Forecasting Bridgewater All Weather Performance in November’s Bond Storm
November’s government bond sell-off resulted in one of the sharpest increases in Treasury yields in recent history and an uptick in fixed income volatility. We use Bridgewater All Weather, one of the largest hedge funds, to illustrate how to quantitative techniques could provide investors with a more dynamic understanding of the potential fund behavior intra-month using only monthly fund data.
November’s government bond sell-off resulted in one of the sharpest increases in Treasury yields in recent history and an uptick in fixed income volatility. While this may be particularly bad news for traditional fixed income funds, risk parity funds should, in theory anyway and to the extent that other asset classes have held their ground, be able to weather such a downturn without major losses.
While the returns of mutual funds employing a risk parity strategy are available on a daily basis, Bridgewater All Weather, the largest risk parity fund in terms of AUM with a reported $70 billion, typically only makes returns available on a monthly basis with a lag following the month’s end. Given the limited public availability of asset class and/or region concentration – and scant information on tactical decisions that drive risk and asset allocation – associated with typically opaque hedge fund reporting, it becomes very hard for investors in risk parity funds to reconcile ongoing market moves with realized returns. While looking at risk parity funds or a 60/40 portfolio can provide guidance, the range of returns observed may be too wide for investors seeking timely month end or intra-month insights into risk parity hedge fund performance. For example, during the month of November, risk parity mutual funds, which report performance daily, lost between -0.47% and -3.62%, while a 60/40 portfolio (60% S&P 500 Index / 40% Bloomberg Barclays US Aggregate Bond Index) experienced a 1.25% gain.
In an effort to provide investors and consultants with a more dynamic understanding of the potential behavior in the largest risk parity vehicle, we employ surveillance methods using MPI’s proprietary analytics and an award-winning technique aimed at providing estimates of daily hedge fund performance using monthly data.
Similar to our approach in a July 2013 post on Bridgewater All Weather, we used the reported time series of returns for the vehicle1 in order to estimate November’s daily returns. To achieve a forecast for the fund’s November performance, we used Dynamic Style Analysis (DSA), MPI’s proprietary factor model, coupled with analyst insight into indices2 that are relevant to risk parity strategies and Bridgewater in particular, similar to what we’ve published before. The model, the same used in the 2013 study linked to above, was trained and tested using 36 monthly returns up to October month end. We then used the most recent (October) month-end index exposures obtained by our model in order to create a synthetic replication portfolio for the following month (November)3. Using the daily return data for the indices in the synthetic portfolio and the prior month-end exposures, we estimated Bridgewater’s daily returns for each day in November up through the end of the month4.
The red line up to October month end above indicates the in-sample estimated performance of our model, based on the calculated monthly betas5, whereas the red line after that indicates the forecasted, out of sample, performance.
The daily forecast implies a November return of -1.92%. This places Bridgewater All Weather in the middle of the range of losses observed across risk parity mutual funds. In a month when segments of the fixed income market experienced losses as severe as -8.9%, and when most other markets and/or asset classes risk parity products invest in also demonstrated negative returns6, a nearly -2% loss is not only expected but is rather modest, particularly when compared with a YTD return of 8.3% (as of October 31st).
The goal of this exercise was twofold. On the one hand we are interested in the daily intra-month returns to evaluate how the fund’s performance would have evolved during one of the most volatile periods for fixed income in recent years. It is interesting to observe not just the volatility the fund experienced, but the magnitude of cumulative daily losses, something not readily available even when monthly returns become available. On the other hand, we try to provide an educated guess regarding the return of the largest risk parity fund during the last month before it becomes available. Such surveillance can serve as an advance notice for investors or for professionals that are interested in monitoring its performance and the behavior of the risk parity category more broadly.
The recent bond market rout and associated volatility have generated a lot of anxiety that may well drag on, particularly given the anticipated rate hike by the Fed, a possible slowing of the ECB’s monetary easing and the presidential transition in the world’s largest fixed income market. It is important for investors in risk parity funds to have surveillance methods and workflows to anticipate and monitor rapid market fluctuations that may impact complex risk-managed strategies, and overall portfolio performance.
- 1 As reported in the HFR database from Hedge Fund Research, Inc.: https://www.hedgefundresearch.com/hfr-database
- 2 Factors include TIPS, Government Debt, Commodities and Equities. Please contact us to obtain a detailed factor breakdown.
- 3 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 insights into 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.
- 4 MPI has no way of knowing the extent of, if any, tactical or strategic steps Bridgewater All Weather may have undertaken intra-month in response to market movements and changes in risk. In response to accusations that risk parity exacerbated market volatility in August 2015, Bridgewater stated that they do not “adjust allocations according to spurts in volatility”, according to the FT. The manager stated in a paper, “We also understand that some managers tend to sell assets when prices fall and buy them when prices rise because they believe that changes in volatility will persist, and volatility tends to rise when prices fall. We do the opposite because we want to rebalance to achieve a constant strategic asset allocation mix.” See the FT story from Sept. 15, 2015: https://www.ft.com/content/97b7da7a-5c18-11e5-9846-de406ccb37f2
- 5 Each day’s in sample return is calculated using that day’s factor returns multiplied by the most recent month’s estimated betas.
- 6 Returns across markets, USD denominated: S&P 500: 4.2%, MSCI EAFE: -1.9%, MSCI EM: -5.1%, Precious metals: -6.7%, Energy: -3.9%, Agriculture: -3.7%, Industrial metals: 9.1%, Global Inflation Bonds: -2.9%, Global EM Sovereign Bonds: -4.1%, 10 year US Treasury: -4.5%, Japan Government: -8.9%, Euro Government: -5.3%, UK Gilts: 1.9%. Inflation Bonds: -2.9%, Global EM Sovereign Bonds: -4.1%, 10 year US Treasury: -4.5%, Japan Government: -8.9%, Euro Government: -5.3%, UK Gilts: 1.9%.