Let's say the investor has a max tolerable drawdown of 33% and an expected return of 5% over the S&P500. Should the portfolio have 10 positions with 10% each, 5 with 20% each? Less? More? Precision is not important, what kind of position sizing makes sense for this portfolio?
There is no drawdown if stock positions are not closed. An investor with a stock portfolio.... Jesus...What is it you are asking, exactly? My brain just went south and west on this. What is your issue?
its called risk aversion, not everything is a function of withdraws or forced selling. perhaps I need to give up on ET, for every good reply there is like 50 piles of crap
After 11K posts, it seems likely you will never give up ET. So bear with us newbies here. A long-only stock portfolio...What drawdown are you talking about here?
You are confusing terminology. Position sizing refers to how many shares of a stock or instrument do you buy to keep risk at an acceptable level. It is usually based off where you put your stop loss. If you are trading that is different than investing. In trading you USUALLY stops even if its a mental stop or area where you would termintae the trade. In investing frequently you do not set stops. Also, what stocks you invest in will make your allocation risk change dramatically. If you only invest in commodities and telecom related stocks you would have more volatility. If you invest in staples and utilities to get your measly 5% return you can forget worrying about drawdown too much because you could probably easily find stocks with yield of just under 5% that will be slow and steady or even several that average 5% yield. If you want 5% growth you are still only talking about slow sticks with likely low volatility. The lower the volatility the fewer stocks you need to diversify (IN GENERAL). As you set a stop loss of where you get out then the larger your stop the fewer shares you buy to limit your risk. So you need to decide are you investing or trading and are you setting hard stops or mental stops or are you NOT setting stops at all? You alocation varies as you choose sticks with higher volatility. Why not just choose 15 good solid stocks with a nice yield and check on it periodically like once a week or so. You should get the best of both world there with less downside risk.
I think that you would need to investigate the volatility of the stock tickers which you select. And the correlation between these. If you select 10 tickers and all move up and down in the same rhythm it wouldn't help to limit your drawdown (= account value fluctuation). However, if all 10 were not correlated to each other your drawdown would be much smaller. So my approach would be to first find some tickers that have a history of not/hardly being correlated. And give a weighing factor to each which is inversely proportional to their volatility. In other words: if the volatility is low I would give it a large position size (and vice versa). Having composed the portfolio I would then look at the return of it and see whether I meet the desired target of (S&P500+5%).
You are trying to work out the best compromise between two opposing effects: Diversification, which says the more stocks the better. Given enough money you should own the entire S&P 500 A target return over benchmark, which will require more a concentrated portfolio to achieve. It's relatively easy to work out the diversification cost of a concentrated portfolio. As it happens I've been doing these kinds of calculations a lot for my new book. In geometric return space you probably lose around 1.2% a year from holding one stock rather than the whole index. It's harder to work out the benefit of concentration, since expected return is a lot more uncertain. 5% over benchmark for a long only portfolio is very punchy. If you really need to achieve that kind of number you're going to need very few stocks, and to be a great stock picker. Since it's a long only portfolio at best you will inherit the max drawdown of the S&P 500 (max drawdown is a function mostly of risk not return); so again 33% is a bit optimistic since the S&P has had a larger drawdown than that twice in the last 20 years. The more concentrated your portfolio is the more likely the max drawdown will be higher. Personally I'd want to own at least one stock per GICS sector, so 11 stocks. That will give you most of the diversification of a larger index, whilst being relatively concentrated with a shot at the 5% over index excess return. GAT
You mentioned "diversification cost", I suppose you really mean "rebalancing cost". Diversification doesn't increase returns but only, decreases volatility. If one notices increases in returns coming from diversification, its likely due to rebalancing This paper explains this well http://www.bfjlaward.com/pdf/25968/65-76_Chambers_JPM_0719.pdf With regards to drawdowns, if the portfolio loses 33%+ due a large S&P500 decline, that's more acceptable. That wouldn't be tolerable would to mainly lose it due to stock picking
If you are using the correct measure of geometric returns, which I was careful to specify, then diversification improves them. This improvement is much larger than the benefits from rebalancing, which are relatively modest. GAT
No kind of portfolio balancing will achieve that you say. You mention +5% annual overachievement over index, that implies Alpha, so you have an alpha generator that can provide this? The other part is that you want to limit MDD to -33%, that can be achieved with hedging using futures or single stock futures, but you also need here an method (black box, algorithm, etc) that tells you then to be hedged and not to be.