"The poor overall performance of mutual fund managers in the past is not due to a lack of stock-picking ability, but rather to institutional factors that encourage them to overdiversify, i.e. pick more stocks than their best alpha-generating ideas. We point out that these factors may include not only the desire to have a very large fund and therefore collect more fees [as detailed in Berk and Green (2004)] but also the desire by both managers and investors to minimize idiosyncratic volatility: Though of course managers are risk averse, investors appear to judge funds irrationally by measures such as Sharpe Ratio or Morningstar rating. Both of these measures penalize idiosyncratic volatility, which is not truly appropriate in a portfolio context."
--- "Best Ideas", a working paper by Harvard professors Randy Cohen, Christopher Polk, and Berhard Silli released in October 2008.
The statement above should be no surprise to value investors. Investors such as Warren Buffett, Bill Ackman, Joel Greenblatt, Mohnish Pabrai, and Eddie Lampert have generally espoused a concentrated portfolio with spectacular results. Given that many of these investors represent those we wish to emulate, the question value investors often ponder is "how concentrated should I be?"
There are several considerations an investor should reflect upon when determining portfolio concentration. One basic consideration is how involved or how often does one look at their portfolio. This may sound like a worthless aspect to consider until one factors in our propensity for loss aversion. People generally feel the pain of a loss twice as intensely as the enjoyment of a gain. All else equal, a 1% loss yields the same feeling as a 2% gain.
If one has a highly concentrated portfolio, overall volatility will generally be higher. Combine that higher volatility with reviewing a portfolio multiple times during the day and even battle-tested value investors can set themselves up to "over-trade" the portfolio and reduce the potential gains. Ahh, but we are value investors and we clearly can realize that market risk/volatility has nothing to do with business risk and risk of permanent capital loss. While it's easy to say that, the reality is it's difficult to stop being human.
In fact, on Tuesday, February 8, 2011, I noticed one of my holdings, Full House Resorts ("FLL"), sustain a loss of roughly 14% on about ten times its daily volume on no discernable news. Is that just noise or is there some material information that can have a significant impact on FLL's future valuation (i.e. impair its long-term business prospects)? As a fundamental value investor, I conduct a lot of stress tests in the form of sensitivity analyses across the income statement. In the case of FLL, certain casino operations have far higher margins relative to others so if something happens to one of these operations, the intrinsic value of FLL can materially change. And when you see a position lose 14% on no news and heavy volume, it's natural to review everything you have and check for something you may have missed in your analysis.
We can all comprehend the behavioral components that can result in self-destructive actions, we can all understand idiosyncratic risk as opposed to market risk, but when one sees red on the screen for sustained periods of time, it can be difficult overtime, particularly with a highly concentrated (~10 positions) portfolio, to stay the course. There are of course people that can handle that volatility but this requires some serious introspection on the part of the investor.
Another aspect to consider with regards to position sizing is the type of value approach one is utilizing. Some investors tend to focus on large cap stocks with low valuation metrics with strong franchises like Microsoft (MSFT), Pfizer (PFE), and Exxon Mobil (XOM). With companies like these, one can probably have a more concentrated basket of roughly 12-15 stocks and perform well.
However, if you are a deep-value microcap/small cap investor, a highly concentrated portfolio can provide tremendous upside potential but also considerable risk of permanent capital loss. As value investors, we like to believe that we conduct rigorous analysis. Running up and down a balance sheet and scrubbing it for liquidation value can provide a tangible backstop value and/or margin of safety. The reality is that despite all of this analysis, even the best investors can be very wrong.
Businesses are dynamic and the changes they undergo influence their intrinsic value. With microcap/small cap companies (<$1B), the operational risk can be pretty high due to customer concentration, manager risk, competition, etc. that is just not comparable to large cap companies. So even if the balance sheet is pristine, the outlook for the business is phenomenal, and it trades for extremely attractive price to cash flow and price to tangible book value metrics, an unseen occurrence can significantly alter the value of the business and cause considerable and possibly permanent loss of capital.
This problem can be compounded when one gets into the deep-value segment as many of these smaller businesses trade for attractive multiples because the businesses themselves are highly distressed. It's for this reason that I generally advocate that deep-value investors spread their investments over 25-35 holdings. In fact, having initially managed a portfolio that was heavily weighted to 5 main holdings (50+% of a portfolio) within a 15-20 stock portfolio, after backtesting my portfolio whereby I increased weightings for the more marginal positions and reduced the weightings of my highest conviction holdings, I found that overall performance would have been better.
One quick response could be that I'm just not a good, high-conviction stock picker. That may very well be the case but what I think is that once we distill the universe into a basket of ideas that meet our rigorous criteria as value investors, drilling even further down and increasing position sizes and shrinking the portfolio provides little marginal benefit under ~25 holdings.
I have found that considering position sizes in terms of potential return/loss to the overall portfolio has helped me significantly. When I establish a 3% position, if it increases by 50%, it will add 1.5% to my portfolio. A 5% position will add 2.5% and a 10% position would add 5%. However, if I'm wrong on a 7-10% position, my portfolio will be down as much as 5% from a 50% drawdown on a 10% portfolio weighting. That requires a lot of "heavy lifting" from other parts of the portfolio to get out of that hole. Presumably there would be other 10% winners to help offset it but consider a nice winner, a 75% return on a 10% position, alongside a 30% drawdown on another 10% position. You're netting out about 4.5% on 20% of the total portfolio which may not be the best risk/return profile on that level of capital.
I can also be wrong as to the magnitude of what I think a 10% size weighting is worth. I may initially think it's a 50 cent dollar and thus worthy of a high conviction (10%) position size when in reality/based on a changes in the business itself, it's an 83 cent dollar. As a result, that 10% position goes up just 20% and adds just 2% to my portfolio. From a portfolio standpoint, that could represent significant opportunity cost one has forfeited to eke out just 2% on 10% of one's portfolio.
This also ties to the variance in regards to the magnitude of stock performance in the short-term (<3 years). Your highest conviction pick may underperform from a stock performance aspect relative to more marginal inclusions in the portfolio. I am sure many investors have seen the business fundamentals of an investment improve yet the stock price refuse to reflect these improvements, leaving the stock even cheaper from a valuation perspective whereby they're left scratching their heads for quite some time.
My investment in Cash Store Financial ("CSF.TO"), formerly known as Rentcash, fits within this category. I owned this stock while it gyrated from $4-5 and felt it was one of the cheapest companies I owned. Great insider ownership and management that was repurchasing shares, very impressive cash flow production and future cash flow production when even factoring in proposed legislation which would negatively impact the business, and just cheap across all valuation metrics I use. This was roughly 10% of my fund but despite improved news and results in 2008 (I bought in H2 2007), it wasn't until mid 2009 that Mr. Market decided it was worth a bit more.
Between the dividends and performance of CSF, the investment worked out well for my investors and me. However, in hindsight committing 10% of the portfolio to this idea was a bit of an opportunity cost. It was a 50 cent dollar when I bought it, and Mr. Market decided it was even a 30 cent dollar by the start of 2010 (I had sold in the $10-12 range) but being right about one specific idea does not mean that I had an optimal portfolio in terms of weighting.
Investors also want to be able to stay in the game and having a highly concentrated portfolio, irrespective of the quality of analysis behind the holdings, can result in big losses which reduces the capital available to be deployed in other promising prospects down the road. A portfolio of roughly 25 holdings with 3-5% position sizes allows a person to take on enough company-specific risk and generate market beating gains without having potential knock-out blows from an errant large position. I have found it's also easier to make better judgments regarding positions with a slightly broader portfolio.
For example, investors are often tied pretty tightly to high conviction picks. I have found it's very difficult to be wholly objective with high conviction ideas because of all of the work invested in the analysis of the company. As a result, it takes a long time to re-assess new information and it can be difficult to determine when to cut the reins. This is clearly not the case with every investor, but I assume most investors have experienced what I have just described.
In 2010, Sprint-Nextel ("S") was one of my highest conviction holdings. It took some time for the various catalysts to emerge and work through the company but overall, S did not do much for my fund and managed accounts. I sold out the entire stake after some level of consternation after Q3 earnings and the sizeable allocation of capital to S cost me performance in 2010. With smaller position sizes, I have personally found that there's less of the behavioral element at play that can hinder portfolio performance.
Position sizing should be clearly tied to one's own behavioral aspects and personality. Some people are very comfortable with a portfolio of just 5-10 holdings (assuming it represents most of their liquid net worth) but the biggest consideration needs to be what and investor believes is an adequate return for the level of concentration.
Consider a five stock portfolio that's equally weighted. Assuming 80% of the portfolio generates a 75% return and one pick was a 90% loss (assume it was an investment in Citigroup in 2008 or a net-net microcap that went bust), the total return is about 42% on that portfolio. As an investor, the question should be can I replicate that performance with a broader portfolio while potentially limiting the downside and risk of capital impairment if one or two holdings goes really bad? That answer and the historical ability of an investor to avoid big mishaps is likely to influence his/her concentration preferences.
--- "Best Ideas", a working paper by Harvard professors Randy Cohen, Christopher Polk, and Berhard Silli released in October 2008.
The statement above should be no surprise to value investors. Investors such as Warren Buffett, Bill Ackman, Joel Greenblatt, Mohnish Pabrai, and Eddie Lampert have generally espoused a concentrated portfolio with spectacular results. Given that many of these investors represent those we wish to emulate, the question value investors often ponder is "how concentrated should I be?"
There are several considerations an investor should reflect upon when determining portfolio concentration. One basic consideration is how involved or how often does one look at their portfolio. This may sound like a worthless aspect to consider until one factors in our propensity for loss aversion. People generally feel the pain of a loss twice as intensely as the enjoyment of a gain. All else equal, a 1% loss yields the same feeling as a 2% gain.
If one has a highly concentrated portfolio, overall volatility will generally be higher. Combine that higher volatility with reviewing a portfolio multiple times during the day and even battle-tested value investors can set themselves up to "over-trade" the portfolio and reduce the potential gains. Ahh, but we are value investors and we clearly can realize that market risk/volatility has nothing to do with business risk and risk of permanent capital loss. While it's easy to say that, the reality is it's difficult to stop being human.
In fact, on Tuesday, February 8, 2011, I noticed one of my holdings, Full House Resorts ("FLL"), sustain a loss of roughly 14% on about ten times its daily volume on no discernable news. Is that just noise or is there some material information that can have a significant impact on FLL's future valuation (i.e. impair its long-term business prospects)? As a fundamental value investor, I conduct a lot of stress tests in the form of sensitivity analyses across the income statement. In the case of FLL, certain casino operations have far higher margins relative to others so if something happens to one of these operations, the intrinsic value of FLL can materially change. And when you see a position lose 14% on no news and heavy volume, it's natural to review everything you have and check for something you may have missed in your analysis.
We can all comprehend the behavioral components that can result in self-destructive actions, we can all understand idiosyncratic risk as opposed to market risk, but when one sees red on the screen for sustained periods of time, it can be difficult overtime, particularly with a highly concentrated (~10 positions) portfolio, to stay the course. There are of course people that can handle that volatility but this requires some serious introspection on the part of the investor.
Another aspect to consider with regards to position sizing is the type of value approach one is utilizing. Some investors tend to focus on large cap stocks with low valuation metrics with strong franchises like Microsoft (MSFT), Pfizer (PFE), and Exxon Mobil (XOM). With companies like these, one can probably have a more concentrated basket of roughly 12-15 stocks and perform well.
However, if you are a deep-value microcap/small cap investor, a highly concentrated portfolio can provide tremendous upside potential but also considerable risk of permanent capital loss. As value investors, we like to believe that we conduct rigorous analysis. Running up and down a balance sheet and scrubbing it for liquidation value can provide a tangible backstop value and/or margin of safety. The reality is that despite all of this analysis, even the best investors can be very wrong.
Businesses are dynamic and the changes they undergo influence their intrinsic value. With microcap/small cap companies (<$1B), the operational risk can be pretty high due to customer concentration, manager risk, competition, etc. that is just not comparable to large cap companies. So even if the balance sheet is pristine, the outlook for the business is phenomenal, and it trades for extremely attractive price to cash flow and price to tangible book value metrics, an unseen occurrence can significantly alter the value of the business and cause considerable and possibly permanent loss of capital.
This problem can be compounded when one gets into the deep-value segment as many of these smaller businesses trade for attractive multiples because the businesses themselves are highly distressed. It's for this reason that I generally advocate that deep-value investors spread their investments over 25-35 holdings. In fact, having initially managed a portfolio that was heavily weighted to 5 main holdings (50+% of a portfolio) within a 15-20 stock portfolio, after backtesting my portfolio whereby I increased weightings for the more marginal positions and reduced the weightings of my highest conviction holdings, I found that overall performance would have been better.
One quick response could be that I'm just not a good, high-conviction stock picker. That may very well be the case but what I think is that once we distill the universe into a basket of ideas that meet our rigorous criteria as value investors, drilling even further down and increasing position sizes and shrinking the portfolio provides little marginal benefit under ~25 holdings.
I have found that considering position sizes in terms of potential return/loss to the overall portfolio has helped me significantly. When I establish a 3% position, if it increases by 50%, it will add 1.5% to my portfolio. A 5% position will add 2.5% and a 10% position would add 5%. However, if I'm wrong on a 7-10% position, my portfolio will be down as much as 5% from a 50% drawdown on a 10% portfolio weighting. That requires a lot of "heavy lifting" from other parts of the portfolio to get out of that hole. Presumably there would be other 10% winners to help offset it but consider a nice winner, a 75% return on a 10% position, alongside a 30% drawdown on another 10% position. You're netting out about 4.5% on 20% of the total portfolio which may not be the best risk/return profile on that level of capital.
I can also be wrong as to the magnitude of what I think a 10% size weighting is worth. I may initially think it's a 50 cent dollar and thus worthy of a high conviction (10%) position size when in reality/based on a changes in the business itself, it's an 83 cent dollar. As a result, that 10% position goes up just 20% and adds just 2% to my portfolio. From a portfolio standpoint, that could represent significant opportunity cost one has forfeited to eke out just 2% on 10% of one's portfolio.
This also ties to the variance in regards to the magnitude of stock performance in the short-term (<3 years). Your highest conviction pick may underperform from a stock performance aspect relative to more marginal inclusions in the portfolio. I am sure many investors have seen the business fundamentals of an investment improve yet the stock price refuse to reflect these improvements, leaving the stock even cheaper from a valuation perspective whereby they're left scratching their heads for quite some time.
My investment in Cash Store Financial ("CSF.TO"), formerly known as Rentcash, fits within this category. I owned this stock while it gyrated from $4-5 and felt it was one of the cheapest companies I owned. Great insider ownership and management that was repurchasing shares, very impressive cash flow production and future cash flow production when even factoring in proposed legislation which would negatively impact the business, and just cheap across all valuation metrics I use. This was roughly 10% of my fund but despite improved news and results in 2008 (I bought in H2 2007), it wasn't until mid 2009 that Mr. Market decided it was worth a bit more.
Between the dividends and performance of CSF, the investment worked out well for my investors and me. However, in hindsight committing 10% of the portfolio to this idea was a bit of an opportunity cost. It was a 50 cent dollar when I bought it, and Mr. Market decided it was even a 30 cent dollar by the start of 2010 (I had sold in the $10-12 range) but being right about one specific idea does not mean that I had an optimal portfolio in terms of weighting.
Investors also want to be able to stay in the game and having a highly concentrated portfolio, irrespective of the quality of analysis behind the holdings, can result in big losses which reduces the capital available to be deployed in other promising prospects down the road. A portfolio of roughly 25 holdings with 3-5% position sizes allows a person to take on enough company-specific risk and generate market beating gains without having potential knock-out blows from an errant large position. I have found it's also easier to make better judgments regarding positions with a slightly broader portfolio.
For example, investors are often tied pretty tightly to high conviction picks. I have found it's very difficult to be wholly objective with high conviction ideas because of all of the work invested in the analysis of the company. As a result, it takes a long time to re-assess new information and it can be difficult to determine when to cut the reins. This is clearly not the case with every investor, but I assume most investors have experienced what I have just described.
In 2010, Sprint-Nextel ("S") was one of my highest conviction holdings. It took some time for the various catalysts to emerge and work through the company but overall, S did not do much for my fund and managed accounts. I sold out the entire stake after some level of consternation after Q3 earnings and the sizeable allocation of capital to S cost me performance in 2010. With smaller position sizes, I have personally found that there's less of the behavioral element at play that can hinder portfolio performance.
Position sizing should be clearly tied to one's own behavioral aspects and personality. Some people are very comfortable with a portfolio of just 5-10 holdings (assuming it represents most of their liquid net worth) but the biggest consideration needs to be what and investor believes is an adequate return for the level of concentration.
Consider a five stock portfolio that's equally weighted. Assuming 80% of the portfolio generates a 75% return and one pick was a 90% loss (assume it was an investment in Citigroup in 2008 or a net-net microcap that went bust), the total return is about 42% on that portfolio. As an investor, the question should be can I replicate that performance with a broader portfolio while potentially limiting the downside and risk of capital impairment if one or two holdings goes really bad? That answer and the historical ability of an investor to avoid big mishaps is likely to influence his/her concentration preferences.
