Last March as the world awoke to the seriousness of the COVID-19 pandemic, asset owners woke up to mounting losses across their various portfolios. More than a decade after the Great Financial Crisis (2008), asset markets were finally experiencing a major crash. A mere five months later, the S&P 500 had recovered its prior peak. These unprecedented times imply a much wider than normal distribution of potential future outcomes. At times like these, it’s imperative to be able to regularly monitor a portfolio’s risk and factor exposures.
Johnathan Crist, CFA FRM is a Sr. Investment Analyst at the Georgia Tech Foundation, Inc., where he is a part of the team managing $2.3 billion in assets. His responsibilities include managing the portfolio allocation, risk, and derivative program.
The reflections here are the results of a several months-long collaboration between Georgia Tech Foundation and Solovis to implement a bottoms-up, seamlessly updated ex-ante risk system, which ultimately became Solovis Risk Analytics. Much of this effort occurred during the most volatile parts of the COVID-19 pandemic.
During good times, portfolio risk is often an after-thought. When asset prices are going up and market volatility is low, risk is more of a check-the-box once- or twice-a-year type process. When it’s time, institutions will dust off their old spreadsheets, update them for changes, incorporate the latest assumptions as best they can and hope for the best. It tends to be much more of a reactive process than a proactive one.
One of the reasons is because there’s a lot of inputs and assumptions that go into a robust analysis of a portfolio’s risk:
- The weights of each investment in the portfolio across all asset classes.
- The covariance matrix of all the investments.
- Returns for a robust and representative set of risk factors in order to decompose the portfolio’s risk into its various drivers.
- A methodology for modeling the variances and correlations of the illiquid assets in the portfolio.
It’s a lot to pull it together even once. Doing so regularly requires coordination across portfolio and functional silos, risk expertise, and significant amounts of automated data.
Johnathan Crist, CFA FRM is a Sr. Investment Analyst at the Georgia Tech Foundation, Inc. where he is a part of the team managing $2.3B in assets. His responsibilities include managing the portfolio allocation, risk, and derivative program. Previously he worked at the UPS Group Trust in the Strategy, Asset Allocation & Risk group. Johnathan graduated from Georgia Tech with a degree in Industrial Engineering and is currently pursuing his MBA with a concentration in Quantitative Finance at the Scheller College of Business.
In 2019, Georgia Tech Foundation (GTF), which manages a more than $2.3 billion endowment, became Solovis’ first Risk Analytics client. At that point, Solovis Risk Analytics consisted only of prototype Python code that printed calculations into an Excel spreadsheet report. But GTF was also a Solovis Portfolio Analytics client. This meant that Solovis could see GTF’s allocation to every stock and bond in its portfolio (including full look-through into the portfolios of many of its managers) throughout time. A risk system built on top of this data had the advantage of being able to generate risk calculations using the portfolio’s actual holdings down to the individual security.
A key drawback of many other risk systems is that they are returns-based rather than positions-based. The consequence of calculating factor exposures and risk using returns is that the historical time series of a manager’s returns reflect shifting styles, asset allocations, and views. Thus, risk metrics calculated using these returns capture the average positioning of the manager or portfolio over the analysis period. However, when conducting a risk analysis, the best predictor of ex-ante risk is the current positions.
Solovis set out to solve this issue for GTF, and other future clients, by building an analytical platform on top of the treasure trove of positions level data it already had access to via Portfolio Analytics.
“Solovis had been our core tool used to monitor our portfolio allocation. As a group, we always challenged ourselves to look through just asset class weights to understand our true exposure and risk. We had integrated our live allocation in Solovis to some crude Excel models to monitor our portfolio risk and equity beta. The launch of Solovis Risk Analytics brought the sophistication of what we were doing to a new level, which resulted in greater trust in the data and better overall portfolio management.”
Solovis had been prototyping risk algorithms and code in collaboration with GTF for much of 2019. An iterative approach was used where features such as scenario analysis and risk proxies for illiquid assets were developed and then beta-tested with GTF. But then as a front-end for its risk application was still under development, the COVID-19 crisis hit. GTF didn’t have time to wait for the final product – it needed risk answers right then.
For those stressful several weeks, Solovis delivered GTF regular risk reports using its already developed risk engine code and Excel spreadsheets. These reports included up-to-date information on its portfolio’s volatility, factor exposures, major risk contributors, and the projected impact of further economic shocks.
“No one is ever truly 100% prepared for the type of event that occurred in February and March of 2020,” Crist said. “This was especially true as we were not quite at a final stage of having the risk platform up and running. Regardless of how far along the project was, the data was vital to the circumstances. It was crucial to understand how our portfolio was positioned, both as the market sold off and in the subsequent rebound.”
This trial by fire of both GTF’s risk management process and Solovis’ new risk software helped GTF make the moves it needed to emerge unscathed from the market volatility. And it turned out to be the best proof of concept Solovis Risk Analytics could have asked for. Over this period, Solovis proved that it could produce insightful risk analyses on an on-demand basis. And these analyses could then be used to help inform the investment process during the worst and most stressful of times.
In the fall of 2020, Solovis completed the front-end of the new Risk Analytics application. Fully integrated with Solovis Portfolio Analytics and accessible on-demand, Georgia Tech Foundation could finally move off spreadsheets for good.
“The events in February and March of 2020 ended up pushing the project along quicker. It provided great fact checking for the parameters and proxies we were assigning to different parts of our portfolio. We were going through a live stress scenario that we could use to compare to what we were seeing in the hypothetical scenarios in Solovis Risk Analytics. Calibrating the model to the true output occurring in markets allowed us to ultimately gain confidence in the data going forward.”
Today: Close Collaboration
Today, GTF remains a happy Solovis Risk Analytics client. Every week, GTF uses Solovis Risk Analytics to monitor its risk and portfolio exposures. Weekly conference calls with Solovis’ R&D team (the team that built the Risk Analytics application) are used to stay on top of things from both an investment risk and modeling perspective.
“Like the relationship we have with Solovis on the Portfolio Analytics side, the Risk Analytics team has become an extension of our staff,” said Crist. “Our weekly calls go beyond just updates on the application, but they cover all topics around risk management from best practices, to future implementation considerations and topical current events. Very rarely do you encounter a team as open to conversation and suggestions as the Risk Analytics team at Solovis.”
One thing in particular that Georgia Tech Foundation likes about its partnership with Solovis is that it extends beyond a simple software vendor-client relationship. To GTF, Solovis is more research partner than vendor. And to Solovis, GTF is an ideation partner and sounding board.
For example, during the fall of 2020, when GTF was interested in better understanding its exposure to nonlinear equity volatility, Solovis’ R&D team helped develop several new volatility-specific factors that are currently used to augment GTF’s base risk factor model.
To learn more about Solovis portfolio analytics and risk capabilities, visit www.solovis.com.