The Future of Investment Manager Due Diligence (and a Look Back at Q1 Performance)

As the optimist plummeting off a skyscraper said, “so far, so good.” Mid-sized companies fared best, while value and growth styles fared about the same in aggregate, although core lagged. This was one of those time periods where the stuff in the middle surprised by not performing in between the stuff on the ends – mid cap outperformed large and small while core underperformed value and growth.  I use Surz Style Pure® style and country definitions throughout this commentary, as described here.

On the sector front, energy stocks fared best, earning 15% in the quarter, as concerns about the Middle East drove up oil prices. Industrials and healthcare were the next best performers, earning 9% and 7% respectively. Rounding out the range of sector results, the remaining sectors returned around 3.5%, far less than the leaders.

Looking outside the US, foreign markets earned less than half as much as the US in US dollars, delivering a 2.7% return, led by Canada and Europe ex-UK, each with 8.5% returns. Europe ex-UK benefitted from a weakening dollar, so currency effects added 5% in the quarter. Several regions lost value in the quarter: Japan, Emerging Markets and Latin America. For Japan, which lost 3.5%, it was a continuation of a decade-long sequence of disappointing returns, in this case caused by the earthquake and nuclear reactor problems. For Emerging Markets and Latin America, this quarter marked a reversal because these regions had been leading foreign markets. As a result, the EAFE index, which had been lagging the total foreign market for some time, was in line with the total market for the quarter.

Now you have a framework for evaluating investment manager performance in the first quarter of 2011. But this is only one small aspect of the overall due diligence process - a process which should be forward looking, focused on developing confidence in managers’ abilities to add value. After all, clients always have the choice between active and passive management. Your recommendations of active managers are very important, which is good.

Investment manager due diligence is no longer a joke

Investment manager selection will become a matter of fiduciary responsibility if the Department of Labor (DoL) has its way. In October of last year the DoL recommended a change to a 35-year fiduciary exclusion. In the future, all advice regarding the selection of securities and/or investment managers would be a fiduciary act subject to fiduciary liability. In other words, advisors need to perform real due diligence. As is usual in fiduciary matters, ignorance is not bliss.

Most client-facing advisors trust others to screen investment managers, and this trust is being implemented more and more through outsourced due diligence. This is a joke. You frequently hear laughable sayings like, “I have never met a money manager who has performed below median,” or, “every manager wins against the right benchmark.” But clients aren’t laughing, and neither is the Department of Labor.

Advisors need to take back control of due diligence by following a straightforward process:

1. Acknowledge that there are problems
2. Care enough to fix them
3. Fix them

The problem

The key problem with manager due diligence is laziness. You just don’t care enough to do the job right. Einstein said “Everything should be as simple as possible, but no simpler.” You’ve made due diligence way too simple. You make poor decisions as a result.  Here are some of the current practices that should change:

• Pigeonholing managers into style boxes. Only index huggers live in a box, and they have tricked you into believing that tracking error is risk. Tracking error measures conviction.
• The 4-corner solution for asset allocation: large-value, large-growth, small-value, small-growth. This is the cart before the horse. Finding talent should come first, then asset allocation. A lot of investment talent is missed because a premium is placed on index hugging. Index-hugger talent is   “optimizing” to a benchmark fit, which requires good software. There’s nothing wrong with this, except other potential talents, like picking good stocks or economic sectors, never get on our radar screens.

• Relying on name-brand indexes without question. Russell, S&P, & MSCI are name brands that generally rely on the price/book ratio for their value-growth classifications, probably because professors Fama and French used it in their widely accepted style research. The current financial crisis, however, caused the book values of some major financial institutions to be grossly overstated because bad loans have not yet been written down. Consequently, these companies are classified as deep value. By contrast, a price/earnings-based classification views these companies as growth. See here for a complete discussion of this topic.
• Peer groups are used to evaluate managers. All CFAs learn the serious problems with peer groups: survivorship, classification and composition biases render peer groups useless, and no one can make these biases go away. But everyone, including CFAs, ignores these biases because it’s easy to leave them unchallenged. Decades of use have led you to ignore well documented deficiencies. Marketers exploit this fact by finding a peer group that makes them look good. Ever meet a manager whose performance is below median?
• Attribution analysis is performed haphazardly. Attribution is the most powerful and forward-looking tool you have but it’s important to use holdings and to get the benchmark right. Otherwise you’ll be misled. Attribution pinpoints the reasons for success or failure, so looking forward you want confidence that the people, process and philosophy that produced the successes are still in place, and that the failures have been identified and are being addressed. Unfortunately most use returns-based analyses for attribution, but returns cannot pinpoint sector and stock decisions. Similarly those who use holdings-based attribution are limited to the popular indexes for their benchmark choices so liberated (non-index hugger) managers cannot be properly analyzed.  You rarely use holdings and always settle for a pigeonholed index as the benchmark.

Why you should care

What you tolerate you lack incentive to change. Nothing will change unless you want it to; you need motivation. Laziness alone is not a reason to change. TV remote controls are here to stay. You need to change the old practices because (1) the old practices don’t work so you select inferior managers (that’s the stick) and (2) differentiating yourself from the competition should bring more clients (that’s the carrot). The old ways cannot tell the simple difference between winners and losers. If you could tell the difference, the active-passive debate could be settled by using active managers where you are confident in their talent and passive strategies in parts of the market where you are not, so-called completeness funds, which fill in voids in your asset allocation. "There can be no transforming of darkness into light and of apathy into movement without emotion." - Carl Jung. You need to get excited about a better future!

In simplest terms, the truth will set you apart, and it’s not that much more difficult than the old ways you’ve been using for the past 40 years.

Solutions

Much has been written about the problems, but little has been offered to fix these problems. There is evidence to show that better managers are in fact selected when these problems are fixed. Here are some recommendations of solutions that work:

• Pigeonholing managers into style boxes. Use custom benchmarks. Indexes are barometers of performance in a market segment, like utilities or large-value stocks. Benchmarks are passive alternatives to active management, capturing the people, process and philosophy of the manager. Blended indexes can serve this purpose but the best indexes for this purpose are mutually exclusive (no security is in more than one index) and exhaustive (the collection of indexes covers the entire market). These are criteria set forth by Dr. William F. Sharpe for returns-based style analysis, plus they apply even more for holdings-based analyses. The Russell, S&P and MSCI indexes do NOT meet these criteria, but Surz Style Pure® indexes and Morningstar Style Indexes do meet these criteria.  
• The 4-corner solution for asset allocation: Find talent wherever you can – look for love in all the right places.  Many talented managers do not live in a style box, despite industry demand for style purity. The first order of business in portfolio construction should be to develop a talent pool where style doesn’t matter as long as the manager adds value relative to whatever he does. Then integrate this talent by optimizing across it and filling in voids with passive indexes where no talent is discovered. Dr. Frank Sortino has been using this approach for a long time, so there is precedent, and it is documented to work. See his recent book: The Sortino Framework for Portfolio Construction.
• Relying on name brand indexes without question. Surz Style Pure® Indexes use the price/earnings ratio as the primary value-growth differentiator, so they do not suffer from the current price/book distortion.  As mentioned above, they are also mutually exclusive and exhaustive. The opposite of mutually exclusive is overlapping membership, as is the case with the Russell and MSCI indexes. Overlapping membership causes a statistical problem in returns-based style analysis called multicollinearity, which inflates the goodness-of-fit measure (i.e., R-squared) and can produce erroneous results called spurious factor loadings, where the style profile is just plain wrong. Similarly, being non--exhaustive like the S&P indexes, which have 1,500 companies selected by committee, means many portfolios have holdings that are not represented. 
• Peer groups are used to evaluate managers. Peer groups should be replaced by portfolio simulations that create all of the portfolios the manager could have held, selecting stocks from a custom benchmark. This is classical hypothesis testing that compares the actual manager return to the universe of returns that could have been earned.  There are no biases in these scientific peer group substitutes, and they’re available days after report period end. 
• Attribution analysis is performed haphazardly. There are only two holdings-based attribution systems that support custom benchmarks – StokTrib from PPCA and Factset. If the benchmark is wrong all of the analytics are wrong. This very important tool in your decision making is worth the investment because it can easily add more value than a typical manager earning hundreds of basis points, and it can add this value over & over again for multiple clients.

A new version of StokTrib is currently in release and can be visited at New Attribution. It provides all of the solutions recommended here, and more.

Read more articles by Ron Surz