KPL's blog on Indian stock markets, science, technology and more
SEBI Registration (Research Analyst) : INH000000743
June 15, 2016
Different position sizing strategies explained
Seth Klarman (Baupost)
You diversify away most of the diversifiable risk by having a portfolio of 20 or 25 positions.
You should be able to tell a great investment from a good investment, so there is no sense in having the same size position with your best idea and your 100th best idea.
A position is defined as the total investment in a company’s securities (which could span different asset classes).
A concentrated position is a ~10% position (every 2 years or so)
A post here shows in the 5.5 years from Oct 95 to Apr 01, Baupost only made two 10+% investments, and five 7-10% investments.
We would own a 10% position in a senior, distressed debt investment where there was a plan in place, where the assets were very safe – either cash or receivables or something where we could count on getting our money back, and where we saw almost no chance of principal loss over a couple of years and a chance of a very high, meaning 20% plus, type of return.
We would not own a 10% position in a common stock that was just plain cheap unless we had a seat on the board and control, because too many bad things can happen.
Most of the time, our most favorite ideas have 3%, 5%, 6% positions.
Position size will increase when a cheap position becomes much, much better a bargain or when there’s a catalyst for the realization of underlying value.
A catalyst gives you a much shorter duration on the investment and greater predictability that you will in fact make money on that investment and aren’t subject to the vagaries of the market and the economy and business over a longer period of time.
New inexperienced managers will have some 20% positions which might even be correlated, that’s absurdly concentrated.
1% positions are too small to take advantage of the relatively few great mispricings that you can find.
Previously had a 10-position policy with each position at 10%. His reasoning then was based on the fact that estimating the probabilities and the odds (i.e. the gain if you win) is error prone, and his own experience is that many times the bottom three to four bets outperform the ones he felt the best about. Hence he decided to weight them all equally.
Recently he realized that if he has 10% positions it’s very hard to recover from a mistake (in the mortgage crisis, his bet on Delta Financial Corp (DFC) got killed).
Since 2008, most positions will be 2-3% (basket trade) or 5% (baseline) of the portfolio and if the seven moons line up he will allocate 10% (home run).
Basket trade: when the risk is slightly elevated he will buy a basket of companies with small weightings.
Sources: here, here, and his Dhandho Investor book).
Say he has $100M.
Top 5 names to be each 3-5% of assets.
Next 10 names to be each 2% of assets.
Want 20-30 positions at 1% each.
Every day, look at those 1% positions, and say I bought them cheap, should I be adding to them, should they be 2 or 3%? For the 5% positions, are they good enough to stay at 5% or should I bring them down to 1%.
Usually enters the position cheap, the stock moves up, sells off some at a high price, the stock pulls back, and it ends up in a position where you don’t want to buy and you don’t want to sell. If you did a good job in your research, these positions will do okay. Constantly evaluate whether you still want to hold the positions, can be impacted by industry news, general economy, level of interest rates, commodity prices, etc.
Lee Ainslie (Maverick Capital)
Each position top out at around 5-8% of the overall portfolio.
Average position size of 2.1%.
Diversifies also by industries, have 6 sector heads (consumer, health care, cyclical, retail, financial, technology) each with 7 analysts.
Calculates the Kelly formula recommended size for all positions including the new position. Then scale it down such that they sum to 100% without changing their relative weights, subject to a maximum of 10% for a position.
To calculate the odds, create base, best, and worst case scenarios, then probability weight them.
Remove correlation among investments by hedging to remove risk exogenous to the thesis of each investment.