MPI Transparency Lab
A free resource from MPI providing unique insight into the world’s largest and most opaque investments
The detailed individual endowment PDF reports are free with registration
MPI Endowment Tracker
*3-year, 5-year and 10-year returns are annualized. Standard deviation estimates are based on either historical or the most recent fund exposures (obtained through dynamic analysis of annual returns) and quarterly returns of risk factors over 10Y trailing window. 70-30 Global Benchmark is a quarterly rebalanced portfolio of 70% MSCI ACWI Index and 30% Bloomberg Barclay Aggregate Bond Index.
05/02/2023: 4Q22 estimates were revised based on the preliminary 4th Qtr 2022 results of the Cambridge Associates’ private investment benchmarks.
05/09/2023: 1Q23 estimates added. We assume zero 1Q23 return for PE, VC and RE until preliminary index returns are published.
Would you like a free report on an endowment that is not in the table?
- Size
- Return
- Attribution
- Risk
- Risk/Return
- Efficiency
- FY2023
*Standard deviation estimates are based on the most recent estimated fund exposures (obtained through dynamic analysis of annual returns) and quarterly returns of risk factors over 10-year trailing estimation window.
*Standard deviation estimates are based on historical fund exposures (obtained through dynamic analysis of annual returns) and quarterly returns of risk factors over 10-year trailing window.
*Sharpe Ratio calculation is based on a) 10-year reported annual returns and b) quarterly standard deviation estimates based on historical fund exposures (obtained through dynamic analysis of annual returns) and quarterly returns of risk factors over 10Y trailing estimation window.
2023 Fiscal year return estimates are based on FY2022 historical endowment exposures (obtained through dynamic analysis of annual returns) and quarterly returns of asset classes. Please click here for more information on MPI estimation methodology. More detailed information on individual endowments can be found in MPI-360 endowment reports above. Official reported returns (when released) could be lower due to management and performance fees paid to private and alternative funds. Fees paid to recent (newer) vintages of private funds could be quite substantial. Please see MPI official release.
2023 Fiscal Year Asset Class Performance
(for fiscal year ending June)
Disclaimer
Some statistics on this page are based on exposure estimates obtained through quantitative analysis and, beyond any public information, MPI does not claim to know or insinuate what the actual strategy, positions or holdings of the funds are, nor are we commenting on the quality or merits of the strategies. Deviations between our analysis and the actual holdings and/or management decisions made by funds are expected and inherent in any quantitative analysis. MPI makes no warranties or guarantees as to the accuracy of this statistical analysis, nor does it take any responsibility for investment decisions made by any parties based on this analysis.
The detailed individual pension PDF reports are free with registration
MPI Pension Tracker
*3-year, 5-year and 10-year returns are annualized. Standard deviation estimates are based on either historical or the most recent fund exposures (obtained through dynamic analysis of annual returns) and quarterly returns of risk factors over 10-year trailing window. 60-40 Global Benchmark is a quarterly rebalanced portfolio of 60% MSCI ACWI Index and 40% Bloomberg Barclay Aggregate Bond Index.
07/11/2023: 2Q23 estimates added based on the preliminary 1st Qtr 2023 results of the Cambridge Associates’ private investment benchmarks.
Would you like a free report on a pension fund that is not in the table?
- Size
- Return
- Attribution
- Risk
- Risk/Return
- Efficiency
- Funding
- Red Flags
- FY2023
*Standard deviation estimates are based on the most recent estimated fund exposures (obtained through dynamic analysis of annual returns) and quarterly returns of risk factors over 10-year trailing estimation window.
*Standard deviation estimates are based on historical fund exposures (obtained through dynamic analysis of annual returns) and quarterly returns of risk factors over 10-year trailing window.
*Sharpe Ratio calculation is based on a) 10-year reported annual returns and b) quarterly standard deviation estimates based on historical fund exposures (obtained through dynamic analysis of annual returns) and quarterly returns of risk factors over 10-year trailing estimation window.
*Standard deviation estimates are based on the most recent estimated fund exposures (obtained through dynamic analysis of annual returns) and quarterly returns of risk factors over 10-year trailing estimation window. Funded ratio source: Public Plans Data;
Underestimation of Pension Risk
Autocorrelation in fund returns frequently leads to underestimation of market risk. In many cases such autocorrelation is caused by allocations to illiquid assets with stale NAVs such as private equity, private credit and real assets. To identify pensions with such issues we compute Durbin-Watson (DW) statistics of their annual returns (lower DW values imply more significant autocorrelation). We provide additional tests for autocorrelation in the table below for statistically inclined.
*Standard deviation estimates and Durbin-Watson statistic are based on trailing 10 years of public annual returns (through June 2022). On a 5% significance level, DW<0.9 indicates strong autocorrelation, 0.9<DW<1.3 – potential autocorrelation and DW>1.3 – no positive autocorrelation. Please see DW Statistical Tables for details. More on Durbin-Watson and liquidity. Note that quarterly returns may exhibit much higher levels of autocorrelation.
The above chart compares volatility estimates using pension reported annual returns (X-axis) with the MPI approach based on quarterly returns of a factor-model proxy portfolio (Y-axis)
Autocorrelation Statistics
(first order autocorrelation tests)
Return estimates are based on FY2022 fund exposures to market benchmarks (obtained through dynamic analysis of annual returns and rebalanced quarterly) and quarterly returns of the benchmarks. In making our estimates of quarterly returns we assume that all pensions report their private asset performance with a quarter lag (e.g., for 2QTR23 performance we use 1QTR23 PE/VC valuations).
Annual return data sources:
- Fiscal years 2001-2022. Public Plans Data. Center for Retirement Research at Boston College, MissionSquare Research Institute, National Association of State Retirement Administrators, and the Government Finance Officers Association.
- Fiscal year 2022: Pension funds’ annual reports.
Disclaimer
Some statistics on this page are based on exposure estimates obtained through quantitative analysis and, beyond any public information, MPI does not claim to know or insinuate what the actual strategy, positions or holdings of the funds are, nor are we commenting on the quality or merits of the strategies. Deviations between our analysis and the actual holdings and/or management decisions made by funds are expected and inherent in any quantitative analysis. MPI makes no warranties or guarantees as to the accuracy of this statistical analysis, nor does it take any responsibility for investment decisions made by any parties based on this analysis.
The detailed individual PDF reports are free with registration
On Demand Endowment Reports
We are constantly adding to our list of available reports. Would you like us to cover a specific endowment/pension?
Disclaimer
Some statistics on this page are based on exposure estimates obtained through quantitative analysis and, beyond any public information, MPI does not claim to know or insinuate what the actual strategy, positions or holdings of the funds are, nor are we commenting on the quality or merits of the strategies. Deviations between our analysis and the actual holdings and/or management decisions made by funds are expected and inherent in any quantitative analysis. MPI makes no warranties or guarantees as to the accuracy of this statistical analysis, nor does it take any responsibility for investment decisions made by any parties based on this analysis.
For 25 years, MPI’s clients have used our tools to gain insight into some of the largest and most sophisticated investment products. Now, for the first time, MPI is making available free, proprietary analyses and reports on specific endowments and pensions to help everyone understand most fundamental potential drivers of their investment behavior.
Driven by our MPI Stylus Pro technology, these MPI-360 transparency reports shine a light on portfolio return, risk, and effective asset-class exposures in a way unavailable anywhere else. Whether you are an institutional investor, asset manager, wealth advisor, researcher, or journalist, you can read below to learn why we believe transparency into these impactful portfolios is important and how we go about pulling back the curtains on their behavior.
And, of course, you can contact us to speak with MPI and learn why MPI is the provider of choice for anyone seeking to understand complex fund and portfolio behavior for which limited data is available.
- 1. Why Transparency?
- 2. Why MPI?
- 3. Methodology
- 4. MPI-360 Reports
- 5. FAQ
1. Why Transparency?
College and charitable endowments, pension funds, hedge funds, sovereign wealth funds…even as frustrated stakeholders call for greater transparency, the world’s largest and most important investments tend to remain shrouded in mystery. Yet the larger these funds get, the more they may not only be affected by markets, but also exert an influence on markets, both public and private. It’s fair to wonder what exactly is inside these tax-sheltered or tax-advantaged investing machines.
MPI remains committed to helping investment professionals, stakeholders, and the public form a clearer picture of these investment giants and their behavior through our endowment transparency and pension transparency reports.
Endowment Transparency
College endowments are famously opaque, providing so little information that their once-a-fiscal-year reports have become a regular autumn spectacle. But seldom do these reports contain much real information about how endowments achieve results, how they decide which risks to take, or how outsiders can independently verify their results.
Disclosures from endowments are minimal, typically consisting of annual performance figures, updates to policy portfolios and general commentary. Endowments often have highly restrictive media policies. Even many institutional consultants struggle to explain top-tier endowment results. Although much has been written on the endowment model for investing by gurus like Yale CIO David Swensen, those seeking to understand tactical allocation shifts, strategy preferences, and manager selection are given little to work with on an ongoing basis.
Pension Transparency
U.S. pension funds have been plagued by chronic underfunding, decreased disclosure, and even allegations of criminal behavior. Stakeholders don’t have the option of moving their money elsewhere. MPI believes that stakeholders in public, union, and other pensions should be able to:
- Validate pension disclosures and reconcile them with actual returns.
- Identify the risks pension funds take.
- Assess whether pension fund risk exposure has resulted in commensurate returns.
- Determine whether current levels of risk are appropriate for a fund’s stated risk tolerance.
2. Why MPI?
MPI is a leader and innovator in helping large investors understand complex portfolio dynamics and the drivers of return and risk.
That’s why government regulators, top institutional investors, and investment management organizations look to MPI as part of their fund due diligence and surveillance toolbox. And it’s why investigative reporters use us to help uncover red flags (like in our analysis of a feeder fund to Bernie Madoff’s strategies.)
The same technologies that drive our products and services were recognized with best academic paper prizes. Here are some of the unique benefits MPI brings to the industry:
- Strategy validation. Was it possible to achieve reported results using reported allocations? Can one trust the returns or allocations? Are there hidden risks?
- A common denominator for comparing opaque investments. Every endowment or pension fund is unique in how it defines asset classes and categories of alternative investments, making apples-to-apples comparisons very difficult. Our tools can standardize allocation types across endowments, allowing a meaningful picture for comparison.
- Netting exposures. Effective equity and fixed income exposures can be netted by shorting, leveraging, or the use of derivatives. Risk managers and CIOs use our factor analysis to uncover hidden exposures that truly drive returns – often unexpectedly.
- Uncovering hidden allocation trends. Harvard’s restructuring after the global financial crisis, Brown’s big bet on tech, Duke’s dive into crypto. Our tools allow a sharper picture of flows in and out of allocation categories, helping reveal which funds’ behavior matches their PR.
- Risk and Efficiency. Endowment and pension press often focuses on returns and ignores the risk and performance efficiency analysis every investment needs. Large endowments are long-horizon investors that lack the constraints of regulated investments. And underfunded pensions could be in danger of chasing returns through high-risk investments. MPI helps uncover an investment’s risk/return efficiency picture even with only a few public data points.
- Performance attribution. Endowments had no problem boasting about asset class contributions to returns in FY2021; after all, it was a record year. But in down years, managers may be less incentivized to explain their results to stakeholders. Our tools help stakeholders form year-to-year performance attribution pictures to keep a close eye on ongoing behavior.
- Meaningful investment rankings. Most public lists focus on pure performance from year to year. But MPI’s inclusion of risk and efficiency metrics helps tell more meaningful stories, shining a light, for example, on smaller endowments and pensions that are more efficient than their larger and more prominent peers.
3. Methodology
When only annual performance figures are reported, a decade’s worth of performance is represented by 10 data points. Traditional static and rolling window methods of regression analysis struggle to find credible insights from such infrequent data. MPI’s Dynamic Style Analysis (DSA), however, is uniquely able to work with such limited data.
DSA improves upon Sharpe’s original returns-based style analysis (RBSA) approach and, using factors indicative of the asset classes deployed by pensions and endowments, provides significant insights into their behavior. With our DSA, MPI is able to explain changes in an institution’s performance over time and to highlight differences across institutional portfolios using a common analytical framework.
MPI’s Dynamic Style Analysis (DSA) Methodology
- Uses publicly available annual returns and public and private market indices/factors.
- Captures changing portfolio exposures in ways other methodologies can’t.
- Cross-validates selected factors and model parameters through rigorous machine-learning algorithms.
- Computes estimated return contributions by asset both for the most recent fiscal year and the entire history.
- Computes risk measures, drawdowns, stress tests and hypothetical shocks using quarterly public proxies and the most recent factor exposures.
For a much deeper dive into the MPI approach and our proprietary methodology, see this CAIA-published paper.
4. MPI-360 Reports
The takeaways from pension and endowment transparency reports found in this Hub are not much different from the ones that MPI clients derive on a daily basis when using our tools to analyze and compare individual mutual funds, hedge funds, PE and VC funds as well as portfolios of funds.
Individual MPI fund reports…
- Uncover trends in asset allocation and exposures.
- Explain more recent and historical results and compare them to benchmarks and peers.
- Provide estimates of risks, drawdowns and efficiency
- Calculate historical stress-tests and hypothetical scenarios
For a selection of many more reports, go to MPI Endowment Tracker and Pension Tracker.
5. Frequently Asked Questions
Asset class exposures in our analysis represent what Prof. William Sharpe called an “effective mix” or what traders usually call “beta-adjusted exposures.” In a multi-manager, multi-strategy, and multi-asset portfolio, both equity and fixed income exposures can be affected (“netted”) by short positions, leverage and derivatives. In addition, overweight in high-beta, more aggressive strategies could result in a higher exposure to the market represented by a generic index such as the S&P 500. And, alternatively, overweight of low beta, value-focused strategies could cause effective equity exposure to be lower.
As a result, exposures obtained through factor analysis of returns may not match reported asset allocation but could provide more accurate information for both risk managers and CIOs as the effective exposures rather than the reported ones are, in fact, driving returns.
We create public market proxies to model private market returns. When we use these proxies for risk calculation, we’re not smoothing them and, therefore, they provide risk assessment as if private investments were marked to market instantaneously. It’s worth mentioning that risk estimates depend significantly on whether the estimation window includes the Financial Crisis of 2008-2009 when real estate in particular experienced a major correction in both public and private markets.
In a multibillion dollar portfolio with dozens of managers, individual specific risk is mostly diversified away. Therefore, it’s mostly about precision of the factor model and its ability to explain composite returns rather than individual manager skill.
By definition, “Selection” is a portion of return that is not explained by our factor model. Selection might include fees, missing or imprecise factor effects, rapid changes in allocation that cannot be captured by annual data, massive write-offs or restructuring, individual success stories such as the Coinbase investment by Duke, etc. Fortunately, multibillion dollar portfolios are not easily turned over quickly, especially when significant portions are in illiquid private investments, and this makes our analyses possible and more accurate. Nevertheless, many try to see manager selection skill behind some endowment success stories. In most cases, such success is driven by a common factor we can identify. For example, we determined overweight of technology as the primary source of the Brown endowment’s winning streak.
Indeed, annual data, the only public disclosure most institutions provide, results in very small samples. Quarterly or monthly data, if available, would allow more accurate and even more granular (more factors) results. But we squeeze the most insight possible out of available data, always with a firm eye on its validity. We use machine learning techniques, including cross-validation, to calibrate our DSA model to make sure it’s not a simple data fitting exercise. These techniques are analogues to out-of-sample replication where past data with a lag is used to project next-period returns, but much more computation intensive.
As mentioned above, asset class exposures obtained through our analyses may differ from stated allocations but, if used to replicate an institution’s results out-of-sample, would deliver very close returns.