Having recently concluded an asset liability review that lasted several months, I recommended to our board that the current policy allocation be retained without changes. Our consultant concurred and the board agreed. During my 25 years in this business, this is the first time I can recall that the asset allocation remained unchanged following the periodic review.

There is always a temptation to make changes to produce a new and improved asset allocation, even though the benefit may be negligible or marginal at best. In fact, in my prior assignment at CalPERS, we found that retaining the 1993 policy through 2010 would have produced a higher return and lower risk than the actual policy, with all the changes, during that period.

Our goals for the recent asset allocation review were as follows:

  1. Increase the risk diversification of the portfolio
  2. Increase the yield component of the returns
  3. Reduce costs
  4. Reduce the downside risk to the funding ratio
  5. Avoid unnecessary complexity.

While we knew it would be difficult to achieve all five, we wanted to accomplish as many as possible.

Weighing the pros and cons

The School Employees Retirement System of Ohio is a $12.5 billion defined benefit public pension fund. The level of risk in our portfolio, as measured by standard deviation within the mean variance framework, seemed appropriate for the long-term funding goals. However, in the current environment, the expected return at this risk level is lower than the fund’s actuarial interest rate. Increasing the risk of the portfolio to reach for a higher return was not prudent, as it also would have increased the downside risk to the funding ratio. Lowering the risk was not an option either, as even lower expected returns would diminish funding progress. Therefore, staying at the current risk level seemed appropriate.

Next, we asked whether a mix of alternative assets could produce a higher expected return at the preferred risk level. Our consultant generated some alternatives with higher returns, between 10 and 30 basis points, with marginal increases in the Sharpe ratios.

To generate this higher return, the consultant included niche assets such as high-yield bonds, emerging-market debt and master limited partnerships with highly constrained allocations to each. In the end, the higher return and higher yield came with higher complexity and less risk diversification, as these assets are all equity-like in their risk characteristics. Also, we didn’t believe they would play a distinctive enough role in the portfolio to qualify as a strategic asset class. We do have exposure to these investments within broader asset classes, but not as a policy allocation.

Because the return assumption has the greatest margin for error, our conviction on these marginal return increases was low. We have higher conviction in the volatility estimate than the return estimate. Deploying new money into these assets at current high valuations in this late-cycle phase was too risky. So aiming for the marginal increase in returns was not worthwhile.

There are two risk-diversifying assets in our mix: fixed income and hedge funds. Hedge funds are a part of our Multi Asset Strategies (MAS) portfolio. Fixed income and MAS have risk contributions that are lower than their weights in the portfolio. All the other assets – public equities, private equity and real assets – are growth oriented, have risk characteristics similar to equities, and have risk contributions that are higher than their weight in the portfolio.

Increasing the fixed income allocation in this environment of rising rates and low expected returns was not promising. Another option was to reduce the 10 per cent allocation to MAS because of its higher fees, but the downside was that we would lose its risk-diversification benefits. In addition, the MAS portfolio has higher net returns than fixed income, with similar volatility. Retaining the MAS allocation improved the efficiency of the portfolio.

Same allocation policy, better efficiency

After assessing all of this information, we decided to keep our existing asset allocation for now. We’ll continue to review it annually; however, we believe we can try to achieve some of our goals by improving the efficiency of the structures within the strategic asset classes.

We are continuing to realign our MAS portfolio by including low-fee options, such as dynamic risk parity and specific liquid alts, without reducing its diversification benefit. Fees are now 1.4 per cent and 17 per cent for the MAS portfolio and we believe we can go lower. We have modified our fixed income structure to include a separate US Treasuries portfolio that can serve as a risk hedge in a market downturn. Allocations to private credit as an opportunistic strategy outside of our policy portfolio also have produced a higher yield than traditional bonds.

I believe we are achieving some of our goals without changing the allocation policy, through better portfolio construction. In this process, we are focusing on defining the role of each strategic asset class in the total fund and aligning the structure with the role. For example, MAS is meant to be a risk diversifier and to generate higher returns than fixed income with similar or lower risk. Our ultimate goal is to refine portfolio construction and management to deliver enhancements when the policy asset allocation may have reached an optimal plateau.

Farouki Majeed is the chief investment officer of the School Employees Retirement System of Ohio.

“Absolute size, by itself, is no indicator of success and achievement, let alone of managerial competence. Being the right size is.” – Peter Drucker, author and management consultant.

 

What is the right size for an investment fund’s internal staff? What are the drivers behind the size of a fund’s front-office operations, responsible for day-to-day investment activities? How big should back-office functions – responsible for governance, operation and support – be in relation to the front office?

To answer such questions, and gain insights into the investment operations of large global funds, CEM Benchmarking undertook a study in 2016 of 26 such organisations with total assets in excess of $2.7 trillion. CEM looked at the full-time equivalent (FTE) headcounts of both front- and back-office staff. The study’s findings confirmed that staffing levels are a poor predictor of total fund cost once external management expenses are included. Further, while our clients often express a belief that the back office should have staffing levels similar to the front office (roughly a 1 to 1 ratio), the findings showed this rule of thumb is, at the very least, overly simplistic and in many cases not useful at all.

For front-office operations, the number of FTE primarily depends on:

Asset size – the more assets funds have, the more front-office FTE they have

Asset mix – the more illiquid assets funds have, the more front-office FTE they have

Implementation style – the more internal management funds have, the more front-office FTE they have.

Within this framework, strong relationships can be seen, and close to 90 per cent of the variation in front-office FTE is explainable through a combination of these factors.

To illustrate the utility of the model, compare the FTE intensity of external active public equity to that of internal private equity. For internal private equity, one FTE would be expected to oversee about $0.2 billion in assets as opposed to about $2.5 billion for external active equity, a 12-fold difference (and this is before considering differences in economies of scale between asset classes, which can be profound).  However, the cost of internal private equity is lower than for external active shares, showing that having fewer staff members does not translate into cost savings.

Front office predicts back office, with some caveats

The survey did show a robust relationship between the size of the back office and the front office. However, the ratio was close to 2.5 to 1, rather than the often-cited 1 to 1. This single figure hides important findings:

There is a base level of back-office FTE required, regardless of the size of a fund’s assets under management

This base level varies dramatically based on fund complexity and the number of functions a fund chooses to do in house. For funds of lower complexity, the base level is about 20 FTE. For complex funds with most back-office functions performed in house, this number grows to 240 FTE

Once this base level has been realised, about 1.2 back-office FTE is required for each additional front-office FTE, regardless of fund complexity and in absence of any changes to the investment program.

Funds with complex investment programs generally exhibited much higher FTE counts than would be expected based solely on their amount of assets. This increased FTE is a poor indicator of cost-effectiveness because:

The largest cost for most investment funds is external manager fees, with the cost of internal FTEs being relatively immaterial

Performing back-office activities in-house is often more cost-effective than outsourcing.

These more complex funds exhibited elevated headcounts across all back-office functions; however, the increase in information technology support was responsible for more than half of the increase. Based on the data (see charts below), it appears there may be a tipping point at which organisations feel it becomes cost-effective to in-source many back-office functions.

The more complex organisations cited several reasons for insourcing back-office functions, including:

Ability to run more robust and customised risk monitoring

Running complex derivatives programs involving a large number of over-the-counter derivatives.

Economies of scale vary by asset class

Another area of interest was economies of scale in the front office. While these were evident, they varied greatly by asset class:

Strong economies of scale were present in the management of indexed public assets, whether they were managed internally or externally

Weaker economies of scale for actively managed public assets, particularly internally managed active fixed income, for which economies were not significant

No economies of scale in the management of real estate or private equity, regardless of whether these assets were managed internally or externally.

Ultimately, asset size, asset mix and implementation style drive front-office FTE. Back-office FTE is driven by front-office FTE, but with important distinctions for larger, more complex funds, which have much higher staffing levels, relative to asset base, than smaller funds. That these higher staffing levels do not result in higher overall costs reflects the fact external management fees remain the biggest cost driver for most funds.

The findings of this study are of interest to funds that want to:

Plan for growth – how your FTE will grow as your assets under management increase

Prepare for change – determining how many FTE you will need to move into new asset classes or change implementation styles

Understand differences – why some organisations have more or less FTE than others.

 

Michael Reid is vice-president, relationship management, at CEM Benchmarking. Alexander Beath is a CEM senior research analyst.

 

The California Public Employees’ Retirement System has defined its 10-year capital market assumptions, which are the essential input for its four-yearly asset liability modelling (ALM) and, ultimately, for the asset allocation in its reference portfolio.

Determining the capital market assumptions is the first item in a six-step process CalPERS uses to set its policy portfolio in ALM workshops.

This year’s assumptions, which the internal staff presented to the board, showed much lower expected returns and higher expected volatility for CalPERS’ strategic investment classes than those developed for 2013.

As a result, one of the fund’s consultants, Pension Consulting Alliance, advised that “these latest inputs will very likely translate into lower long-term expected compound returns for the range of strategic portfolio options the investment committee will consider during the 2017 ALM process”.

The capital market assumptions the CalPERS internal team proposed, after input from its consultants and other players in the industry are:

Asset class                  compound return (%)           volatility (% standard deviation)

Global equity              6.8                                           17

Private equity              8.3                                           25.5

Fixed income              3                                              6.6

Real assets                   5.8                                          12.6

Inflation                      2.8                                           8

Liquidity                     2                                              1

 

The return estimates the $323 billion fund derived were modest compared with some others in the industry. For example, in private equity, its return estimate of 8.3 per cent is significantly below the projected return assumption of its private-equity consultant, Meketa, which forecasts 9.6 per cent. This is primarily because CalPERS has a lower expected return for global equity than Meketa. Both expect a premium for private equity above global equity of about 1.4 or 1.5 per cent.

All asset assumptions are based on inflation and a real risk-free rate, with assets then divided into three categories – low, medium or high risk premium.

Input from one of CalPERS’ other consultants, Wilshire, shows return prospects across all asset classes have been declining for decades, following the downward trend in interest rates. Core fixed income remains low, and the risk premium for every other asset class is anchored to fixed income.

Once the capital market assumptions have been agreed upon, the process for developing a new strategic asset allocation will begin. As at June 30, 2016, CalPERS asset allocation was:

 

Public equity               51.1%

Income                        20.3%

Real estate                   9.3%

Private equity              8.9%

Inflation                      6.0%

Liquidity                     1.5%

Infrastructure              0.9%

Forestland                   0.7%

MSCI’s long-awaited decision to include China’s A shares in its Emerging Markets and All Country World indices will affect more than $1.6 trillion in funds that track the MSCI Emerging Markets Index.

One large institutional investor, who has had a long-term commitment to onshore Chinese assets, says that although the Chinese economy faces serious issues, it has achieved a scale that makes its eventual integration into global financial markets a foregone conclusion.

“This should be a significant stepping stone to the real prize – inclusion into major global fixed income indices,” the investor says. “Foreign inflows in the onshore Chinese government bond market have the potential to counterbalance domestic capital outflows, which would improve onshore financial stability.”

MSCI is including 222 China A large-cap stocks, which will make up about 0.73 per cent of its Emerging Markets Index.

Remy Briand, MSCI managing director and chair of the MSCI index policy committee, says the expansion of the Shanghai-Hong Kong Stock Connect program has been a game changer for the accessibility of China A shares and contributed to inclusion’s approval.

Before making the decision, MSCI conducted an extensive consultation that included a large number of asset owners, asset managers, brokers and other market participants.

“International investors have embraced the positive changes in the accessibility of the China A shares market over the last few years and now all conditions are set for MSCI to proceed with the first step of the inclusion,” Briand says. “MSCI will reflect a higher representation of China A shares in the MSCI Emerging Markets Index when there is further alignment with international market accessibility standards, sustained accessibility within Stock Connect is proven, and international institutional investors gain further experience in the market. MSCI is very hopeful that the momentum of positive change witnessed in China over the past years will continue to accelerate.”

MSCI said in a statement: “This decision has broad support from international institutional investors with whom MSCI consulted, primarily as a result of the positive impact on the accessibility of the China A market of both the Stock Connect program and the loosening by the local Chinese stock exchanges of pre-approval requirements that can restrict the creation of index-linked investment vehicles globally.”

Combined, the Shanghai and Shenzhen exchanges make up the second-largest sharemarket in the world, after the US, but China represents only about 1 to 2 per cent of the world’s stock indices. The country generally is under-represented in investors’ portfolios.

In the last couple of years, the inclusion of China A shares has been delayed, due in part to investors’ concerns about the liquidity and replication risk that may result from potential renewed voluntary suspensions in trading of mainland Chinese companies on the local stock exchanges.

MSCI says international institutional investors welcome the expansion of Stock Connect and view it as a more flexible access framework than the current QFII and RQFII regimes. They also welcome the decrease in the number of suspended China A shares, but still consider the number high, compared with other markets. Investors encouraged Chinese authorities and exchanges to consider additional measures to address the issue.

During the consultation, the vast majority of institutional investors approved the proposal to include large-cap shares that are not in trading suspension. Many also recommended that MSCI include China A large-cap shares of companies that already have in the MSCI China Index equivalent shares trading in Hong Kong. Doing so meant an increase from the original 169 China A shares to 222.

Last year, when MSCI delayed China’s inclusion in the Emerging Market Index, Briand said investors had indicated they wanted further improvements in accessibility. This included the ability to move funds in and out of China and clarity on stock suspensions.

MSCI will launch its China A International Large-cap Provisional Index today, followed by additional global and regional provisional indices, including the China and Emerging Markets Provisional indices, in August 2017.

The $140 billion Teacher Retirement System of Texas is working successfully with its hedge fund managers on a new compensation structure that promises to “radically realign” the fees the fund pays. Speaking at his last TRS board meeting, departing chief investment officer Britt Harris explained that TRS hedge fund managers are taking up the so-called 1-or-30 model.

Under the model, TRS pays hedge fund performance fees only after managers meet an agreed upon hurdle rate. Managers can earn whichever is greater – either a 1 per cent management fee or a 30 per cent cut of the alpha or performance after benchmark.

“We get at least 70 per cent of the alpha, gross, after a hurdle rate; we’ve approached 22 managers and 17 or 18 are on board,” said Harris, who is leaving to become CIO of the $40 billion University of Texas Investment Management Company, UTIMCO, also based in Austin, Texas.

Under the model, in years when the fund underperforms, the management fee gets paid back to investors from the following year’s performance fees.

It’s a switch from the industry practice of charging investors 2 per cent of assets and 20 per cent of profits. Given TRS’s respected status among peers, it could set a precedent. Indeed, the fund has pledged to “lead the industry in improving terms for better alignment in a low-return environment”.

Hedge funds will have a place

Investors withdrew $106 billion from the $3 trillion hedge fund industry last year, in the largest annual outflows since 2009, data provider eVestment states. Yet Harris, who briefly served as chief executive at hedge fund giant Bridgewater Associates – of which TRS bought a minority stake in 2012 – believes hedge funds still make up an important part of the TRS portfolio when markets get more challenging.

“We haven’t needed hedge funds,” he said. “The market has gone pretty much straight up for eight years and, in general, hedge funds haven’t performed that well. There will come a time when we will need hedge funds, when markets go in the other direction. But right now, hedge funds are losing assets and are open to better compensation structures. We are collaborating with them to make sure that happens.”

The HFRI Fund of Funds Composite Index returned 3.2 per cent over the last five years, trailing the 7 per cent to 8 per cent return most US pension funds target. However, Harris does note that some strategies have performed well in the tougher climate, like event-driven and restructuring funds, both of which have “shot forward”.

TRS has been investing in hedge funds since 2001, deploying capital to AQR Capital Management, Fir Tree, GoldenTree Asset Management, MKP Capital Management and PDT Partners, among others. The fund has $10.6 billion in hedge fund investments, accounting for 8.3 per cent of its total assets, and paid $203.5 million in management and performance fees to them, directly or indirectly, in the fiscal year to August 2016.

Focus on fees across the fund

At the trustee meeting, Harris and his team touched on ways to improve the monitoring of fees, relative to alpha generation, across the fund. The fund has previously pledged to customise all its investment strategies where appropriate, to better reflect TRS’s needs than off-the-shelf products could. Harris also noted in his presentation the success of co-investment and principle investing at driving down fees.

“We have $11 billion in principle investments in private equity and real estate,” he said. “Both those portfolios have performed at or above the peer benchmark. On the private-equity side, there are essentially no fees, and on the real-estate side it is a significantly reduced fee.”

Harris leaves the fund in robust financial health after 11 years as chief investment officer.

TRS told its trustees of strong first-quarter results, with investment earnings at “a pretty remarkable” $6 billion for the first three months of 2017, equating to a 5 per cent return for the fund. Performance has been driven particularly by the emerging market equity allocation, where MSCI benchmark returns have come in at 11.4 per cent for the first quarter of 2017. TRS has $14 billion of exposure to emerging markets, or 10 per cent of fund assets.

TRS’s investment team also attributed the strong results so far to “asset allocation positioning”. Namely, being underweight US Treasuries in the stable value portfolio and, therefore, avoiding their relatively poor performance. Strong returns in the high-end core real-estate portfolio also pushed the result northwards, although Harris’s investment team now expects that asset class to slacken off.

“The expected return for real estate has gone from high towards low,” said Harris, who joined TRS when the pension fund was valued at $100 billion. Shortly after his arrival, that value dropped to $67 billion, due to being ravaged by the 2007-08 financial crisis. Today, he leaves with assets under management doubled from that time.

Harris also attributes the strong returns to historically low volatility.

“Market conditions, the volatility of individual asset classes and the correlation between asset classes has allowed us, and other people, to generate these returns at half the risk over this period of time. This will not last.”

TRS’s fund is divided between a global equity portfolio that accounts for 58.7 per cent of assets and a stable value portfolio accounting for 16.5 per cent of assets, half of which is an allocation to US Treasuries. Real return accounts for 19.8 per cent of assets, comprising real estate and other inflation-hedging holdings; while risk parity accounts for 5 per cent of assets under management.

During the two years to the end of December 2016, Japan’s Government Pension Investment Fund, the biggest investor in the world, decreased its domestic bonds exposure by 10 per cent, re-allocating the assets to domestic and international equities.

This has been a relatively quick move away from bonds, considering the extent of GPIF’s bond portfolio and the size of the fund. At the end of 2008, the fund had more than 75 per cent of its assets in domestic bonds, with only 6.6 per cent in international equities and 9.4 per cent in domestic equities.

Since December 2014, domestic equities have increased by 4 percentage points, to 23.76 per cent, at the end of December 2016, and international equities have increased by 3.5 percentage points, to 23.16 per cent of the fund.

The fund has $1.3 trillion in assets. It now invests in more than 2120 listed Japanese equities; the largest holding, by dollar investment, is Toyota, at 188,430,900 shares.

Globally, GPIF has holdings in 2596 companies, with the largest including Microsoft, Verizon, Johnson & Johnson, Exxon Mobil, Facebook, GE, Nestle, Wells Fargo, and Procter and Gamble.

As the fund has increased its allocation to equities, it has also become interested in stewardship. This month, it asked all of its external asset managers to disclose the details of their proxy voting records on behalf of GPIF.

In a statement, GPIF president Norihiro Takahashi, said: “GPIF believes that disclosure of the details of proxy voting records is very much essential for institutional investors to fulfil own stewardship responsibilities in order to deepen corporate governance reform and move its focus from ‘form’ to ‘substance’ as Japan’s Stewardship Code indicates. GPIF shall continue to enhance the mid- to long-term investment returns for our beneficiaries through improvement of corporate value and fostering sustainable growth of investee companies.”

As previously reported, in 2016, all of the fund’s external asset managers exercised their voting rights.

GPIF uses managers rather than investing directly, because its size makes it too influential. It generally limits a stock owning to 7 per cent. The fund has previously stated that its external managers with poor governance will get a smaller part of the cheque.