A CFA Institute post on volatility...

Here is a blog post from the CFA institute about whether volatility is risk. That is a good question and something we have touched on before.  It is a very short cover and has some constructive, additive points on the topic.  On the other hand it biases, in my opinion, against spenders a tiny bit though it does, thankfully and to his credit, not exclude them from the analysis.  It's just that he says that for them it's more or less not that big of a deal. I disagree.  It is entirely possible that spending and expected returns will in fact dwarf the effects of vol but neither is vol trivial for us. It seems dumb for me to "digest" a one page article but digest we will (it'll take one minute to read so click thru too). Here are some of the main points from the linked post:

1. Volatility is one of the biggest risks in investing according to conventional financial wisdom. A small minority of investors, mostly value investors — a group to which I belong — take a different view. We think it is the probability of permanent capital loss, not volatility, that constitutes the real risk.  Neither perspective is entirely correct.
2. Warren Buffett famously said that as a long-term investor he would “much rather earn a lumpy 15% over time than a smooth 12%.”
3. Long-Term Investors with Strong “Stomachs” -- volatility is not a risk
4. Short-Term Investors -- Volatility is a primary risk
5. Long-Term Investors with Weak “Stomachs” -- should treat volatility as a risk for behavioral reasons
6. Long-Term Investors Who Consistently Spend Small Portions -- Volatility matters to some degree, but it is not the main risk.

Ok, let's take this one at a time (although I will admit that a former boss who successfully, by his wits and energy, accumulated hundreds of millions once told me that by the time you are responding to someone else point by point you are way behind or maybe even have already lost. None, zero, of his fairy dust landed on my head, in case you were wondering). 

1.  He is correct in this statement when he arrives at the last sentence.

2.  Warren must (tongue in cheek) not know finance theory.  He should really specify how lumpy is lumpy Let's look at the expected value of two strategies with the return and lumps (I pick the lumps since he does not say) he described. Here is a sim of geometric returns over time (25 periods) with: a): 15% return and 30% std deviation in blue, and b) 12% return and 15 % std deviation in red:

I don't know but that "12 % smooth" is starting to look ok.  My guess however is that he (Warren) is really talking about 15 vs 12 at the end of a process rather than the beginning in which case that should be clarified and after which I am pretty much on board with Warren before we talk about spending.

3.  In the context of the response to # 2 I'm thinking he is not 100% correct here. This should be clarified as well.

4.  Yeah, I think this is right

5. Yeah, I think this is right

6.  I think he is right-ish here but I also think that it is both underplayed a bit and also belied by his own simulation where he shows the risk can drop by orders of magnitude for spenders with a less "lumpy" portfolio.  Since he uses a fixed duration, just for fun let's up the ante a bit and use my new JPA tool which hews to the known theory of estimating ruin risk.  I'll "see" his two scenarios (s1 = 8% ret, 15% std dev, 5% spend; s2 = 8% return, 10% std dev, 5% spend) and raise with stochastic longevity so that we can get a more analytically rigorous take on this. We'll assume a 55 year old because we like early retirees and assume a mode and dispersion of longevity in a gompertz formula of 90 and 9 which is somewhere near the shape of the SOA annuitant mortality distribution for someone around this age. Now let's re-run his sims on JPA:

Scenario 1 - Lifetime ruin risk-- him (14%) me (12.5%)

Scenario 2 - Lifetime ruin risk -- him (3%) me(2.4%)

Actually this is interesting and I was kinda surprised. It looks like we are playing the same game and I thought I'd have some bigger edge in analysis and I don't[1]. But either way I still think that a drop of that magnitude in fail risk for lower vol cannot be ignored by spenders.  The question is maybe whether spending reduction or vol reduction will get a bigger lift for the effort. My guess is spending so he is probably right in the end. As a side note, he does not say anything about whether his returns are net (inflation, taxes fees) or whether they are arithmetic or geometric. If we ding his what-I-assume-to-be "linear and un-netted" 8% down to 5% and then re-run the lifetime fail risk is over 30%. That would get my attention but is not really the point of this discussion, is it?

My general comment, which I've made before, is that yes permanent loss of capital is risk more so than "volatility as volatility" but that: a) for portfolios without spending the long term effects of volatility on geometric returns can in fact be a form of loss (expected opportunity loss but loss nonetheless) depending on where one is starting with one's expectations, and that b) for portfolios with a spending constraint, volatility can precipitate permanent loss of capital in terms of an earlier or bigger fail via sequence risk or just the pernicious effects of lumpy returns.  Volatility is risk in both of these cases because we are losing real money one way or another.  One just has to decide how much (vol) is too much or how much less is best and that we have not answered. 

---------------------------------------

[1] here is a better view of changes in lifetime fail rate estimates for changes in vol for some scenario I was playing around with last week.