"Systematic alternative risk," depending on how you look at it, goes by quite a few names these days and from what I can tell it appears to be a very current thing -- if not perhaps overly current -- in institutional finance circles. Other names for this phenomenon can include (broadly speaking now): dynamic beta, hedge fund beta, exotic beta, smart beta, "tilts" of various flavors (e.g., value, size, and maybe especially momentum), etc. etc. There may or may not be important differences among these names, of course, especially with respect to "smart beta,"
but setting aside the overused term "smart beta" for a moment, many of the others at least tend to share some common features. These features generally include things like: an overlay of systematic execution rules on top of liquid instruments; long term exposure to definable, real, and compensate-able risk that cannot be diversified away; low or non-correlation with other mainstream beta (e.g., equity and credit risk...) plays; positive expected returns over the long run; generally lower costs and higher liquidity than hedge fund alpha (or pseudo-alpha I guess you could call it in many cases); relative transparency of purpose and execution; attempts to avoid (hopefully) big or sustained draw-downs; and higher trading capacity over the market as a whole for most investors. Or, to risk a really big over-generalization, in my view they all share at least four things: 1) positive return expectation, 2) low relative volatility, 3) low correlation with other asset classes, and 4) accessibility in terms of cost, liquidity, availability and capacity. This gives the systematic alt risk approach the potential to shift an efficient frontier up and to the left for the small guy. That's a minor holy grail. A grail that I want.
but setting aside the overused term "smart beta" for a moment, many of the others at least tend to share some common features. These features generally include things like: an overlay of systematic execution rules on top of liquid instruments; long term exposure to definable, real, and compensate-able risk that cannot be diversified away; low or non-correlation with other mainstream beta (e.g., equity and credit risk...) plays; positive expected returns over the long run; generally lower costs and higher liquidity than hedge fund alpha (or pseudo-alpha I guess you could call it in many cases); relative transparency of purpose and execution; attempts to avoid (hopefully) big or sustained draw-downs; and higher trading capacity over the market as a whole for most investors. Or, to risk a really big over-generalization, in my view they all share at least four things: 1) positive return expectation, 2) low relative volatility, 3) low correlation with other asset classes, and 4) accessibility in terms of cost, liquidity, availability and capacity. This gives the systematic alt risk approach the potential to shift an efficient frontier up and to the left for the small guy. That's a minor holy grail. A grail that I want.
The various alt risk premia these systematic approaches attempt to capture emerge from a whole slew of factors that sometimes, but not always, differentiate themselves from standard market risk and they may or may not come from sources like: pure accident or serendipity, exploitable behavioral tics (e.g., under/over reaction of market participants, fear of leverage, etc.), the relative willingness of investors to take or not take volatility/liquidity/fat-tailed (i.e., non normally distributed) risk, the exploitability of overlooked (orphaned) low-risk assets, and sometimes they come from a generalized aversion by investors to taking "long" risk in declining assets and short risk in rising assets (which is back to the behavioral thing). There may be other sources but who knows? I am certainly no expert or academic on this.
Some examples of alternative risk strategies, if one were to happen to Google this topic, can include things like: equity value (long/short equities with high/low fundamental values), GDP weighted sovereign bonds or long/short the yield or term structure of sovereign bonds, volume weighted commodities, tilts towards backwardation in commodity futures, various flavors of weighted real assets, currency carry based on purchasing power parity, catastrophe bonds, merger or convertible arbitrage, volatility variance swaps, short option vol, systematic trading rules that favor momentum or value or size or low volatility or whatever (I'm a whatever, more momentum than not plus a couple other things), and a whole bunch of other strategies that can extend as far as the eye can see it seems.
The question for me then is whether one has to be an institutional investor or a deep pocketed blueblood in order to capture some of the same value that your average underskilled-and-overpaid-slacker-2-and-20 hedge fund managers are trying to capture (often, by the way, they get what they call alpha on the sly from alternative systematic risk that is disguised as alpha but hush now...) but for a way lot less money and with much freer access to your own capital. This is a really important question for me since, if I didn't make my point in the last sentence, let's be clear that I think 96% (I made that up but it probably isn’t that far off) of managers are not worth their cost in fees and the abridged liquidity they force you to accept compared to the other 4% of managers who might actually be worth it. The problem then is that those 4% of other managers usually have their funds closed to new money and if they haven’t closed, then you and I (or at least I for sure) probably can't afford their minimums in my lifetime. That's why, based on personal experience I think the answer is yes, a retail investor can (should?) do it themselves, but it is a qualified yes. My caveat is that for any randomly selected 100 retail investors, I think that maybe one or two (or five, or…) have the attributes to make it work while the rest should probably stay closer to low cost indexing. On most days I would probably admit that even the "one or two" including myself probably should index as well. Generally speaking, though, the attributes of the "one or two" would have to look something like this for it to work:
- Interest in finance and trading as a non-casual undertaking
- Quite a bit of free time on your hands
- Really good discipline in following rules
- Has a low cost platform
- Has a strategy with firm, clear rules
- Has the patience of Job and a zen-like disinterest in market ups and downs
- Blind to the obvious: indexing and allocation are what really work
- Strong stomach for draw-downs
- Strong stomach for co-skewness (bad things happening all at the same time)
- Awareness that sometimes strategies work and sometimes they don't
But if one does happen to have these attributes then I don't think it is all that hard or expensive (in my opinion anyway; the example below costs maybe 20 bps excluding personal time). In my case, over a two year time frame (ok, I know two years is laughably short but that's what I have; I have older data but there are some data problems for me before 2014) from 2014 to 2015 I applied a systematic alternative approach to a set of simple, liquid ETFs and mutual funds which were split into two strategies, one focused more on traded fixed income and one focused more on a macro theme (e.g., currency, interest rates, market indices, geo-country themes, commodities, etc). I will say that some instruments such as options were used along with some spot forex trades but these turned out to be minor in the overall program. For these two strategies the risk premia arose, I think and I guess, mostly from behavioral sources related to momentum as well as other market behavioral tics. The exact details of the strategies are not as important in this note, though, than the fact that they: a) are theoretically accessible to and achievable by retail mortals, b) use very available and very liquid assets, c) are relatively inexpensive to execute, d) are not particularly time intensive, e) have more or less positive return and low volatility expectations, and, f) generally have accretive correlation attributes.
Note, by the way, that this was not an exercise in day trading or asset allocation, it was systematic long term rules-based investing. It was also, fwiw, not back-tested. Sometimes these things are hard to test and have to be done empirically in real time which is a little bit of a risk. In addition, I have to say that it feels to me like this approach might not work in a year or two -- but my main point here is that a non-institutional investor can create and operate strategies that nudge the efficient frontier up and left without either day trading or paying a second or third or fourth tier manager extortionate fees for a substandard and illiquid investment. This type of approach might even, I think, take on more importance now since valuations in many asset classes seem to be at historic highs (90th percentile for both stocks and bonds as of late 2015). In a situation like this I really, really don't want to hand my personal family risk over to a high fee, hard-to-sell anything. I want my own hands on my own capital with systematic rules to absorb some of the bigger shocks I know are coming. My guess is that others do too and that systematic alternative risk -- along with the bread and butter of allocation and diversification, of course -- will become more interesting rather than less in the coming decade.
Back to the strategy. Here's what it looked like after two years of effort:
The legend is as follows:
X axis - Standard deviation of returns, annualized
Y axis - Compound annual return
AGG - Aggregate Bond ETF
SPY - US Large Cap equity ETF
EFA - International developed equity ETF ex- US and Canada
IYR - Real estate equity ETF
- Yellow line - Efficient frontier of SPY and AGG, allocations 0-100%
- Dark blue dots - Efficient frontier of various allocations of AGG SPY IYR and EFA
- Red dot - Advisor managed sub-portfolio, conservatively allocated
- White dot - Asset allocation ETF with about 40% equity exposure
- Light blue dot - Model portfolio (~50/50 AGG/SPY and a small bit of EFA and IYR)
- Green dot - Strategy A: systematic macro style approach
- Yellow dot - Strategy B: systematic fixed income style approach
- Black dot - Weighted average of Strategy A and B, a proxy for the whole program
Are there any conclusions that can be taken from this? Maybe. First, it's easy to see that diversifying between stocks and bonds (the yellow line) can work to lower risk or improve returns per unit of risk depending on how one allocates…but we knew that already. Second, it's easy to see that adding additional asset classes (blue dots) to the two asset model can enhance (or destroy) the efficient frontier…but we knew that already too. Third, it looks like almost any solution might have been better than what a paid advisor executed for me (the red dot). Fourth, if I had happened to have a habit of making a fetish of returns to the exclusion of all else then I probably went down the wrong path but I'm pretty sure I didn't. Fifth, either I got really lucky (which I probably have been) or harvesting a broader set of return sources than equity/bonds by diversifying into systematic alternative risk programs can punch out the efficient frontier a little bit (little bit, i.e., I'm not looking for big-play home runs here) in a reasonable, constructive, cost-effective, and accessible way. I'll flatter myself and assume the latter. Or at least until it stops working.
Some additional reading:
There is a lot out there but here are a couple random choices
No comments:
Post a Comment