We’re often asked by capital allocators, what type of investors are you?  While one can take this as an existential dilemma, these firms “bucket” investment managers based upon a descriptive framework.  It helps with top-down asset allocation, as well as focusing efforts on understanding a manager’s approach to risk management.  These buckets are often centered upon asset class, geography, or strategy.  While the first two are fairly straightforward, thinking deeply about strategy buckets often poses unique insight into how an organization approach risk.  To help explain why, let’s walk through a thought experiment in portfolio construction.


Suppose a large pension has put out an RFP for an external manager.  The product the manager must deliver has explicit guidelines.  Its aim should be to outperform the S&P 500 while minimizing the relative tracking error.  The pension’s thinking is they want to replace an internally managed index account, but hope to achieve alpha instead of replacing it with an index ETF.  Their key concern is to have a firm sense of the beta of the active oriented product.  There is one strict rule for the portfolio to make this pension comfortable with the beta management: the portfolio must include all 500 constituents of the S&P and no other securities.  It must also be fully invested with no leverage.  Seems odd?  Maybe, but bear with us for a second.  The example will show us how different investment strategies may lend themselves to this type of mandate, while others may not.   Let’s walk through some examples.

Out-performance Edge?

Let’s start with value investors.  These firms often restrict themselves to a narrow circle of competence, thereby making investment diligence on hundreds of securities quite challenging.  These investors seek to determine the intrinsic value of securities through forms of fundamental analysis.  Scaling that approach across hundreds of securities can present significant cognitive and organizational difficulty.  This is not to be said it can’t be done, but it wouldn’t be a natural fit.

Contrast this to macro-economic investors, who formulate top-down perspectives rather than performing fundamental valuation.  A view on rates, crack spreads, or employment slack can be applied to certain sectors or swaths of the market, but other securities may be inapplicable to these macroeconomic trades.  As such, macro investors often express their theses through broad instruments such as ETFs or futures rather than individual securities.  The process behind formulating macroeconomic projections doesn’t often coincide with individual security selection.

Quantitative investors however are well situated to create such portfolios.  Factor-based investors specifically identify patterns through data that can be systematically implemented across a universe of securities.  These patterns are based upon each security’s computed sensitivity to the factor in question.  This makes constructing a portfolio of many constituents a lot more manageable, as the investment process is centered upon relative weightings.

Now, fundamental managers would scoff at the idea of being pigeonholed into owning 500 securities.  Same with macro firms.  It defies their concept of investment domain.  Hell, this RFP hypothesis is hardly the type of thing you see from a pension, but it is a good tool for highlighting a difference in thinking.

As silly as it may seem, some allocators are beginning to believe that their ability to manage an entire portfolio’s risk simply by cobbling together allocations to unique and talented gunslingers is a bit unmanageable, and can have unintended consequences.  Some don’t mind, can stand periodic volatility, and focus on the talent above all else.  They may also seek to manage beta through passive investments to complement their actives.  But the fact remains, the art of allocation isn’t just picking horses; it’s also managing the aggregate vrisk.  

Risk Management / Portfolio Optimization

Returning to our imaginary RFP, we talked above about the ability for different strategies to generate alpha through an investable domain restriction, but what about managing the relative risk?  This involves portfolio optimization, which is in our view, the most often overlooked aspect of evaluating investments.  The ability to systematically manage risk, which in this case is defined as variance, is not applicable to many strategies. Many of these strategies would view the very attempt as a fool’s errand (“over optimization”).  For example, a common criticism of mean-variance optimization when applied to active management is that it spits out vastly different optimal weights when faced with all-too reasonable uncertainty in expected returns.  This is anathema to a fundamental investor’s sizing of positions upon perceived valuation compared to downside risk, irrespective of whether they use Bayesian frameworks to systematize those views.

Fundamentalists balk at this fragility, let alone the idea that managing variance versus a benchmark is a form of sound risk management.   As one of my partners likes to say, the malaise of the hedge fund industry was delivering what many allocators wanted: little to no volatility and uncorrelated returns (which for the last market cycle has unfortunately been a challenging sell).  Most fundamental investors think of risk along the lines of permanent loss of capital rather than volatility.  Macro investors often augment their view of risk in terms of tail-risk, often computed by metrics like VaR, or through complex scenario analysis.  This is especially the case when instruments they use express non-linear payoffs or leverage.

Quant investors focus on statistical determinants of risk, such as mathematically friendly variance, which can be plugged into a variety of algorithms and equations.  An optimization problem like the S&P 500 portfolio is almost like a puzzle.  You need to fit pieces on a priori risk/return characteristics to achieve the optimal aggregate risk/return, not too dissimilar to what the capital allocator must ultimately do.  Quantitative financial literature has tomes on the evolution of this space, much of which began with Markowitz.  The fact that the model portfolio has 500 constituents is probably an enabler to this thinking, given that it diminishes the impact of individual standard error in impacting aggregate portfolio optimization.

So Who Manages Risk and How?

This is not to say that factor-based portfolio construction is superior to macroeconomic or fundamental strategies.  Intuition and a respect for tail-risk are hallmarks of proper risk management.  Given the non-uniformity of market outcomes, skeptics of quant approaches point to many a blow-up, that was thought of as statistically impossible, to substantiate this view.

But there is something said about thinking each way: risk managing for permanent loss through optimal security selection, understanding non-linear payoff the way that macro economic investors do, while leveraging the benefit of systematic implementation in a holistic risk framework.  Each has its strengths, and each has its weaknesses. Perhaps the key to finding investment solutions for the future is combining the best aspects of each.  Now what strategy is that!?