I was watching business television and an articulate gentleman, the top dog for a rather large retirement fund, was the guest. Perhaps I’ve been living too long in a cocoon, but while discussing the current investment strategy of this large portfolio he made frequent references to their “risk budget.” Admittedly I have not been paying a whole lot of attention lately to what I call consultant-speak, so I had to do five minutes of research to discover what the heck he was talking about.

When I determined that the “risk budget” is merely a bastardization of a Markowitz-type portfolio optimization model, I was reluctant to keep reading but I did try. We all had to learn the stuff at school – theory developed in the early 70’s – but I find it hard to believe it’s still being used. For the layman, the idea is to use expected returns (for bonds, equities, real estate or whatever asset classes are deemed to provide returns with various degrees of risk), volatility (more risk assigned to the higher returns based on historical data), and covariances (different asset classes are not perfectly correlated allowing for diversification) to solve – using simultaneous equations – the portfolio mix of assets which optimizes the expected portfolio return for the amount of risk that can be tolerated.

Flip it around and massage it some and you can also express the whole exercize in terms of a risk budget. How far off the rails can actual (measured) risk go before you can no longer safely fund liabilities. The budget part defines the range of unanticipated risk that can be tolerated, and if suddenly more risk is being experienced than bargained for then presumably a committee decides what actions must be taken (policy adjustments) to get back on budget.

Whew! Sounds beautiful doesn’t it? After twenty years of examining returns, covariances and such I’m more inclined to adhere to this other model…..i.e. garbage in/garbage out. Historical data provides only very high level guidance and not much more than you can figure out by casual observation. Yes stocks offer better returns and are more volatile than bonds but only over extremely long time periods at that. Useful? Once you acknowledge that there’s no way to determine what the world will be like at the crucial ending point it all becomes rather academic. When is the ending point anyway? Twenty years? Ten years? Six months (bonus time)?

The same guest suggested that based on loads of research from third party sources, U.S. equities were becoming more attractive and that their weighting in the fund could be increased without compromising the risk budget.

The S&P 500 rate of return over the past year exceeded +27%. Doesn’t it seem a bit strange that a model that influences decision-making would suddenly find the return/risk profile of an asset class more attractive after the asset bested a 10-year Treasury bond by a multiple of 13.5 times? I’m just a simple fellow, and although properly schooled in quantitative techniques (finer details now long forgotten perhaps) I’ve learned from experience not to be a slave to statistics and theory. Simpleton that I am, even I can see a fly in this ointment.

But human nature demands an explanation (theory) of some sort – in reality the decision to take more risk is now based purely on a warm and fuzzy interpretation of data. Oddly, the models simply quantify those features we find warm and fuzzy – the U.S. stock market has ‘stabilized’ (translation…we missed some decent returns but now feel more confident) as the data clearly proves. As complex as they seem, the purpose of these models – designed by people who desire comfort – is to corroborate or justify the decision. Just as the same models justified the decision not to delve into the stock market before it paid handsomely. If today’s decision turns out to be wrong, just blame the model again.

Now I don’t begrudge the consultants – everyone has to make a living. A consultant espousing the virtues of budgeting for “risk” certainly sounds more contemporary than one selling the old Efficient Markets Hypothesis – even if it is the same thing but with a trendier handle.

Over the course of two months, bearishness has given way to warm and fuzzy especially concerning the U.S. stock market – which I predicted was the place to be (read my book) since mid 2010. A robust earnings recovery was well underway even if large segments of the economy were in dire straights. The evidence was aplenty.

And I was pounding the table in July (see When to rotate into cyclicals? NOW!)- the only model needed was a thinking brain – because analysts had clearly gone overboard revising their numbers down – more concerned about Europe and China than business fundamentals. NOW??? A proper risk budget should factor in slightly more risk for U.S. equities in the short term, not less in my judgement.

What models cannot incorporate are future anomalous events. The Bernanke Twist is on again, which will not be nearly as effective now that institutional investors will be looking to sell their bonds to divert money into riskier (yet suddenly more comfortable) assets. Holding all things constant, this should be good for the stock markets but nothing is ever constant. Selling will cause mid-term yields to rise unexpectedly (an error factor in those models) like it or not and suddenly owners of massive bond portfolios will be just a bit poorer as bond prices drop. Dividends and earnings will be slightly less valuable (slight downward adjustment in valuation or prices) for a brief period – we call these corrections – until organic economic growth picks up the slack.

Models & risk budgets keep everyone busy, and frankly most of the job postings I’ve seen for the past couple of years are looking to recruit math and stats-savvy younger folk who will collect and analyze mountains of data. More jobs can’t be a bad thing, unless of course the models are used solely as a substitute for responsibility in which case one should shove them where the sun don’t shine.