Enhancing cash flow forecasting in private equity with machine learning models

| Private Equity | Perspectives
Natalia Sigrist
Partner, Private Equity

In contrast to liquid equities, private equity is characterised by scarce and low-frequency data with limited past history of observations. Consequently, while machine learning has gained widespread adoption in the analysis and forecasting of public equities returns, its application in private markets is more problematic.

Nonetheless, there exists an unprecedented demand within the private equity industry to leverage the power of machine learning. In the current market environment where liquidity is more irregular, investors are calling for improved cash flow visibility and more frequent updates to forecasts.

Being an early user of machine learning techniques in our investments, we have conducted extensive research in the realm of cash flow forecasting. Our work involved testing five machine learning models, employing two data scaling techniques, incorporating forecasts at varying frequencies and encompassing a multitude of independent economic variables.

To read the full paper written by Natalia Sigrist, Marco Perfetto and Alexandra Kovrigina, please contact: