Infinity Q: Too Much Alpha
The suspension of redemptions and planned liquidation of the Infinity Q Diversified Alpha fund (IQDNX, IQDAX) – a $1.8 billion hedge fund-like multi-strategy liquid alternatives mutual fund that was started by investment staff from the family office of a private equity titan – has sent shockwaves through the fund management industry. Using MPI's quantitative surveillance framework we discover a slew of red flags that could have alerted the fund's investors.
Volatility and Shockwaves
The suspension of redemptions and planned liquidation of the Infinity Q Diversified Alpha fund (IQDNX, IQDAX) – a $1.8 billion hedge fund-like multi-strategy liquid alternatives mutual fund that was started by investment staff from the family office of a private equity titan – has sent shockwaves through the fund management industry. It is the latest episode to have investors checking their assumptions about the behavior of their alternatives managers, reviewing plans to increase allocations to alts as fixed income allocations have taken losses amid rates’ recent liftoff and generally wondering if their portfolios need to be simplified and return expectations trimmed.
On February 22nd, the SEC announced that James Velissaris – the founder and CIO of Infinity Q Capital Management, which was funded by and spun out of Wildcat Capital Management, the family office belonging to David Bonderman, co-founder of top five private equity firm TPG Capital – had been “adjusting certain parameters” of the third-party pricing models that value the swaps and other derivatives reportedly comprising 18% of the Infinity Q Diversified Alpha fund portfolio, a strategy that was “launched to offer the investment strategies managed by Wildcat to external investors”.
Infinity Q confirmed Velissaris had accessed and altered the models and that “it was unable to conclude that these adjustments were reasonable, and, further, that it was unable to verify that the values it had previously determined for the Swaps were reflective of fair value”.
While shocking, warning signs had been in the air. In December, according to Bloomberg News, shareholders were sent notice that the Infinity Q Diversified Alpha fund would hard close the fund to new money at the end of 2020. While soft closes are common for funds mindful of capacity issues, hard closures are rare.
Additionally, Jeffrey Ptak, Head of Global Manager Research at Morningstar, found a pattern of tardiness in Diversified Alpha’s annual report filings due to audit delays. Their independent auditor BDO had this to say in 2015: “While performing our audit, we became aware that the Fund did not have adequate controls over trade allocations and trade corrections, including procedures to notify the Fund’s administrator when trade errors occurred. We believe this condition is a material weakness in internal control over financial reporting.”
An investigation is underway and the findings should ultimately tell for how long management had been “adjusting certain parameters” of the external pricing provider’s models to determine the value of Tier 3 swap assets. These swaps reportedly accounted for 18% of the portfolio so it remains to be seen what impact it had on returns, as well as the extent of losses that Infinity Q Diversified investors will face.
There is a lot yet to emerge and much that is unknown at present. But the episode inspired us to take a look at the Diversified Alpha Fund.
Diversified Alpha Track Record
Below is Diversified Alpha’s performance track record excerpted from Infinity Q Diversified Alpha’s pitch book from September 2020. The Sharpe Ratio, which measures return per unit of risk, would have shown the fund to be superior to either the Credit Suisse Hedge Fund Index or the Credit Suisse Global Macro index as its returns were significantly higher than those indices yet it took similar risk. This is an extremely attractive quality to have in a product such as this.
In the spring of 2020, when stocks and hedge funds around the world were experiencing significant declines as the true scope of the COVID-19 pandemic was being realized, the Infinity Q Diversified Alpha fund was doing the exact opposite, having strong positive returns. Further investigating the fund’s performance, it was also positive at the end of 2018 when US stocks were nearing a bear market, approaching 20% off of their peak. This behavior is also extremely desirable in a portfolio as it serves to offset stock losses in a portfolio thereby reducing overall portfolio risk.
Other Desirable Characteristics
The pitch book goes on to check all of the other important boxes for things that investors are looking for such as having a low correlation and beta with the S&P 500 of -0.12 and -0.06 respectively. In discussing risk, the pitch book shows risk management underlying all of their investment process with risk allocated to various strategies in a precise manner to a tenth of a percent. An Institutional Investor article cites a potential investor in the fund claiming that representatives of Infinity Q indicated that “they would constrain losses to 2 percent of net asset value.”
A diligent fiduciary could have looked to one of Infinity Q Diversified Alpha’s SEC filed list of quarterly holdings from which this schedule of assets, liabilities and other financial instruments are listed according to whether they are categorized as level 1, 2 or 3 assets.
A quick summary of what these levels indicate as described in the SEC filing:
- Level 1 Quoted prices in active markets for identical securities.
- Level 2 Observable inputs other than quoted prices included in level 1 that are observable for the asset or liability either directly or indirectly.
- Level 3 Significant unobservable inputs, including the Fund’s own assumptions in determining fair value of investments.
The fund shows significant level 3 assets but an investor could have easily written them off as they are much smaller in magnitude than the assets on their balance sheet. But this may belie the true risk these level 3 assets represent because they are options and swaps as opposed to stocks and bonds so their payoff can be dramatically different than stocks and bonds. Further, because these level 3 assets are often valued using a quantitative model, the inputs to the models are subjective and therefore one can significantly alter the valuation by changing model inputs. What we have is a balance sheet with extensive positions in options and swaps for which their risk is inherently difficult to measure and whose valuations are subjective.
Swaps, Options and Futures
If one dug into the holdings of Infinity Q Diversified Alpha, this is what they would have found for variance swaps but with two additional pages detailing all of their variance swap positions. Looking into dispersion swap contracts and correlation swap contracts would yield similar long lists. When it comes to futures, one would have found contracts for crude oil, soybeans, corn, cattle, wheat, coffee, etc.
Even with this detailed holdings disclosure for the fund’s hundreds of positions, it would have been extremely time consuming and challenging to both determine how much risk the fund took historically to achieve the returns they did and how much risk the fund was taking at present. Further, holdings change regularly so this kind of in depth analysis would have been required on an ongoing basis to monitor the fund and its risk. This kind of ongoing supervision would have been both costly and time consuming.
A Factor Approach
It is possible, however, to step away from the long list of Infinity Q Diversified Alpha’s holdings and instead take a top-down view of the fund using returns-based analysis. Over the past 30 years, investors have been relying on returns-based factor or style analysis to reconcile performance with information provided by fund companies, to detect hidden or unwanted exposures and to assess manager skill or the lack thereof. Quantitative analysis is one of the three pillars of fund due diligence, along with holdings-based and qualitative analysis.
This top-down approach seeks to identify a combination of known risk factors that best explain the returns of a fund. This method also identifies the portion of returns that are not explained by exposure to the systematic factors, which are often referred to as selection returns.
Style Analysis of Alternatives
Applying a quantitative factor approach to a multi-strategy alternative mutual fund like Infinity Q Diversified Alpha fund is more complicated than for a traditional mutual fund which owns only stocks and bonds. The fund implemented a combination of hedge fund-like strategies, such as equity long/short, global macro, managed futures and volatilities across global asset classes.
Complex hedge fund strategies, however, have been successfully replicated with factor models both in academic research (such as Fischer et. al. (RFS 2016)) and in the investment industry, with hedge fund tracking indices. To analyze Diversified Alpha, we adopted a similar approach with common risk and style-based factors used in Fischer, Hanauer and Heigermoser’s paper and add additional volatility factors in an effort to capture the fund’s volatility strategy exposures.
Dynamic Style Analysis
To analyze the Infinity Q Diversified Alpha fund with a quantitative factor approach, we utilized our proprietary Dynamic Style Analysis (DSA) model, which is available in the MPI Stylus Pro research platform. DSA uses machine learning to identify similarities in data patterns between the fund’s monthly returns and those of possible factor/strategy exposures and selects factors that maximize predictive power. This dynamic approach also has the added benefit of being able to identify style shifts more quickly than traditional regression analysis, as well as handling leverage and derivatives in a more precise way.
Given a set of factors that we selected for the analysis, the chart below depicts the factors our DSA model estimates the fund had exposure to historically. We call this a Style Portfolio or Factor-Tracking Portfolio.
In the context of the Diversified Alpha fund, all of the hundreds of positions reported in their quarterly SEC holdings report can be summarized into this handful of dynamic factor exposures. As depicted in the factor exposures chart, there is a material change in fund’s factor exposures over time. The application of quantitative factor analysis in the due diligence process is contingent on being able to precisely detect changes in factor exposures over time, as well as the ability to capture implied leverage and short exposure. Because traditional returns-based style analysis methods often fall short in providing accurate and timely information on complex investment product like this, a more advanced model is critical.
In order to quantify how well this factor model explained the Infinity Q Diversified Alpha fund’s return behavior, two metrics were calculated. The first metric is the style R-squared (R2) which quantifies how much of the fund’s variance was explained by the factor model. The next metric is the predicted style R-Squared (PR2), which, as a predictive analytic, describes how well the factor model was able to predict what occurred in each subsequent period. The difference between the predicted return (using the factor exposures from each prior period) and the actual return is quantified with the PR2 metric. The phrases “in sample” (R2) versus “out of sample” (PR2) are generally used to make this comparison. MPI’s PR2 computation technique is borrowed from machine learning and is often used as part of an optimization to select the subset of factors that will maximize predictability. While the model captured 77.9% of the fund’s return variability (R-squared), the Predicted R-squared of 41.4% is considered on the low side for a traditional mutual fund. However, in the case of Diversified Alpha – a multi-strategy alternatives mutual fund that incorporates hedge fund type strategies, a lower PR2 number is not unusual given the complexity of the strategy.
Fund Versus Model
Using this factor model, we plot the return of the fund versus the return of the style portfolio or factor-tracking portfolio. We see that the style portfolio is generally directionally correct in terms of upticks and downticks (due to the solid explanatory power of our model or R-squared) but the Infinity Q Diversified Alpha fund appears to outperform the portfolio based on systematic factor exposures by a wide margin.
Selection, or alpha, accounts for the difference between the fund’s return and the return of the model. While investors are typically attracted to high selection, any statistically significant alpha, positive or negative, should be carefully investigated. For example, a significant negative selection could signal hidden fees or expenses, while significant positive selection could mean a missing factor, NAV manipulation (as was the case with the infamous Manhattan Fund) or legitimate manager skill.
In the case of Infinity Q Diversified Alpha, our DSA analysis suggests that the majority of the fund’s returns over the past 6+ years remain unexplained, quantitatively speaking.
The factor-tracking portfolio fluctuates around zero while the fund keeps generating massive positive returns, albeit in the same pattern as the modeled factor-tracking portfolio. Although it may be perfectly legitimate, we consider it to be a potential red flag when a fund has such large and consistent positive alpha, especially in comparison to its peers.
Factor Exposures and Attribution Analysis
We see that the three most persistent long exposures in the fund were equity momentum, emerging equity, and VIX Mid Term while the most persistent short exposures were equity value, term spread and commodities. Over the life of the fund, the biggest contributors to positive returns according to our analysis were the fund’s short exposure in equity value which contributed 7.53% to total returns and its long position in equity momentum which contributed 5.01% to total returns. The biggest negative contributors to returns were the fund’s short exposure in term spread which contributed -6.75% to returns and the short position in commodities which contributed -4.12% to returns.
The anomaly in this attribution chart is that the fund returned 38.80% (5.3% annually) yet the factor exposures contributed very little to the total return. The fact that the fund used many non-linear return factors, including options and swaps among others, certainly contributed to some of the unexplained returns. Even with these non-linear factors, the selection, or unexplained returns, comprised 36.63% (5% annually) of the fund’s total return, almost all of the fund’s returns.
We can zero in on the fund’s apparent skill by plotting the monthly returns of the fund vs. returns unexplained by the style or factor portfolio. We see that the selection return was quite uniform and it was infrequent that there were subsequent months with negative selection since 2018 (yellow shaded area in chart below).
Cumulative selection return (Green line in the chart below) representing the difference between the fund’s performance and that of its factor model (Red and Blue lines in Cumulative Performance chart above) is climbing steadily month-after-month. If one has a high degree of confidence in their factor selection and the robustness of their quantitative analysis, such persistent positive selection return could call into question the reporting of the fund or other concerns. When checking the holdings of the fund, the presence of significant level 3 assets could have provided management with an opportunity to apply some degree of control over how these assets were valued which could potentially be a source of these persistent positive selection returns.
In its fund surveillance applications, the MPI Stylus platform facilitates ongoing quantitative monitoring of large numbers of funds to provide advance warnings and identify areas of concern. Traditional mutual fund investors are usually concerned with tracking error to benchmarks, overall risk, style drift, etc. The red flags that are most relevant for hedge funds and alternative mutual funds surveillance are often focused on the following areas (among others):
- Implied Leverage (as estimated by the model) and changes in Implied Leverage
- Predictive Power of the model and change in Predictive Power (e.g., if a model that worked suddenly stopped working)
- Unexplained Performance and magnitude and Consistency of Unexplained Return
Quantitatively comparing a fund to its peers is an important step of both due diligence for prospective investors and ongoing fund monitoring for investors. So as the next step, we applied our standard surveillance screen that included some of the flags mentioned above to Morningstar’s Multialternatives category, the mutual fund group that Infinity Q Diversified Alpha was assigned by the fund data provider.
An Infinity of Quantitative Flags
The chart below shows Diversified Alpha’s peer funds in the Multialternatives category and plots the number of red flags (X-axis) that each fund raised in comparison to the respective fund’s AUM (Y-axis). Infinity Q Diversified Alpha had 6 flags, the maximum, and was the largest fund triggering more than two flags. In fact, the fund was the 6th largest fund in the Multialternatives category, a testament to fundraising prowess, likely a function of the attractiveness of its stated returns through periods of equity market volatility and drawdowns since inception. Across the Multialternatives peer group, we note that the highest concentration of flags was in the predictive power of the factor model and unexplained performance buckets. Again, a fund that triggers 6 quantitative red flags indicates that one should do additional due diligence and investigation to better understand such behavior but, by itself, does not indicate any malfeasance on the part of the manager.
How Much of an Outlier?
We further examined the fund and found that both the unexplained performance and its significance were by far the largest in its peer group while the explanatory power of the factor model at 80% was on-par with its peer group’s average. In the chart below we show each fund in the category plotted with the selection return unexplained by the model (Y-axis) versus its t-statistic (X-axis), an indicator of selection significance. Higher t-stats indicate the consistency of monthly return that doesn’t come from factor exposures but rather from the manager’s perceived ability to select securities (the unexplained portion).
One doesn’t need to look at statistical tables to identify the big outlier in the chart. The funds with significant negative selection, located in the bottom left quadrant with t-stats less than -2, are likely affected by high management fees, assuming that they also have solid R2 values. But there is only one outlier with very significant positive selection, indicated by a t-stat of 3.6. This means that 5% of the fund’s 5.3% total annual return were consistently unexplained by the factor model. Yet this same factor model did a good job explaining the month-to-month movements of the fund, even its exceptional performance during the COVID crisis.
Infinity Q Diversified Alpha’s stated historical performance was quite spectacular. It gave investors much of what they are looking for in an alternatives allocation: positive returns with little volatility in calm markets, little to no correlation with the core asset classes of equities and bonds, and strongly positive performance during periods of market distress and drawdown. Infinity Q Diversified Alpha came with the pedigree of a hedge fund strategy and portfolio management team both originating from the family office of a private equity titan. The mutual fund wrapper on the strategy satisfied advisors’ daily liquidity needs and lower appetite for exorbitant performance fees (though Infinity Q Diversified Alpha still charged in excess of a 2% management fee, standard for hedge funds).
The fund’s strategy was multifaceted and complex with reported holdings that were exotic, cryptic and extensive. A position-based analysis of this fund would have been very challenging for even experienced analysts given the hundreds of esoteric financial instruments including swaps, options, and futures reported to the SEC.
Using the demonstrated top-down quantitative approach with a precise model, it is possible to identify a handful of factor exposures that appear to have driven the return behavior of Infinity Q Diversified Alpha over time. Even more importantly, it is possible to isolate the selection return of the fund and see how consistently positive it was. Such rare consistency of alpha could potentially be the mark of persistent skill and selection, but typically sets off a quantitative red flag and should merit further attention. Of course, one must always view alpha in regards to the validity of the DSA model to ensure that key factors were not omitted or simply not able to be well captured. Reviewing these quantitative results would leave many advisors and analysts wondering “is this performance too good to be true?” Using a quantitative surveillance framework such as we present with our tools, one could have seen that Infinity Q Diversified Alpha was a true outlier, setting off more flags than any other peer fund of its size.
An SEC investigation is underway and time will ultimately tell if any malfeasance occurred and what impact it had on returns, as well as the extent of any losses that Infinity Q Diversified Alpha investors will face once an NAV is established. The fund has announced that it will liquidate, returning all capital to investors. Based on our quantitative analysis, the large and consistent unexplained returns have existed for some time.
Those who are currently invested in the fund know they have taken a hit but are unaware of the damage it caused yet. There are other funds out there with behavior that appears to be well explained by systematic factors but still seem to provide large and overly-consistent alpha over time. Fiduciaries and asset owners should continue to explore better quantitative solutions to quickly and precisely identify performance anomalies and hidden risks, especially in complex alternative investment categories. Such idiosyncrasies may indicate the fund’s performance may not be as stellar as it first appears.
The MPI Stylus Pro research platform can give investors in complex, multistrategy liquid alternatives and hedge funds an ability to identify these performance anomalies prior to investing, and either further investigate the fund or avoid it all together.
For liquid alternatives asset managers, their Chief Risk Officers and fund boards, Infinity Q Diversified Alpha will likely serve as another cautionary tale, perhaps alongside the infamous Third Avenue Focused Credit Fund (though very different instances). As 40 Act fund sponsors prepare to comply over the course of the next 12 months for the SEC’s 2020 Derivatives Rule – whereby funds using derivatives in excess of 10% of net assets will have to institute a standardized risk management framework and employ a derivatives risk manager who will be appointed by and reports to the fund board, the final story of Infinity Q Diversified Alpha may end up being a teachable moment for fund managers, risk officers and fund trustees.
With a consensus forming that the four decades long bond bull market may be behind us, the case of Infinity Q Diversified Alpha is particularly pertinent for advisors and investors who are looking for portfolio diversifiers and shock absorbers in a potentially new regime, leading them to increasingly look at hedge funds and alternative mutual funds as solutions to help meet return expectations.
The analysis above was performed while the Infinity Q Diversified Alpha fund was calculating the true NAV of its assets. Details have since emerged in regards to the valuation of the fund. A website has been set up to disseminate information to investors in the fund with the most recent disclosure being this update as of March 26, 2021 which stated the following:
On February 18, 2021, the last day on which the Fund calculated a net asset value (“NAV”), the Fund’s stated NAV was $1,727,194,948.50, compared to an asset value (before considering liabilities and other deductions necessary to calculate an NAV) as of March 25 of $1,249,485,022.
Remarkably, in our analysis we identified a significant portion of the fund’s return that couldn’t be explained by the factor model resulting in a wide gap between the model and the fund’s reported performance. We alluded that this gap could be the result of alleged price manipulations which precipitated the suspension of redemptions.
When we incorporated the reported downward NAV adjustment of -27.7% as the February 2021 return, the fund’s adjusted return fell precisely to the level of our factor tracking portfolio (Blue line) as shown in the chart below. This means that our factor model most likely represented the true NAVs of the fund all along before any alleged manipulations took place.
 For the past 6+ years MPI’s Hedge Fund Index unit has been using similar techniques to successfully track the Eurekahedge 50, an index of 50 top hedge funds and more recently an index of top 20 CTAs – all with liquid ETFs – thus creating an ultimate investable benchmark for liquid alternative funds. This serves as yet another proof that common ideas of hedge fund managers can be identified as dynamic betas.
 Full list of factors used in DSA Analysis:
- Broad Equity: Fama-French Excess Return on the Market
- Equity Size: Fama-French Small minus Big
- Equity Value: Fama-French High minus Low
- Equity Momentum: Fama-French Momentum
- Broad Bond: Bloomberg Barclays U.S. Aggregate Bond
- Credit Spread: Bloomberg Barclays U.S. Corporate Investment Grade – ICE BofAML US Treasury Index
- High Yield Spread: Bloomberg Barclays U.S. High Yield Corporate Bond – ICE BofAML AAA US Corporate Index
- Term Spread: ICE BofA 10+ Year US Treasury Index – ICE BofAML 1-3 Year US Treasury Index
Currency and Commodity
- Commodity: Bloomberg Commodity Index
- Gold: Bloomberg Gold SubIndex
- Currency: Deutsche Bank Long US Dollar Index
- Inflation: Bloomberg Barclays U.S. Tips – ICE BofAML US Treasury Index
- Trend following: Credit Suisse Managed Futures Liquid TR
- FX Carry: Russell Conscious Currency Carry TR
- Emerging Equity: MSCI Emerging Markets
- Volatility: CBOE Volatility Index
 DISCLAIMER: MPI conducts performance-based analyses and, beyond any public information, does not claim to know or insinuate what the actual strategy, positions or holdings of the funds discussed are, nor are we commenting on the quality or merits of the strategies. This analysis is purely returns-based and does not reflect actual holdings. 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.
 DSA tracks these time varying factor exposures with a true dynamic and distribution-assumption-free model unlike traditional rolling regression windows and Kalman filter approaches.
 Fund returns are net of 2.24% fees. This means that before-fees monthly unexplained returns are approximately 0.18% higher without any negative ones over the past two years.
 All funds’ returns are net of fees. Given that Infinity Q Diversified Alpha’s management fees were above 2%, the fund’s unexplained gross returns were 2% higher than their unexplained net returns of 5%, or 7%.