Paying the Top Guns of Institutional Investing

By Charles Skorina

Last month in Part One of this report we focused on relative performance.  We ranked 107 CIOs by trailing 5-year returns.

Now, we focus on how much institutions pay these excellent people.

The bare comp numbers lead us to the tricky and perennial question of whether their pay is properly aligned to their performance (or vice-versa), and we offer some analysis and opinion from the point of view of working headhunters.

We also consider the cost of an OCIO firm relative to an in-house CIO-led investment office.

Now, on to the charts!

N.B.1.: Data from 990 filings for calendar year 2014, disclosures, media, our estimates

N.B.2.: (Name) in parenthesis was CIO for all or most of 2014

N.B.3.:  Dollar figures in italics are Skorina’s estimates

 

Clarifications and caveats:

… Where’s my school?

If your school isn’t on the list, it’s probably because it uses a committee-and-consultant model and has no CIO.  Or, it’s too small. (Our lower-bound cutoff is ~$500 million AUM).  These are the same 107 institutions reported in Part One of this report in July, 2017.

…Why are assets and returns reported for 2016, while compensation is reported for 2014?

All data in this chart is latest available as of mid-August 2017.  Most of our compensation data is from federal tax filings (IRS Form 990) for FY2015.

Filers report compensation for the latest complete calendar year within the fiscal year, which is usually calendar year 2014.  This lag is annoying but unavoidable given current IRS policy.

Compensation for calendar year 2015 will become publicly available within the next few months, at which time we will issue an update.

In fourteen cases, the schools are not 990 filers and data is from other official or semi-official sources.  These are usually public universities (including Canadians) without parallel foundations managing their endowment funds.

In seven cases, we have made our own estimates of total comp based on our professional judgement and industry knowledge.  We think these are fair approximations, but we make no guarantees.  They are printed in bold italics, like this.

With a few exceptions, the returns are trailing 5-years annualized as of June 30, 2016.

We haven’t festooned our charts with a lot of footnotes.  But, if readers have questions or corrections regarding specific numbers, we’ll be glad to respond.

Assets under management (AUM) are generally for June 30, 2016, again with a few exceptions for schools with nonstandard fiscal years.  These AUM numbers are usually identical to those reported to Commonfund-NACUBO.

…The reported CIO (or OCIO) is as of late June, 2017.

Since CIOs come and go, this individual or firm is not necessarily responsible for performance over the entire trailing 5-year period.  And, because of the time lag in IRS reporting, compensation may pertain to a predecessor CIO who was in office for all or most of calendar 2014.  In those cases, the predecessor CIO is printed in parentheses.

We take CY2014 comp as the best available estimate of current compensation, whether or not the job has turned over.  CIOs tend to get year-over-year raises in line with university administrators generally, and for individual performance.  Average CIO pay in 2017 is probably at least 5 percent higher than in 2014.

…Fees paid to OCIO firms are included in this chart, although they are not strictly comparable to compensation of individual CIOs.

This issue is discussed below with a separate breakout of OCIO numbers.

————————————————–

The Problem with Harvard

Harvard Management Company is sui generis in the endowment world, and its CIO (actually, a CEO) has made twice as much as the next-highest-paying endowment chief in recent years.

Even if we used 3-year averages, the number would be conspicuously high.  And, as we and others have pointed out, it’s not easy to justify that generous salary given HMC’s mediocre performance over the past five years.

See: The Harvard Management Company: Time for some creative destruction?  http://www.charlesskorina.com/?p=3631

A lot of that comp has been performance bonuses banked by the incumbent in prior years and paid out three or four years later.

We won’t know Narv Narvekar’s comp for calendar 2016 for many months.  We know he doesn’t have any banked bonuses like Ms. Mendillo or Mr. Blyth.  But he may have received some kind of signing bonus.

On balance, we think his pay will be much lower than the $13.8 million we cite in the chart.  In fact, we think the bonuses for the senior staff at HMC generally will be reduced under a new formula.

But we are sticking to Ms. Mendillos’s 2014 comp as our official guess in order to be consistent in our methodology.

 

AUM, not performance, drives pay

In the next chart we’ve broken out pay into base, bonus, and total (“Base” includes “other” items in W2 comp which are usually minor.)

We’re down to just 80 CIOs here, having stripped out OCIO firms and also individual CIOs for whom we have only a single, total dollar amount. OCIOs are addressed in a separate section down below.

Note: more than a quarter of the CIOs on this list apparently have no bonus arrangement in their contracts. They are mostly at the smaller funds.

N.B.1.: Data from 990 filings for calendar year 2014, disclosures, media, our estimates

N.B.2.: (Name) in parenthesis was CIO for all or most of 2014

N.B.3.:  Dollar figures in italics are Skorina’s estimates

Bonuses are usually tied to a rolling multi-year average portfolio return.  A three-year window is typical to smooth our single-year aberrations.

Logically, we should see higher bonuses (as percent of total comp) for CIOs with relatively higher five-year returns.  Those with below-average returns should have lower bonuses in their paychecks.  This isn’t complicated.

But, our attempts to find any such pattern in this cross-sectional dataset fizzled. We could probably find such a pattern longitudinally (across time) for individual CIOs.  But we don’t yet have a good multi-year dataset for pay. (We’re working on it!)

Returns have only a very weak correlation to either bonuses or to total comp according to our regression analyses.  The R-squared statistic is minuscule: around 0.02.

On the other hand, there is a pretty robust relationship between AUM and total pay, with an R-squared statistic north of 0.6.

Size matters: bigger firms pay more … way more!

In the larger corporate world CEO pay is an object of great interest and controversy for obvious reasons.  But the relationship of size to compensation looks just like what we see in our set of endowment CIO data.

Kevin Hallock at Cornell University is one of the go-to experts in this field.  He’s chair of their department of Labor Economics and director of their Institute for Compensation Studies.  In papers with his students and colleagues he’s studied CEO pay for many years.

He says: “It doesn’t matter whether company size is measured as assets, market value, sales revenue or number of employees — bigger firms pay more … way more.”

“We can isolate the impact of all kinds of other characteristics (e.g., industry, return on assets, profitability, research and development expense, etc.) and even use complicated statistical techniques to remove the influence of “unmeasureable” characteristics, and the size-to-pay link remains intact.”

This isn’t just crony capitalists taking care of their board-room buddies, either.  The same relationships are found in non-profits (e.g., endowments and foundations) and labor unions.

Quintiles: from top to bottom

There is a simpler, more intuitive way to show these relationships.

We can condense the whole list (sorted by either AUM or returns) into five chunks (quintiles) and readily see whether comps are moving in step with the alleged explanatory variables.

We use median values instead of means to eliminate the tug of outliers like Harvard.  And, we drop the 11 OCIOs from our list of 107, leaving just the 96 individual CIOs.

Among the top quintile with its 19 mega-endowments, the median AUM is $9.6 billion, and the median total comp is almost $2 million.

As we move down toward the smaller funds, median AUM drops to $0.66 billion while comp goes to $260 thousand.

The CIOs running the smallest funds – which are twenty times smaller than the ones in the top quintile – make only about 13 percent as much as the CIOs running the big funds.  And the comp drops pretty smoothly quintile by quintile:

 

Looking at quintiles for 5-year RETURN and their corresponding comp numbers, we see a much looser correlation.

In the middle quintile, where returns are about average for all large endowments, comp is actually slightly higher than for the much higher-performing top quintile!  In the top half of this ranking, comp seems to be almost divorced from performance.

In the bottom half of the return ranking, there is moderate correlation of comp to return.

But the low performers still make more than 40 percent as much as the high-performing CIOs.

In other words, the drop in compensation from the quintile to the lowest quintile is much steeper for AUM than it is for performance.

 

————————————————–

 

OCIOs: the cost of outsourcing

There are 11 OCIO firms included in our big list of 107 at the top of this letter.  Here we break them out as a separate group.

 

This is our most recent complete OCIO list, as of September 2016.

The fees paid to OCIO firms aren’t really comparable to the pay of an individual CIO, but we thought it would be useful to include them in our master compensation chart for comparison purposes.

When a fund hires an OCIO they are, in effect, hiring a whole virtual investment office: the equivalent of perhaps a half-dozen full-time employees plus overhead.

For fiscal year 2014, large endowments reporting to NACUBO-Commonfund said their median internal costs of portfolio management were 48 basis points on AUM (i.e., 0.048 percent).

Because they have economies of scale not available to individual institutions, we would expect OCIOs to be able to offer fees below the client’s internal cost.  So, we would expect to see fees somewhat below 48 bps.

In fact, we see that fees in our small sample range from 8 to 41 bps, averaging about 25 bps.  Some older surveys suggest that OCIOs ask for between 30 and 100 bps.  Well, they can ask.  But we think some of those circulated numbers are optimistic.  The space is crowded and competitive, and getting more so.

We think 25 to 40 bps is closer to current actual fees.  We note that even an established, top-of-the-league firm like Alice Handy’s Investure seems to be getting no more than 41 bps in FY2015.

The two firms charging only 8 bps are special cases.  We think Hirtle gave Berea College a low-ball rate years ago on the grounds that Berea serves low-income Appalachian students, charges no tuition, and operates on a very lean budget.

The number for Verger really only represents the comp for Jim Dunn himself.  Verger is a proper North Carolina LLC, and also an SEC-registered RIA, seeking clients beyond Wake Forest.

But it appears that WFU is the majority equity-holder and is required by accounting rules to report Verger as a consolidated entity.  And that implies that it has to report Mr. Dunn as an employee on its IRS 990.  We think WFU’s total fee to Verger is two or three times that number, even though it isn’t disclosed.

If fees were the only consideration, the outsourcing decision would be a no-brainer.  But, of course, the decision is much more complex than that, an issue we will discuss in greater detail in our next Skorina’s ultimate outsourcer (OCIO) list.

Fiddles and finance: Navigating an inefficient market

Our amateur analysis of this dataset has convinced us of what we already intuitively thought: the market for investment management talent is pretty inefficient.  Paying a bigger salary doesn’t necessarily ensure that an institution gets better performance.

Fund managers are used to trading in relatively efficient markets.  Corners of it (such as U.S. treasuries) are so ultra-efficient that even quants wielding super-computers can’t find arbitrage opportunities.  Treasuries are uniform commodities, information flows at the speed of thought, and transaction costs are very low.

This is much less true in other corners of the investment universe, but there is still usually enough data and trading volume to give investors at least the illusion of rational cause-and-effect relationships.

In truly inefficient markets, things are very different.  Consider 18th-century Italian violins, for example.

A violin made by Guarneri in 1743, known as Il Canone Guarneri (Guarneri’s “cannon,” for its volume) was played by the 19th-century virtuoso Nicolo Paganini, and is worth about $4,000,000.

Fiddles like this are the opposite of uniform commodities.  They have histories and names; they are celebrities in their own right.  On the other hand, you can find good, though relatively obscure, instruments of the same vintage on E-Bay or Amazon for “only” a few tens of thousands of dollars.  They were never owned or played by anyone famous, so they don’t command a big price at auction.

Prices are also driven by larger market forces.  Collectibles in general, as an asset class, rise and fall in favor like any other asset class.

We would argue that the market we work in — the market for investment talent — is situated closer to the fiddle-end of the spectrum than to the T-Bill-end.  It’s lumpy, discontinuous and not terribly transparent.

Anyone eyeballing our big chart up above will easily pick out odd couples of CIOs who have pretty similar performance at funds of roughly the same size, but who have quite disparate paychecks.

Geography; personal chemistry; and, above all, the timing of hiring, all influence the value of CIOs in our inefficient market.  And, some schools just prefer to pay more of the CIO’s comp as bonus and less as base.  Others prefer the inverse.  This may have to do with the background of the board or committee members.

One important factor in market efficiency is transaction cost.  From the institution’s POV, hiring a CIO is expensive in both time and money.  And, as headhunters, our fee is a transaction cost.

We’re a very efficient firm, to be sure.  But, our inefficient market suits us just fine.

————————————————–

Things don’t just happen; things are made to happen

— John F. Kennedy

Modern financial economics argues that changes in asset prices and (therefore) returns are essentially random phenomena.  This Random Walk Hypothesis is consistent with – although not exactly equivalent to – the efficient-market hypothesis (EMH).

These concepts are powerful, elegant and mathematically tractable, and they’ve won Nobel prizes.  Sometimes, they even seem to correspond to reality.

Still, not everyone – theorist or practitioner — takes those ideas straight.  (Those that do buy index funds).

We are professionally obliged to believe with President Kennedy that people, including chief investment officers, can, to a significant extent, make things happen.

Burton Malkiel may have confidently issued the 11th edition of his seminal A Random Walk Down Wall Street.  But up at MIT, Andrew Lo, a rising star in financial econ, who’s first book: A Non-Random Walk Down Wall Street, came out in 1999, has published a brand-new opus:  Adaptive Markets: Financial Evolution at the Speed of Thought (2017).  It’s not only up to date, but surprisingly readable.

If you’re not ready to spring for the book, then a warm-up to his more recent work is a 2004 paper in the Journal of Portfolio Management: The Adaptive Markets Hypothesis: Market efficiency from an evolutionary perspective.

We’re hopelessly unqualified to evaluate his work.  But we can make an observation about its contingent usefulness to investors.

Assuming that Dr. Lo’s AMH turns out to be closer to reality than classical EMH, then it implies that investors can arrive at profitable investment strategies, at least for a while.

The profitability of strategies will wax and wane as the environment and the number of competitors change.  Arbitrage opportunities will open and close over time.  Then new strategies must be crafted to meet new conditions and exploit new opportunities.

An Efficient Market is a special case within the larger framework of Dr. Lo’s Adaptive Market.  A near-EM in his world is a possible, but relatively unusual situation.

In fact, Dr. Lo describes a world that is intuitively familiar to working investors.

The rationality of investors may be “bounded,” but it is still be good enough to find profitable investments, as long as they stay nimble and aren’t afraid to innovate.

This is a great relief to us, since it implies that paying people to manage money is not irrational.  And, by implication, paying us to recruit them might be a very good idea.

Charles Skorina works with leaders of endowments, foundations, and institutional asset managers to recruit Board Members, Executives Officers, Chief Investment Officers and Fund Managers. Mr. Skorina also publishes The Skorina Letter, a widely read professional publication providing news, research and analysis on institutional asset managers and tax-exempt funds. Prior to founding CASCo, Mr. Skorina worked for JP MorganChase in New York City and Chicago and for Ernst & Young in Washington, D.C.