Replicating the returns of private equity and venture capital, two murky asset classes, is possible. Rather than investing in private companies directly, there are indexes that seek to replicate the industries' returns by constructing portfolios of liquid, U.S. large-cap equities. These portfolios are designed to mirror the characteristics and returns of the private markets.
The first step in creating a replication strategy is to create baseline research indexes. The indexes must first contain detailed transaction and valuation data on the universe of U.S. private equity- and venture capital-backed companies. Individual companies are aggregated to provide an industry level valuation of all U.S. private equity and venture capital companies (effectively creating industry benchmarks). This approach allows research indexes to reflect the gross returns of the industry, i.e., before the respective fund manager fees and expenses are deducted. Also, the research indexes are market-cap weighted.
Building tracking portfolios
Armed with a detailed historical profile of the private equity/venture capital industry return characteristics established through the research indexes, tracking portfolios are then created to mirror the risk and return characteristics of the respective asset classes. They are designed to track their respective research indexes in the medium-term due to reporting delays and other noise inherent to private company data. Tracking portfolios are constructed using liquid, U.S. large-cap equities. Modest leverage is commonly used so the tracking portfolios' risk loadings match those of the private equity and venture capital industries in aggregate.
While private equity and venture capital returns are not, themselves, predictable, the risk exposures of the asset classes can be predicted, and it is possible to create portfolios that mimic these exposures. A consensus forecasting approach averages the desired portfolio weightings generated by a portfolio of non-linear predictive models. Each model aims to replicate the risks and return characteristics of the private equity or venture capital research index. Models are designed to minimize the overlap of information.
It is critical to note that since the research indexes are highly diversified, idiosyncratic risks at the individual company level are neutralized, leaving only systematic risk.
The underlying economic relations between predictive variables and the risks within resulting portfolios are selected because they robustly predict exposures of underlying private equity and venture capital firms in different market environments. The models are designed to consistently track underlying exposures using modest levels of dynamic portfolio leverage and inter-sector leverage adjustments. Averaging the portfolio recommendations across models reduces model risks and individual models are prevented from taking extreme positions. Portfolio parameters are estimated in-sample and evaluated using out-of-sample verification data.
3 main steps to construct a replication portfolio
1. Sector distribution matching. Each quarter, the market-cap sector weightings of the research indexes are matched using individual stocks from publicly traded sector indexes. Intra-quarter sector weightings self-adjust based on relative market performance.
2. Portfolio level beta matching. On a monthly basis, the portfolio risk characteristics of the research indexes are estimated using proprietary models. The overall risk level is replicated in the tracking indexes through modest leverage at the portfolio level. The leverage factor is reset monthly and can go below one.
3. Sector-level beta refinement. On a monthly basis, the same proprietary models are used to estimate the risk characteristics at the individual equity sector levels. Leverage adjustments, up or down, are made to each sector to better reflect risk estimates. This action tightens portfolio tracking, and the net leverage effect of adjustments across all sectors is zero.
The resulting tracker index portfolios are composed of highly liquid U.S. equities, allowing the portfolios to trade efficiently and be scalable.
Arthur R. Bushonville is founder and CEO and Jeffrey F. Knupp is president of DSC Quantitative Group, Chicago. This content represents the views of the authors. It was submitted and edited under P&I guidelines, but is not a product of P&I's editorial team.