If you have a good portfolio – i.e. did the “right” things and diversified internationally, tilted to value, etc. it hasn’t worked well recently. This is where the mettle of quality advisors is tested. Can we keep clients on-board and on-track, continuing to do the right things even when it hasn’t worked for a while?
Of course, there is a difference between perseverance and stubbornness, but I think this is perseverance. Diversification and value have too much evidence – evidence that is pervasive (in lots of markets and asset classes), persistent (in lots of time periods), robust (to various specifications), and economically meaningful (makes you money, not just a statistically significant t-stat).
This is similar to the late 90’s in a way, but there the pain was brief and acute (huge underperformance for about four years), here it is more of a chronic and dull pain – it just goes on, and on …
I thought it would be helpful to everyone if I walked through how to think about it. This will be a little long, but hopefully useful.
First, assume we have two asset classes, Stocks and Bonds. Assume there is no serial correlation in the returns (i.e. no momentum or reversals). Bonds have a lower expected return than Stocks, but less risk too. The mix in that case will entirely depend on risk tolerance – the psychological risk tolerance, the time horizon doesn’t matter. Remember in our set-up I specified no serial correlation. In reality there is negative serial correlation in the short run (i.e. reversals in daily returns), positive serial correlation in the medium run (i.e. momentum over a year or two) and negative serial correlation again in the longer run (i.e. reversals in the five to ten year period). The first and last of those are not tradable (transaction costs kill daily unless you are a market maker, and after a run up stocks may have lower long-run expected returns but it will still almost always be higher than bonds). So, suppose we settle on 60% Stocks and 40% Bonds.
For the 60% stocks suppose we have two options:
- US Stocks
- Int’l Stocks – same expected return as US Stocks, same risk as US Stocks, but not perfectly correlated.
If we are merely trying to maximize risk-adjusted return, it is clear we should split Stocks 50/50 between US and Int’l. But there are two reasons not to:
- Most people benchmark (at least partially) off of their family/friends/neighbor’s returns. And those folks are overweight US. So going 50/50 won’t maximize happiness since a shortfall compared to others will be more painful than a surplus compared to others would be pleasurable. This is the psychological reason (Kahneman and Tversky’s Prospect Theory combined with Framing).
- This is subtly different from the previous point. We are competing against others for retirement resources so having a portfolio different from everyone else increases the risk of being able to obtain those resources. In other words, suppose my neighbor Bob will invests in US only and at retirement will be buying assisted living services. I invest 50/50 US and Int’l and will end up with either more or less than Bob at retirement. If I “win” I can buy more assisted living than him, but if I lose I get less. But here’s the subtlety, since the “Bobs” in the US outnumber me when US wins it will push up the prices of assisted living (more competition and more willingness to pay on the demand side) so it is entirely rational to partially hedge (since “not losing” is more important than “winning”). This is the objective reason.
So, given that, in my hypothetical example it might make sense to be 40% Bonds, 40% US Stocks, and 20% Int’l Stocks. (This isn’t an asset allocation recommendation, just an exposition of the thinking.)
Now, let’s take it one step further. Suppose we have two choices for our US Stocks:
- US Core
- US Value – higher expected return than US Core, same risk as US Core, and perfectly correlated (in reality the risk is actually lower, and the correlation isn’t perfect, but I’m making a point)
In that case, from a purely mathematical perspective you should invest all of it in US Value. Time horizon is irrelevant – do you want higher expected returns or lower ones? Everything else (in my set up) is the same! There are three reasons not to be so extreme however:
- Periods of underperformance are likely more painful than periods of outperformance are pleasurable. So, it would make sense from a psychological perspective to not go 100% US Value.
- In reality, since US Value is also lower risk, we should do even more US Value (potentially even if we had to short US Core to do so).
- In reality, since the correlation isn’t perfect having some US Core makes sense from a diversification perspective.
Those last two issues are offsetting and in practice not as big a deal. (Russell 3000 Value is 95% correlated to Russell 3000 – that’s not perfect correlation but it’s pretty close.)
Make sense? Time horizon is irrelevant except for from a psychological perspective where people are benchmarking off of undiversified (US) or untilted (no factor exposure) portfolios. Which they are. But I can’t easily provide a mathematical answer to a psychological question. It is a function of how well we can manage client expectations (and maybe how sophisticated the clients are). So that’s why I said, “This is where the mettle of quality advisors is tested. Can we keep clients on-board and on-track, continuing to do the right things even when it hasn’t worked for a while?”
Good portfolios lose regularly!
“The best way to measure your investing success is not by whether you’re beating the market but by whether you’ve put in place a financial plan and a behavioral discipline that are likely to get you where you want to go.”
“In the end, how your investments behave is much less important than how you behave.”
Both of those quotes are from The Intelligent Investor by Benjamin Graham (1965 edition).