Time for a new model?
What if the data available today, combined with advanced computing power, could help limited partners manage their private equity portfolios more easily? That was the challenge a group of researchers set for themselves in a bid to help investors understand the asset class’s inherently unpredictable cash flows. Here, one of the researchers outlines the new approach, while a seasoned investor offers his response.
Meet the researcher and investor
 |
Alex Billias is chief operating officer at Bella Private Markets, where he has worked on a wide variety of research and strategic consulting projects. He also leads the development of quantitative tools for simulating portfolio cash flows and benchmarking performance. Prior to Bella, he conducted physics research for the ISOLDE facility at CERN in Geneva. Alex Billias |
|
| . | ||
 |
Patrick Sherwood is a principal at GroveStreet, focusing on fund and direct investment activities across sectors. He is also involved in analysing and managing client portfolios. Before GroveStreet, he was managing director of investments at the Wallace Foundation, having previously worked for the Yale Investments Office. Patrick Sherwood |
In Takahashi-Alexander Revisited: Modelling Private Equity Portfolio Outcomes Using Historical Simulations, a group of researchers at Bella Private Markets (Dawson Beutler, Alex Billias, Sam Holt and TzuHwan Seet) led by Harvard Business School professor Josh Lerner, outline a
new model that uses simulations to help LPs forecast cash flows and valuations in their PE portfolios.
The new model preserves the simplicity and intuition of the Takahashi-Alexander model developed in 2001, but addresses what the authors say are its limitations – that it requires users to create and input assumptions, and that it provides only a single estimate for a period’s expected capital calls, distributions, and net asset values.
The new model uses information from an existing portfolio and matches its funds to historical funds across a range of time periods to create thousands of simulated portfolios that structurally mirror the existing one. It matches funds according to criteria such as vintage year, geography, fund strategy and size. The model then weights and rescales the sampled funds’ cash flows and valuations to match the exposures in the real portfolio.
By running simulations across historical periods and funds, the model offers a range of performance, cash flow and NAV forecasts for the real portfolio. The authors say this approach allows the model to be customised in order to understand the effect of market conditions, for instance by restricting the model to crisis periods such as the dotcom bubble to see the impact of a similar crash on a portfolio of VC funds.