Thursday, August 24, 2006

2005 401(k) Plan Asset Allocation Account Balances

The EBRI/ICI Participant-Directed Retirement Plan Data Collection Project is the largest, most representative repository of information about individual 401(k) plan participant accounts. As of December 31, 2005, the EBRI/ICI database includes statistical information about: 17.6 million 401(k) plan participants, in 47,256 employer-sponsored 401(k) plans, holding $1.0 trillion in assets. The 2005 EBRI/ICI database covers approximately 37 percent of the universe of 401(k) plan participants, 11 percent of plans, and 42 percent of 401(k) plan assets.

Wednesday, August 23, 2006

Pension Protection Act of 2006 - Summary Resource

The Center On Federal Financial Institutions (COFFI) has drafted an exceptionally understandable Summary of the Provisons of the Pension Protection Act 2006 which impact the finances of the PBGC ( single employer defined benefit plans). They conclude it is unlikely that the Act will restore the PBGC to solvency.

They also publish a Primer on Pensions and the PBGC which is likewise exceptional in its usability. This document should be added to the DB plan fiduciary's required reading list.

Tuesday, August 22, 2006

A New Investment Measure that Predicts Performance

Martijn Cremers and Antii Petajisto from the Yale School of Management are on to something with their study entitled How Active Is Your Fund Manager? A new Measure that Predicts Performance. Their notion that active investment management decisions should be bifurcated and measured differentially by both stock selection and factor timing (aka market timing) correlates well with how the consultant community tries to measures active manager performance. The study suggets that funds whose alpha source is stock selection outperform funds that generate alpha by sector timing and funds that are closet indexers. The authors introduce a measure they call Active Share which indicates how different a funds stock holdings are from its benchmark index and therefore how actively the fund is managed. Based on what appears to fairly rigorous statistical analysis, the active share measure appears to have a unique relationship with fund performance and demonstrates some persistence over time. According to this study, funds with the smallest stock overlap with their benchmark(ie stockpicker), the smallest assets and the best 1 year performance seem attractive, outperforming their benchmarks by 6% per year. This study appears to advance academic work on active management though the calculations are based on fund holdings which can make repetitive use uneconomical.

Monday, August 21, 2006

NYC Pensions

NYT ran an interesting article Sunday on how differences in actuarial assumptions for New York City's pension plans can lead to significantly different pictures of a pension plan's financial status. It is not very clear in the article how the different assumptions were used by the Plan's actuary though the alternative liability valuation methodology reportedly used a risk free asset rate which would currently be in the vicinity of 5%.

The NCPERS 2005 Comprehensive Annual Report had a similar disclosure for last year (2005). The disclosure entitled Other Measures of Funded Status (page 173 ofthe report) clarifies the alternative funded ratio. The Plan's market value of assets is divided by the Market Value of Accumulated Benefit Obligations. In this calculation the plan's assumed rate of investment return is equal to a spread of US Treasury Security yields with durations consistent with expected fund payments (ranging from 4%-6% in this exhibit).

For this measure of funding to be relevant to readers, the exhibit notes prepared by the actuary suggest we suspend the fundamental laws of modern portfolio theory. We must ignore asset allocation..a decision that significantly influences a portfolio's returns and ignore the basic risk & return tenants upon which every institutional investor builds their portfolios.

2nd Quarter Pension Portfolio Performance Benchmarks

In addition to measuring your portfolio performance against market benchmarks it is informative to look at performance against institutional trust universe medians. Several are published quarterly and the more useful ones provide general asset allocation information as well. Care should be taken in drawing comparative performance conclusions however because the size, fees, objectives (foundations/endowments, public, corporate) must be considered.

For second quarter 2006:
the ICC median Master Trust returned -1%. This universe includes 21,000 portfolios with an aggregate market value of approximately $41.7 trillion.

the Mercer (Mellon) Us Plan sponsor Median Universe return was -1.1% for the second quarter. Asset allocations are not provided.

Wilshire provides return and asset allocation information from the TUCS universe. 2nd quarter information not yet available.

Thursday, August 10, 2006

Persistence of Manager Alpha

Mercer has published a practitioner study on the Persistence of Manager Performance. It concludes that the quest for good managers defined by past superior returns vs peer medians will invariably yield a high percentage of managers who underperform in subsequent periods. The author conjectures that the cyclical behavior of substyles within a mandate ( ie GAARP vs momentum) are more likely drivers of superior results than manager skill. These findings lead to the thought that investors might be better served finding good managers with poor recent track records (Bill Nygren, John Rodgers, Charlie Driefus anyone?). The study concludes that benchmark out-performance may be more persistent than peer universe out-performance. This may align with academic studies suggesting that momentum effects can drive short term performance persistence (1 yr).

Tuesday, August 08, 2006

Manager Alpha - Large Cap Value Funds

It is interesting to review the list of mutual funds that have provided the most and least alpha to investors over time (calculated over the longest period since fund inception or 20 years). For example, this list includes the 10 Large Cap Value funds that have provided the highest positive and negative percentage contributions to the asset weighted style and market benchmark alphas summarized in the August 2 post. These 20 funds provided over 50% of the asset weighted exposure for the 1047 LCV funds included in the study. Only 245 of the 1047 funds had a positive historical asset weighted market alpha. 802 funds had an asset weighted market alpha of Zero or below. The median asset weighted market alpha value was 0 (the largest 100 funds =81% of total value). The mean asset weighted market alpha was .10% indicating a skew to positive alpha.
Greatest LCV Asset Weighted Alpha Adding Funds

Fund NameStyle Alpha %Market Alpha %
Dodge & Cox Stock80.3%21.7%
American Funds ICA79.9%18.8%
Oakmark Select38.8%7.9%
Mutual Shares.57.16
T Rowe Price Equity Income13.2%6.4%
American Funds-Fund Inv27.4%5.51%
American Funds-Wash Mutual2.9%4.5%
Excelsior Value and Restruc26.1%4.15%
American Funds-Amer.Mutual3.1%4.02%
Oakmark I7.2% 3.23%

Greatest LCV Asset Weighted Alpha Detracting Funds
Fund NameStyle Alpha %Market Alpha %
Vanguard Windsor -18.9%-2.75
Fidelity Equity Income-27.7%-2.39
Pioneer Value-10.8%-1.75%
Vanguard Windsor Admin-11.9%-1.7%
SEI Inst Large Cap Value-8.5% -1.37%
Lord Abbett Affiliated-9.4%-1.14%
Van Kampen Comstock-9.6%-.96
Vanguard Value Index-3.7%-.95%
Putnam Fnd for G&I-8.6%-.8%
Consulting Group LCV-3.9%-.75%

(rank ordered by Market- positive and negative % sum to zero for the category)

Would be happy to post results from other fund categories if there is any interest.

Monday, August 07, 2006

DOL Proposes ERISA Plan Vendors Disclose Outside Revenues

In July the DOL proposed a number of changes to the Form 5500. The Department has proposed modifying Schedule C reporting requirements in an effort to clarify them and to ensure that plans obtain the information they need to assess the reasonableness of the compensation paid for services rendered, taking into account revenue sharing and other financial relationships or arrangements and potential conflicts of interest that might affect the services provided.

The proposal would require a range of investment service providers to disclose information about all the revenues and compensation they receive in connection with providing services to ERISA plans. This requirement could mean that all revenue sharing agreements (12b-1's, sub TA fees, commisions , soft dollars) must be reported to ERISA plans. Currently, many providers decline to disclose this information. Under the proposal, service providers who do not disclose this information must be identified and reported to the DOL. To emphasize the fiduciary responsibility to understand and control costs the DOl comments:

The Department believes that an annual review of such expenses is part of a plan fiduciary’s on-going obligation to monitor service provider arrangements with the plan. Requiring the reporting of such information
should emphasize that monitoring obligation.

Friday, August 04, 2006

Pension Protection Act of 2006 Passes Senate

The US Senate approved the House version of the Pension Protection Act of 2006 by a vote of 93-5. Major provisions of the legislation are ssummarized by Plansponsor.

In addition to addressing defined benefit and cash balance plans, the Bill exempts the provision of investment advice to plan participants by potentially conflicted advisors from the ERISA prohibited transaction category provided;
  • fees for the advisory services are not directly correlated to the investment recommended or a computer model is used which meets certain general criteria,
  • the advice arrangement is specifically authroized by a plan fiduciary,
  • the advice arrangement is audited annually,
  • certain disclosure and presentation requirements are met, and
  • plan fiduciaries remain responsible for the prudent selection and review of the advice providers.

In some way this legislation may protect or at least provide legislative remedies for eggregiously biased or unsuitable advise. Unfortunately, the proprietary product bias for many DC plans is already build into the structure and scope of their plan investment lineup. Advice will always be limited to the short list of options available in each plan.

The industry motivation to provide plan advice is to leverage institutional relationships (ie with the plan sponsors) into individual relationships with the participants in hopes of capturing the explosive growth of assets that will undoubtedly be leaving (talked out of) DC plans as boomers retire. Once in an IRA , outside of ERISA, advice can better serve the needs of the advisor's shareholders.

It will be interesting to see how aggressively and most importantly how economically DC plan providers develop and solicit participant advice arrangements. While we agree that many DC plan participants would benefit from better core asset allocations, the cost of that improvement could easily be offset by the explict cost of advise and the implicit biases to active management and proprietary investment products.

Wednesday, August 02, 2006

Active Manager Alpha Study



A recent study in the debate over the superiority of active investment management vs. passive management concluded that while both strategies have their strengths, the “average” mutual fund either met or exceeded its category benchmark in 54 out of 63 Morningstar categories over the last 3 years. Though not explicitly stated, this study suggests active investment management is a superior strategy. These results are contrary to informed thinking on the topic which suggests that investment management skill is a precious commodity which is not widely available. As William Sharpe put it, “properly measured, the average actively managed dollar must under-perform the average passively managed dollar, net of costs”. Curious about the difference, we conducted a similar study on equity mutual funds in an effort to reconcile these two perspectives.

To measure the benefit of active management many investment managers and investors subtract the return of their benchmark from the return of their fund and consider the remainder “alpha”. However, this calculation does not account for the level of market risk implicit in the fund which may be different from the benchmark. Beta and Alpha were introduced to quantify the nature of the investment market risk / return relationship. Every fund return can be decomposed into an element of Beta and Alpha. Beta is a measure of a fund’s sensitivity to market movements and hence its exposure to market risk. Alpha is a measure of return in excess of the returns implied by the market risk or Beta of a fund. Alpha is often interpreted as the value added by the active manager. Another way to think about this is that a Fund’s Beta represents market exposures which are inexpensively available via indexing while a fund’s Alpha represents unique value which is deserving of an investment management fee. In our study we used Alpha as the metric of active management.

It is critical to utilize proper benchmarks in evaluating active management. Morningstar style boxes and the standard Russell market index benchmarks don’t always reflect the investment universes or active management strategies employed by funds classified or benchmarked to them. Therefore, in addition to using standard Russell market indices, we used style based returns methodology to develop customized style benchmarks for each fund in the study. Generally, the resulting higher category R-squareds generated by the style benchmarks over the market benchmarks indicate they provide a more reliable alpha estimate.

In past studies of the average actively managed fund, the “average” was variously expressed as; a median return, a mean return or an asset weighted return. We computed a mean average as well as an asset weighted category alpha to highlight any differences in these “averages”.

Measuring fund performance against conventional benchmarks appears to systematically overstate the case for active management. Fund alpha computed using more relevant style benchmarks is generally lower than alpha computed using the conventional Russell market benchmarks under both average and asset weighted measures. This also suggest that investors ‘on average’ may be paying active management fees for what is in part beta exposure outside of the conventional asset class benchmarks.

Smaller cap and growth funds may provide better opportunities for Alpha.This study supports the notion that less efficient asset classes such as small cap equity may hold more promise for the average active manager. Similar studies suggest that growth style investing may provide more active management opportunity than value investing. Our data seems to support this general observation.

Measuring alpha on an asset weighted basis provides a more encouraging perspective on active management than on a mean average basis. The results of the two “average” methodologies are quite different. This difference in how empirical evidence supports the superiority of active vs. passive may account for the biases held on either side of the debate. The study we mentioned earlier used an asset weighted average. It is intuitive that higher alpha funds should and apparently do attract more assets over time than lower alpha funds. It also unfortunately follows that alpha tends to deteriorate as asset size increases. At this level, the debate seems to turn on which measure best represents the average active manager, a mathematical mean or an asset weighted average.

Our intuition is that the average alpha values in the study better reflect the challenges and the opportunities available to the “average” investor seeking active management outperformance.

Through this work we can see that the outcome of historical reviews of fund results and the resulting connotations about the potential for active management are sensitive to the benchmarks utilized and the mathematical definition of the “average” active manager. The next question, of course, is whether any average or specific past performance can be useful for investment decision making? For many reasons, the sole utilization of past returns in selecting investments is not as valuable as its broad use would suggest. However, past performance in concert with other quantitative and qualitative information can be helpful in narrowing an investor’s opportunity set to a comparatively small set of actively managed funds with the potential to outperform an appropriate market index.