Sunday, October 24, 2010

Leo de Bever on When The Music Stops

Late Friday afternoon, I had a chance to speak with Leo de Bever, CEO & CIO of the Alberta Investment Management Corporation (AIMCo). As always, the conversation was fascinating as we delved deep into issues affecting pension funds.

A few weeks ago, AIMCo released its Annual Report 2009/2010. AIMCo's fiscal year ends March 31st, just like that of CPPIB and PSPIB. AIMCo’s total fund return was 12.0% for the year ended March 31, 2010, outperforming the AIMCo Composite Benchmark by 1.2% and the AIMCo Client Composite Benchmark by 1.5% (click on image below to enlarge):

AIMCo’s Balanced Fund Composite return was 17.8% for the year ended March 31, 2010, outperforming the Balanced Fund Client Composite Benchmark of 16.9% by 0.9%.

What I found particularly interesting was the section covering changes to performance benchmarks:

Prior to July 1, 2009, AIMCo measured performance against the asset-weighted composite of client investment policy benchmarks (i.e., the Total AIMCo Client Composite Benchmark). At one point, this measure had as many as 90 separate benchmark components. Some were minor variants of listed market benchmarks. Others reflected long-term inflation-adjusted total return aspirations that already included an expected return from active management. This structure stemmed from a historical focus on filling client demand for standalone investment “products”. It is not well-suited to managing total portfolio risk and return, or for measuring the effectiveness of active management.

A client’s policy benchmark should measure the net return from passively investing policy asset mix in listed bond and stock market indices. Unlisted allocations should be linked to the closest listed return and risk proxy. On July 1, 2009, AIMCo adopted internal performance management benchmarks consistent with this principle.

The 2010 fiscal year Total AIMCo Composite Benchmark reflects the Total AIMCo Client Composite Benchmark prior to July 1, 2009, and the new, market-based measures thereafter. The difference between these two measures will diminish over time. As client investment policies come up for revision, we are making considerable progress towards reducing benchmark complexity and setting separate active return goals.

We made these changes because there is merit in simplicity and conceptual consistency. We would have been better off last year using the AIMCo Client Composite Benchmark, but we believe the new benchmarks are the right benchmarks for the long term. Generally speaking, the absolute targets in the AIMCo Client Composite Benchmark will pose a higher hurdle for active management when markets are declining; whereas the new AIMCo benchmarks are more challenging when markets are rising.

Because this is a transition year, the remainder of the Investment Performance section reports on Balanced Fund actual performance and client composite benchmarks, as we have a longer return history for our Balanced Funds. Future annual reports will report only on the AIMCo Total Fund Composite.

As you can see below (click on image to enlarge), the benchmarks used for AIMCo's policy portfolio are transparent and mostly based on passive listed indexes:

Mr. de Bever was telling me how the benchmark for private equity (PE) used to be CPI + 8%, but this wasn't realistic. "Return on market risk in the 1990s was CPI + 7%. Since 2000, it's CPI + 2%, so you have to adjust your expectations accordingly".

We got into a long discussion on policy portfolios and active management. "My goal is to effectively deal with the end issue -- to deliver the best risk-adjusted returns. Policy determines 80% of the outcome." He referred to work done by Bruce Clark and John Campbell at Arrowstreet Capital (also see their extensive research on-line): "Mean reversion means there are better or worse times to be in equities but the problem is how do you determine that?"

I told him that right now, I see everyone nervously playing the Bernanke put because the fear of underperfomance is overwhelming the fear of another Black Swan event. Here is where the discussion really got interesting. Every senior pension fund manager and asset manager is facing the same concerns, and should pay attention to this part.

"Banks do not mark their commercial real estate to market. Quantitative easing (QE) is all about giving banks enough of a cushion to absorb these losses. For Bernanke, keeping the system afloat takes precedence over everything else. Not sure he's wrong but he's solving one crisis by sowing the seeds of another."

He referred to Carmen Reinhart and Ken Rogoff's book, This Time Is Different: Eight Centuries of Financial Folly and told me that he's worried that something might happen over the next few years. "What am I suppose to do? If I'm too conservative, I risk underperforming. But markets are schizoid right now and not at fair value. Markets are rewarding indiscriminately -- it's a very tough environment to operate."

I told him the problem is that you you can't afford to underperform for too long or else you risk losing your job. That's why everyone is playing up QE even though they're nervous it won't do a thing to bolster the real economy. There he agreed telling me that consumption is holding up because most people are defaulting on their loans.

And on underperforming the markets and your peers, he mentioned the "Brinson effect" referring to Gary Brinson, Chairman and CEO of UBS Brinson (see an interview here). "You might eventually be right but it could cost you your job. Even long-term investors have low tolerance for short-term underperformance. They're long-term investors as long as it makes money in the short-run."

"The problem now is if you're too early, you're worse off. Prince once said as long as the music keeps playing, I'll keep dancing. That's what's happening right now. But when it all comes together, I'm afraid that some catalyst -- like the shooting of the Archduke in Sarajevo -- will blow all this up."

Mr. de Bever's concerns are reflected in the discussion of economic and financial conditions in AIMCo's annual report:

In 2009, massive public intervention restored liquidity to a banking system paralyzed by the crisis of 2008. Credit spreads narrowed, and both bond and stock markets recouped a large part of the previous year’s losses. Output in nearly all countries recovered from recession, but at vastly different speeds. Recovery was, to a large extent, fuelled by expansionary fiscal and monetary policy. The question now is how soon increased private spending will allow public stimulus to take a back seat.

In setting our investment strategy, we have assumed that increased private spending may not happen for some time. High consumption and overinvestment in housing financed by cheap credit were a large part of the problem. Therefore, they are unlikely to be a significant part of the immediate solution. In North America, corporate profits are strong, but investment remains weak because of high excess capacity in many sectors. Many firms used the recession to improve productivity, so labour market recovery will be slow. Currency devaluation is, at best, a stop-gap for more fundamental adjustments.

Overlaying all of this is the longer-term fiscal pressure on health and pension systems from an aging population and the growing realization that some of the policies put in place many decades ago may not be sustainable. One senses a hunger for simple and painless solutions to this complex set of issues. That is probably naïve. Whatever political consensus eventually emerges will take time and involve unpopular changes to taxes and public spending. This is not the greatest recipe for robust focus on economic growth.

Anyone’s ability to predict the three- to five-year return-on-capital with any accuracy is very limited. If markets were predictable, the return-on-equities would be both much lower and more stable. The best a long-term investor like AIMCo can do is to identify superior investments that largely stand on their own, but reflect our best estimate of what the future will bring.

In listed markets, the near-term outlook for stocks is mixed despite strong earnings growth because of doubts about the strength of the economic recovery. That same worry makes the near-term outlook for bonds attractive. However, the longer fiscal stimulus and low regulated interest rates are needed, the higher the medium-term risk is for higher inflation and higher interest rates. Given the inflation sensitivity of our clients’ obligations, we continue to look for assets with long-term inflation sensitive returns.

We would like to find effective ways to participate in the rapid growth of Asia and South America. However, economic growth has a distressingly low correlation with high investment returns. In many cases, it may be more rewarding to invest in sectors that export to regions experiencing rapid growth.

In private equity and debt, we are still seeing good opportunities with the recapitalization of companies that are fundamentally sound but have overextended themselves in the period just prior to the 2008 crisis. The next few years are promising for direct investments in private equity.

Interestingly, the returns on active management in fiscal year 2010 came mostly from internally managed listed equities:

For calendar 2009 and 2010, the Board and management agreed on a stretch target of $500 million for net value-added. This represents about 1% of balanced fund assets. This $500 million is allocated to various asset classes, based on where we expect opportunities to be the most attractive and on AIMCo’s estimated capacity to exploit those opportunities.

For example, the calendar 2009 fixed income target was unusually large because we saw good prospects for capitalizing on 2008 market dislocation. Equities are given a large allocation in calendar 2010 because of improved capacity to focus on incremental return. Managers are given an active risk budget appropriate to their value-added target.

Fiscal 2010 value-added by active management above the Total AIMCo Composite Benchmark was $748 million. Short-term value-added is inherently volatile and active program effectiveness should be judged over longer periods

The positive fiscal 2010 results came mostly from internally managed listed asset classes. Last year was not a good year for unlisted assets. Real estate and infrastructure have done well over longer periods. Our externally managed private equity portfolio suffered from high costs and heavy exposure to vintage year investments obtained near the peak of the last equity boom.

We discussed the cost of external managers, and private markets at length. Here, Mr. de Bever is unflinching: "Go to the discussion on page 24-25 of the annual report. External managers charge management and performance fees. Our internal managers are paid for value added over a listed benchmark. You can literally replace external managers at 1/3 the cost. If you look at how partnerships are structured, a third of the cost goes to paying managers, a third goes to marketing funds, and a third goes to somebody who put up the capital for the fund."

"It really pays to go internal, but when I got interviewed on the results, most reporters didn't focus on the fact that we cut external manager fees down from $175 million in year 1 to $126 million in year 2, but rather on what we paid internal managers."

I then told him that most hedge fund and PE managers will argue that unlike them, pension fund managers do not have skin in the game, no hurdle rate or high-water mark to deal with, and that they don't take stupid risks with other people's money an then fall back on a four-year rolling return when they blow up.

"A lot of hedge funds close shop when they're underwater and reopen under a different name (very true). As far as pension fund officers' compensation, you have to be careful. That four-year rolling return which our long-term compensation is based on is not an absolute return but a relative return to the policy portfolio. It probably should be extended to seven or ten years but Revenue Canada rules made that impossible in the past. And for the most part, pension funds operate under strict risk parameters."

I then brought up 2008 and the blow-ups at large Canadian pension funds revolving non-bank asset-backed commercial paper (ABCP) and credit portfolios. "ABCP was a classic duration mismatch. The extra yield did not justify the underlying risk. As far as credit portfolios blowing up, there was a governance problem between the groups measuring risk and those managing risk. I think 2008 was a wake-up call and a lot of those issues were rectified. You cannot blame risk officers for those credit blow-ups."

I brought up the tyranny of quants and the illusion of normality. "There is a famous dictum: models are to be used but never to be believed. You need a healthy dose of skepticism and you need to separate out risk measurement from risk management. Returns are not normally distributed. We're seeing extreme events more often."

As far as compensation in the investment and pension industry: "It's true that people in the investment industry get paid a lot, but when you consider the tough operating environment and the unstable nature of our industry, then I'm not so sure the compensation is as outrageous as some people think."

And on picking the top external managers, Mr. de Bever had this to share: "I'm not sure you can pick top managers so easily. You can have 1000 Warren Buffett wannabes and at the end of 30 years one of them will have delivered exceptional returns. I operate under two assumptions: (1) try to beat the markets as schizoid as they are and (2) never assume you're better than the market because that's when you're going to fall flat on your face."

The discussion on private markets was particularly interesting. I told him that as more money floods into alternative asset classes like private equity, real estate and infrastructure, returns have been coming down.

"Underfunded plans are shoving money into these asset classes but they fail to appreciate that returns have to be earned. As more money goes into these asset classes, the market becomes more efficient, the opportunity set shrinks and returns come down." To make his point, he pointed out to what's going on in infrastructure right now. In commercial real estate, however, he sees opportunities in some regions where current pricing is "at very low values".

We had an interesting discussion on bonds and underfunded pension plans. I told him it's crazy to see pension funds cutting risk now by going into long-term bonds. The push into liability-driven investing is exacerbating the bond bubble because many pension funds are (erroneously) assuming that they're cutting risk by going into bonds.

"This is the dynamic instability problem. I recently showed our clients what will happen to a 30-year bond if interest rates go up 1% from here. With a duration of 16, you'll suffer a capital loss of 16% and wipe out your gains very quickly. I'm more comfortable with 5-year government bonds." He added: "Most of the money in the bond market was made since 1980. In the last 30 years bonds have made almost as much as stocks, which doesn't make sense."

Finally, Mr. de Bever shared these thoughts with me on the state of global pension funds: "Plan sponsors do not appreciate the difficulty in making money in this environment. Funding ratios in Canada should be increased to 125% to reflect a 60/40 stock/bond split. You can't improve risk-adjusted returns by taking more risk ( In other words: most pension plans do not understand that the price for taking more risk to get higher return is greater volatility of return). The cost of pensions have gone up by a factor of 4 and contributions have not gone up the same. Look at France, people are in denial."

And he left me with this stark warning: "Some time in the next five years, politicians will have to deal with a funding and labour crisis." I guess that's when the music stops playing for good.

[I thank Mr. de Bever for graciously spending so much time with me covering these important issues. Any errors/ omissions in this post are entirely my responsibility and will be corrected if needed.]

No comments:

Post a Comment