Concentration is the most undermined risk working against long-term wealth generation. The reason you have not heard about it is that both active and passive investment management indulge in concentrated portfolios and don’t think there is anything wrong with the approach. In addition, human active asset managers’ skill of stock selection is rare, and hard to measure and passive indexes are concentrated because of poor design [1]. This is why it is convenient to ignore concentration and allow investors to be misled by performance.
On 27th April 2021, the portfolio advisor carried the article, “James Anderson and Nick Train leave Warren Buffett in the dust over the last 20 years” [2]. James Anderson has retired, and Nick is still at the top of the table with 1600% returns over the last 20 years. The article focussed on top performers and showed Warren Buffet at the bottom of the table with a 500% return for the period, compared to 300% returns of the S&P 500 for the respective period.
If we separate the managers' skills from the market returns, the world’s best averaged around 10% annual excess returns over the 20 years. And year over year, those margins compound and reach a multiplier greater than 5 times the total market return. [FTSE 350 or S&P 500]
However, there is more than what meets the eye. Nick Train manages Finsbury Growth and Investment Trust [FGT], which carries only 22 names with the top 10 representing more than 84% of the weight. [3] FGT outperformed 16 out of 20 years, however in 30 months [Starting September 2020 till May 2022], FGT gave up five years of outperformance.
Fundamentalists and the investing public at large may find it sacrilegious to bundle Warren Buffet’s performance with other Active managers, but in the end, Warren is not outside the scope of performance measurement. Berkshire Hathaway [BRK] has delivered a similar annualized excess return of 10% above the S&P500 since 1965. However, the last 20 years have seen BRK matching SPY. Anyone who has invested in BRK from Jan 1999 till July 2020 could have done equally well by buying the SPY. Warren is not a stranger to the competition with the SPY and famously won the wager [4] regarding Hedge fund underperformance against the SPY. But what may be puzzling for the common investor is how Warren can underperform the SPY consistently.
Both passive and active, investment management is a concentration contest. SPY’s size-weighted method concentrates 30% of the weight on 2% of the stocks [10 out of 500] [5]. Active managers can charge a fee only if they are differentiated from the markets. This forces them to become stock selectors. The effort and resources spent in identifying ideas, naturally lead to concentration. The overweighting, game of intuition [amplification] works till it eventually fails, either as a crash or as a drag on performance.
Concentration by very nature is clustered and hence cannot deliver a homogeneous, consistent outperformance. Like an artist, there are spurts of genius, which depend on many variables. This can disillusion the investor seeking consistency if he does not reconcile with the volatility of the art form. Asset owners with a longer horizon could assimilate the risk, but for small investors concentrated activity could be a risk mismatch. Investors should also understand that if there was no idiosyncratic selection, there would be no showcase of expertise and hence the human manager would seem like a quant machine, without a cult, without celebration, without religion, just with a process.
Though the concentration characteristics are similar between SPY and the Warren-Nick-Active method, the two concentrations follow contrastingly different approaches. The SPY method is concentrated because of a convenient construction design, while the Active method is concentrated because of the human need for idiosyncraticity, to explain the cause for an effect and feel good when the prognosis is correct. There is nothing wrong with connecting cause with effect. The problem is regarding the divergences when the cause does not create the desired effect, situations when the effect drives the cause [6], or simply a function of the nature of information [7]. Subjective analysis can be a tiring job and the more the assets grow, the manager could hit capacity constraints. After all, there is a limit to performance. How many billion dollars can you allocate to a 22-stock portfolio?
Investing machines don’t suffer from such legacy thinking as they understand that concentration increases risk which increases the chance of performance surprises. Machines are also capable to understand that there is an inflection point when concentration can start hurting the Information Ratio. Above all this comes asset allocation. You can’t expect a genius stock selector to first do a tremendous job of finding a needle in the haystack and then also conjure up a dynamic asset allocation model that tactically understands when to increase or decrease an allocation's overall weight. Great stock selectors work against monumental odds and with smart machines they can increase their chances of success. Otherwise, they are preparing themselves for an underperformance shock or systematic destruction of a brilliantly cultivated legacy of half a century.
The chart below is a Cartesian summary of 15000 portfolios plotting Annualized Excess Volatility and Annualized Excess Returns run on a set of 5 statistical factors, for 3 different holding periods, on 9 scoring systems from 2016 till 2021 on S&P 100 components. The chart breaks up the data into four quadrants. Q2 is the most desirable quadrant depicting high annualized excess returns with negative annualized excess volatility with a small chance of 17.48%. Q1 is the high annualized excess return – high annualized excess volatility quadrant with a large 48.76% chance. Q1 quadrant pulls most asset managers into the profession [Higher Risk – Higher Return]. Q3 is the underperformance quadrant with 15.38% probability. Here the asset manager is underperforming but with an overall low risk. The Q4 is the uncertainty quadrant which kills an asset manager's performance as there is negative annualized excess return but high annualized excess volatility. Add concentration and overweighed allocation to the components and you can visualize "blow up" risk. A similar Cartesian chart works for most other benchmarks and groups of components. There is no amount of due diligence that can eliminate Q4 and Q3 quadrants risk. When an asset manager relies on human skill and naïve concentrations to overcome 33.65% chance of underperformance, he[she] is playing a risky game with his[her] reputation and investors’ money.
Machines are relentless. They do not age. They do not tire. They can stay objective [if you train them to detect bias]. They are designed to learn from failure. They don’t have an ego. Unlike machines, human reputations are easier to build and come with a glass jaw [Easily broken]. Machines on the other have to toil to become mainstream and become indispensable. Google page rank algorithm did not become mainstream in a day, but it changed everything.
Till human beings will live, they will juice the proverbial informational orange [8]. Idiosyncraticity will never die, which means markets would continue to oscillate between inefficiency and efficiency. Investors will continue to run after high-performing funds, without worrying about concentration risk, which will continue to be an off-the-shelf product that would never go out of fashion. And this is why even if the machine would open up its framework and explain how it does, what it does, it would never take subjective active managers completely out of business. High risks will always come with enough high returns for enough periods, to trap a segment of naïve investors all the time.
Sustainable alpha does not need concentration. Machines will invariably solve Granger’s challenge [9] i.e. an open method that delivers alpha that does not get arbitraged away eventually will bury the efficient market hypothesis [EMH]. Machines will transform finance into Science. The early adopters have already ushered in this new generation of investing because they understand that it is not Physics that is dangerous to your wealth [10], but concentration.
Bibliography
[1] Pal. M, The S&P500 Myth, SSRN, August 2022
[2] Habergham. H, James Anderson and Nick Train leave Warren Buffett in the dust over the last 20 years, Portfolio Advisor, April 2021
[3] Finsbury Growth & Income Trust PLS, Factsheet, June 2022
[4] Collie. B, Warren Buffet and the million-dollar wager, Russell Investments, July 2017
[5] Weights of S&P 500 components
[6] Pal. M, The Observer Effect, SSRN, February 2022
[7] Pal. M, Reversion Diversion Hypothesis, SSRN, July 2015
[8] Pal. M, Information is not an Orange, AlphaBlock, April 2022
[9] Granger. C, Forecasting Stock market Prices, International Journal of Forecasting, 1992
[10] Pal. M, How Physics Solved Your Wealth Problem, SSRN, October 2016