Sunday, June 15, 2008

The ABCP's of Pension Governance

The asset-backed commercial paper (ABCP) crisis that hit Canada last August highlights the need to look deeper into a subject that receives little attention in the financial press, namely, pension governance.

Several of the largest public pension plans in Canada, including the
Caisse de dépôt et placement du Québec, Ontario Teachers Pension Plan, and PSP Investments were among the institutions that held asset backed commercial paper in their books. (see Diane Urquhart's Another Made-in-Canada Defective Investment Product for a critical review of the Third Party ABCP market).

So what does the ABCP crisis have to do with pension governance? It turns out, a lot. Pension governance is the system of structures and processes implemented to ensure both the compliance with laws and the effective and efficient administration and investment of the pension plan.

In essence, pension governance covers the "who" and the "what" of decision making at pension funds. Who is ultimately responsible for investment decisions and what are the processes used to make sure investment and operational risks are properly managed, ensuring that the pension promise will be met without taking undue risks.

The six key areas of pension governance are oversight, legislative compliance, plan funding, asset management, benefit administration, and communication. In this post, I will focus on a few topics concerning asset management given that this area receives the most attention in the the media. Having worked at two of the largest public pension funds across public and private markets, including alternative investments such as hedge funds and private equity, I am well placed to highlight the main deficiencies that I see in the way pension funds report their performance.

The media typically covers annual reports in a very summary fashion. For example, we all read that the Caisse
de dépôt et placement du Québec (Caisse) realized gains of 5.6% in 2007 while the Ontario Teachers' Pension Plan (OTPP) earned 4.5% annual rate of return. OTPP's annual return was 2.2% above its composite benchmark while the Caisse's annual return was 0.6% below its composite benchmark (0.7% above its composite benchmark excluding the write-downs from their ABCP holdings).

Are you confused? What does this all mean? How do you properly compare the performance of big pension funds like the Caisse and OTPP? The short answer to this question is that you need to know what are the risk-adjusted returns of the pension investments and what expenses were incurred to attain these returns. A higher annual return can be achieved taking higher risks and paying higher fees, so beware of headline numbers. Also, keep in mind that some pension funds have different periods covering their fiscal years. The Caisse and OTPP's fiscal years end in December 31st while PSP Investments and the Canada Pension Plan Investment Board's (CPPIB) fiscal years end March 31st.

In order to properly compare the performance of pension funds, you need to know what are the underlying benchmarks of each and every investment activity underlying the overall returns and the annual standard deviations of these benchmarks. Using this information, you can derive the Information Ratio of each pension fund.
The Information Ratio measures the excess return of an investment manager divided by the amount of risk the manager takes relative to a benchmark. It is used in the analysis of performance of pension funds, mutual funds, hedge funds, etc.

The term "alpha" typically refers to the
annualized excess investment return over benchmark. The policy benchmark portfolio (the 'asset mix') is usually approved by the Board of Directors and senior management at the pension fund is responsible to beat the composite benchmark index. Typical benchmarks for stocks and bonds here in Canada include the S&P/TSX, the S&P500 and the Scotia Capital Bond Universe. Typical benchmarks for inflation-sensitive assets like commodities, real estate and infrastructure are some spread over inflation. The Board of Directors approve the policy portfolio and they approve the weights and ranges in each asset class. Any deviations from those weights are treated as active management decision, where the pension fund's senior management assumes the risk of being overweight or underweight a certain asset class.

Senior managers of pension funds are concerned with beating the benchmark for their asset class, beating the overall policy benchmark portfolio and achieving the actuarial rate of return over a period of time to make sure that the pension liabilities will be met. But their first interest is to beat their asset class's benchmark so they can receive their bonus at the end of the year. This is important to understand and I will come back to it later.

The problem is that very few pension funds explicitly report their benchmarks for each and every investment activity, their alpha targets and their annual risk risk budgets. For example, PSP Investments publicly states that its Board has set the following investment objectives on both absolute and relative performance:
  1. Absolute Performance: Achieving a return (net of expenses) at least equal to the actuarial rate of return as determined by the Chief Actuary of Canada; and
  2. Relative Performance: Achieving a return exceeding the Policy Benchmark return by 0.50% net of expenses.

But what is the approved risk budget to attain 0.5% net of expenses over the policy benchmark? In its Statement of Investment Policies, Standards and Procedures, PSP fails to disclose the benchmarks for its real estate and infrastructure investments citing "competitive reasons". Interestingly, on page 67 of the Annual Report 2007, we see that Real Estate accounted for the bulk of the alpha in FY2007, earning 36.5%, a whopping 29.8% above the benchmark return of 6.7%!!! Similarly, for FY2006, Real Estate returned 21.6% or 13.3% above its benchmark return of 8.3%. (Notice how the public market asset classes delivered close to benchmark returns).

Obviously, there is a clear "risk'" disconnect between the benchmark for real estate and the investments made into the real estate portfolio. If this is the case, then why not disclose the benchmark for real estate? Why is this important? For the simple reason that performance is tied in to compensation. If you look at the top five salaries paid out in 2007 at PSP Investments (page 50), you will see that 3 of them came in the real estate group.

When you are paying out these type of salaries and bonuses at a public pension fund, the public has a right to know what are the benchmarks used to evaluate the performance of each internal and external manager. Moreover, independent specialized consultants should conduct thorough performance and operational audits in each investment activity so stakeholders know that appropriate risks were taken to beat their benchmarks. These reports should go directly to the Board or government supervisors and their content should be publicly disclosed on the pension fund's website.

I am using PSP Investments as an example here but I can just as easily use other well known public pension funds in Canada (including the ones I mentioned above) to make my point. Over the last five years, the bulk of the "alpha" from every large Canadian (and global) pension fund came from private market investments into real estate, private equity, and infrastructure. In many cases, the performance of these asset classes trounced the benchmark by 10%, 20%,30% or more!!!

If compensation is tied to performance and benchmarks, doesn't the public have a right to know whether or not the benchmarks used to evaluate this performance accurately reflect the risks taken by the investment manager(s)?

The dirtiest secret in the pension fund world is that benchmarks used to reference the performance of private investments and hedge fund activities in public pension funds are grossly underestimating the risks taken by the managers to achieve their returns. Moreover, most of the "alpha" from these investment activities is just "beta" of the underlying asset class. Why are pension executives being compensated for what is essentially beta?!?!?

There is a disconnect between public market benchmarks and private market benchmarks. Most pension funds use well known public market benchmarks like the S&P 500 to evaluate the performance of their internal and external managers. Public market benchmarks are well known and for the most part, they accurately reflect the risks that investment managers are taking.

But there are no standard private market benchmarks; these investments are illiquid and valued on a quarterly basis with lags. This leads to some serious issues. In particular, if the underlying benchmark does not reflect the risk of private market investments, a pension fund can wipe out its entire risk budget if real estate or private equity gets hit hard in any given year, which is not hard to fathom in the current environment.

Given my training in economic theory, I believe in opportunity cost. Private market benchmarks should be based on public market benchmarks with a spread for illiquidity and leverage. Most pension funds use a spread over inflation, which sounds reasonable until you look at the risks they are taking to achieve their returns in private markets.

Benchmark issues also happen in public markets, but to a much lesser extent
(how many U.S. large cap investment managers do you know that beat the S&P 500 by double digits?). The money market guys that were buying ABCP to "juice up" their returns over T-bills easily beat their benchmarks for years until one day that market seized up (who would have thought the safest asset class, cash, would get whacked by contagion effects from the U.S. credit crisis?). The fund of hedge fund manager who invested in illiquid hedge fund strategies easily beat his benchmark of a spread over T-bills until certain strategies blew up once the credit crisis developed.

Let me be clear: there are serious issues in pension fund governance that need to be exposed. Pension funds should begin by disclosing the benchmarks for each and every investment activity underlying the overall returns and clearly identify who is responsible for each activity. They should stop bundling investment activities together and clearly report returns and benchmarks for internal and external managers. Only then will the key stakeholders understand whether or not the appropriate risks were taken to reach the returns, justifying the compensation being paid out.

It is up to the Board of Directors of each pension fund to ensure that each benchmark accurately reflects the investment risks of each investment activity, otherwise you risk vast discrepancies in compensation between your private market managers and public market managers. Pensions funds should be proactive and publicly post a document that clearly explains the benchmarks for each and every investment activity and whether or not they accurately reflect the risks of the underlying investment activity of both internal and external managers (for example, even though they are not perfect, see Ohio Public Employees Retirement System's investment policies).

Stakeholders do not have the expertise nor the time to scrutinize annual reports from large public pension funds. My aim here is to assist government supervisors, journalists, and the general public to demystify pension fund activities and to shed some light on issues that do not get the attention they warrant.

No comments:

Post a Comment