 year returns and shorting the stocks that will have the worst 5 year returns is about the best strategy I can ever imagine.

      • ptuomov February 2, 2016 at 3:48 pm

        In a long-short context, here’s one thing that you could try to see whether risk control is important or not to the outcome of this experiment.

        After you have sorted stocks into ten perfect foresight portfolios and assigned each of them a decile assignment (as you did above), form ten portfolios from your overall top decile stocks by creating sector portfolios (say the ten GICS 2 sectors) out of the top ranked stocks. Then form your final long portfolio as the equal weighted portfolio of these ten sector-specific top deciles (averaging ten portfolios, one per sector). Similarly, for the short side.

        Another way to say this is that I’d try reweighting the stocks in your top decile not to be equal weighted but such that each sector has the same amount of weight (and equal weighted within sector.)

        My guess is that your average return will go down only a little but your standard deviation and drawdown will both go down a lot. But it’s just a guess.
        This is not the only way to use the perfect foresight information to get less volatile results, but it’s the simplest that I can come up with.

        [A correction: the stocks in the original portfolios are cap weighted, not equal weighted as I wrote above. If cap weighting is considered important, one could modify my proposal to allocate equal amount of market cap in each sector and then cap weight within sectors.]

  2. MrLovingKindness February 2, 2016 at 4:47 pm

    An omniscient being would short and long the best stocks every day (maybe every hour, minute or second, depending on the trading drag), and thus, avoid the draw downs completely (assuming the being knows the highest and lowest price points each day).

    • Wesley R. Gray, PhD February 2, 2016 at 6:55 pm

      Now that would be great, but he’s need to convince the broker to give him free trading costs — which I’m sure he could do with a little negotiating. ha!

      • Novice February 19, 2016 at 7:43 am

        Regardless of the level of transaction cost, it does not make sense to make “value-weighted portfolios sorted on future top 5-year performers” as the best performing stock of the future would outperform this portfolio in every period (5-year periods with this rebalancing strategy).

      • Thomas June 15, 2017 at 7:32 am

        Just following up on this … have you looked at what (God_best) max draw down looks like with annual rebalance?

        • Wesley Gray, PhD
          Wesley Gray, PhD June 15, 2017 at 8:16 am

          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

          • Thomas June 15, 2017 at 10:08 am

            Thanks, certainly an interesting way to frame the discussion!

  3. Steve February 2, 2016 at 10:03 pm

    I’m with you – this was an exciting one…thanks for sharing! I would not have predicted that result. Now matter how good your system, you’ve got to be doing this long-term.

    Also interesting to see that the strategies we talk about sit around the 7-9 decile mark.

  4. OVVO Financial February 2, 2016 at 11:23 pm

    Why would God disregard utility?

    “When early mathematicians first formulated principles of behavior in chance situations, they assumed the proper objective of the individual was to maximize expected money return. It was later found that this objective can be incredibly bad.

    The expected utility rule was proposed as a substitute for the expected return rule.”

    Pg. 207 here: http://cowles.yale.edu/sites/default/files/files/pub/mon/m16-all.pdf

  5. Ashley February 3, 2016 at 6:05 pm

    My other takeaway from this is — If you’re reaching for the biggest returns, you’re going to be taking the biggest risks. I wonder how “God” would do if he factored in some measure of relative volatility while building the portfolio?

  6. JS February 4, 2016 at 10:16 am

    The fourth worst long-only portfolio (03/31/2000 to 03/31/2001) had a portfolio return of -34.03%. This is the exact same portfolio return for the 8th worst long/short portfolio (which happens to be the exact same time period). While it is certainly theoretically possible, it seems highly unlikely to the second decimal. Is it possible there is an error? It is also possible I missed something in the methodology.

    • Wesley R. Gray, PhD February 4, 2016 at 11:00 am

      Hey JS,

      Nice catch. Updated the L/S table. We have to handjam these things out of our systems to get them in to the blog.

      BTW, if you have the capabilities to replicate the study, let us know what you find. Happy to share details on methodology, etc. We know of no other investigations into this topic so we are “pioneering” the research effort.

      Would be great to see what others find.

  7. Patrick Luby February 4, 2016 at 12:08 pm

    Thank you…fascinating article. Have you done a similar analysis on asset allocation?

    • Wesley R. Gray, PhD February 4, 2016 at 2:58 pm

      Hi Patrick,

      We haven’t done a study like this in that context, but in general, strategies that work, tend to have major pain points at some point…that is a recurring theme in our research.

      When we get some R&D bandwidth we’ll explore some of the questions folks have posted on the blog…

    • JS February 5, 2016 at 4:57 pm

      Not exact, but Figure 1 in this firm’s blog does a good job of contextualizing the extreme nature of the “failure” of a diversified portfolio in 2015 (which is, I expect, the most imminent reason for asking). There are several exhibits that could be helpful.

      http://gestaltu.com/2016/02/navigating-active-asset-allocation-when-diversification-fails.html/

      I am not affiliated with the firm or article, just thought it might help.

  8. Prateek Sharma February 11, 2016 at 5:38 am

    The value-weighted portfolio holds a very special significance, if the two fund separation theorem holds, that is all risky investments are based on a unique risky portfolio, than value-weighted portfolio is the only portfolio that will clear the market. Under some (highly questionable) assumptions, the value-weighted portfolio is the optimum portfolio, one with maximum expected Sharpe ratio. However, as soon as you make a person clairvoyant, which means that he/she can predict, with certainty, which stocks will get what return in the future, the notion of risk is eliminated. In this certain world, value-weighting would not make any sense, and neither would any expectations about the Sharpe ratio. In fact, the optimum strategy would be to invest the entire capital in a single asset that “you know” would yield the highest returns for your investment horizon (5 years, 10 years, 20 years , whatever it may be). If shorting is allowed, the optimum strategy would be to short the single asset that is going to perform the worst, and go long the single asset that is going to perform the best.

  9. Kevin Osborne February 12, 2016 at 3:30 pm

    Investors are terrible at timing active investments because they don’t have enough information or don’t know how to value that information. Yet small investors have a huge advantage over the big funds because of the nimble trading available to them as opposed to moving a large position that alone changes the value of a stock. If an investor learns how the big boys choose their stocks it is possible to piggy back on their moves, in and out, and make very good returns, certainly better than the average market, even if generally long. (although shorting when the clouds appear will help considerably)

    • Yo February 15, 2016 at 2:11 pm

      that’s called high frequency trading

      • Kevin Osborne February 15, 2016 at 2:16 pm

        Thank you.

  10. Mark Dodson February 16, 2016 at 4:44 pm

    Wes – this is great stuff. Consulting services departments of wirehouses often use trailing 3yr (and 5yr) comparisons with the S&P 500 as one of their primary method for determining what managers get hired and fired from their platforms? Knowing your lookahead period was 5yrs, leaving this option out, I’m still curious how often this perfect 5yr foresight portfolio found itself lagging the S&P 500 after a random three year period.Would it be possible to see calendar year returns for these strategies as well?

    • Wesley R. Gray, PhD February 17, 2016 at 10:10 am

      had to post in 2 sections because it is too long.
      Also posting 1yr rolling cagr. most of the time it wins, but even God can lose.

      • Wesley R. Gray, PhD February 17, 2016 at 10:15 am

        the rolling cagr chart is kinda fascinating. In normal markets the thing kinda grinds, but when the S&P blows up you get massive spikes…so one hypothesis is you need chaos for active mgmt to really shine and as buffett says, “figure out who is swimming naked.” Who knows. Main point is market volatility and market psychology is incredibly complex.

        • Mark Dodson February 18, 2016 at 4:01 pm

          Wes – thanks for taking time to share all this data and the chart. I appreciate it!

  11. Hannibal Smith March 20, 2016 at 1:37 am

    Hate to be a party pooper, but there’s nothing surprising at all about these results. Being right is not the same thing as making money as Ned Davis rightly wrote an entire book about.

    The proper way to have constructed the deciles if you had perfect hindsight was on a risk/reward basis. Then “God” wouldn’t have been fired since he would have beat his tracking index. A better question is why the hell would “God” be a “professional” manager? No, he’d be an independent investor.

  12. MAR April 14, 2016 at 11:14 am

    I think what is interesting from this data is that someone like Buffett has close to a 20% CAGR return over a long period of time, which based on your data seems impossible. Assuming here that 30% top decile returns come down a few % due to real transaction costs etc. For someone like Julien Robertson who had a reported 32% CARG return over 20 years your data would imply statistically it was almost impossible to achieve. They’re not God but they’re pretty close.

    • Hannibal Smith April 14, 2016 at 1:21 pm

      Buffett ran a hedge fund for the first 20 years or so, so those returns are baked into the cake for later CAGR. Very misleading. And Robertson was a product of the pre-1990 era where data was not widely electronic and available. The trick is whether or not these two can still do their anarchronisms in this day and age… they’ll have to adapt and so far Buffett seems to be failing at it. I don’t think Robertson is a player anymore; just his cubs.

      • praxeologue August 8, 2016 at 9:49 am

        I think a point rarely covered about Buffett’s 20% compound return is the ‘free’ leverage from investing the insurance float. If you look at his unlevered returns as an investor they work out to about 12% a year for 40 years vs 9% for the S+P (see here http://taussigcapital.com/A_Tale_of_Two_Capital_Structures.pdf ) … this is not to downplay his investment acumen as his decision to have an insurer and invest the float was all deliberate on his part so he gets 100% credit. But, we shouldnt compare an unlevered fund managers record to Buffett/BRK, it isnt fair comparison.

Comments are closed.