A summary of the most recent academic research on illiquid assets presented at the Inquire Europe and Inquire UK joint seminar in Amsterdam in March 2016.
Illiquid assets – what are they?
The recent joint Inquire Europe and Inquire UK seminar brought together over a hundred investment professionals and academics for the latest from academia on illiquid assets. These are assets that are more difficult to trade and include, for example, private equity, real estate, hedge funds, infrastructure funds and loans. Illiquid assets cannot be sold quickly because of a lack of ready and willing investors to purchase them. In other words, illiquid assets cannot be readily converted into cash. For this reason, investors expect to earn an illiquidity premium when investing in illiquid assets.
The Yale Model is an early example of a institutional investor seeking to capture the illiquidity premium. The model, developed by David Swensen and Dean Takahashi in managing the Yale endowment fund, associates liquidity with lower returns and the endowment fund is characterised by a heavy exposure to less liquid asset classes such as private equity.
The Yales’ University Investments Office outlines the approach as follows:
The heavy allocation to non-traditional asset classes stems from their return potential and diversifying power. Today’s actual and target portfolios have significantly higher expected returns and lower volatility than the 1990 portfolio. Alternative assets, by their very nature, tend to be less efficiently priced than traditional marketable securities, providing an opportunity to exploit market inefficiencies through active management. The Endowment’s long time horizon is well suited to exploiting illiquid, less efficient markets such as venture capital, leveraged buyouts, oil and gas, timber, and real estate.
However, Tim Jenkinson from Saïd Business School, University of Oxford, gave evidence that private equity returns have been converging towards the returns of public equities, a longer term trend associated with the maturity of the asset class. His conclusion was true for both buyout funds, which focus of acquiring large parts of companies from parent company or from private owners, and for venture capital funds, that seed early-stage emerging growth companies.
Moreover, performance persistence for buyout private equity has largely disappeared where excess returns are now competed away and manager returns mean‐revert. This is, however, appears to still be less the case for venture capital.
However, in his view, the idea that one can just mimic private equity returns by leveraging up public companies clearly under-estimates the role general partners play in managing leverage risk. From a behavioural perspective, he sees the fact that private equity funds are illiquid as an advantage for investors since they cannot easily make panic sales after big stock market falls, saving them from realising poor performances.
In turn, Ludovic Phalippou, also from Saïd Business School, University of Oxford, suggested that returns to private equity, especially for Buyout firms and to a lesser extent for Venture Capital firms are much more volatile than the time series of consultants tend to suggest, comparing well with the performance of the small and mid-cap segments of equity public markets and much more correlated with public market returns than previously thought. He came to this finding by introducing a new methodology to estimate the historical time series of returns to private equity based on the cash contributions and distributions accruing to limited partners.
Morten Sørensen from the Copenhagen Business School found large amount of long-term persistence in private equity performance, in particular in buyouts. But because individual firm performance is noisy he also found little investable persistence, much like Time Jenkinson. It is thus difficult for investors to identify top private equity funds with the highest expected future performance.
Robert Bianchi from the Griffith University proposed a five factor asset pricing model that seems to capture well the variation of a number of return indices of listed infrastructure companies in the US and/or the Americas from the mid-90s onward. This five factor model starts with the Carhart model and adds an additional factor: the utilities industry duly orthogonalised against the other factors (correlation with other factors is removed). Robert claimed that this extended Carhart model pretty much explains what goes on in listed infrastructure indices. It is also interesting to highlight the fact he found levels of beta in range of 0.7 to 0.8, high for an asset class with a reputation for being quite defensive.
Jacob Sagi from the University of North Carolina, Kenan-Flagler Business School, derived an equilibrium illiquid asset pricing model where investors vary in how they value the income stream from real estate properties. He observed that when assets are illiquid investors only hold the asset over short periods of time if much better offers happen to come along. Owners periodically receive bids for their property from investors but gains from a trade only exist if the valuation of bidders (net of transaction costs) exceeds that of owners. Jacob used purchase and sale data from NCREIF to compute holding period price-appreciation returns for commercial properties to calibrate the model and get additional insights into the illiquidity of commercial real estate. He found that transaction prices are 11.4% lower than the average owner’s private value of the asset. This can be viewed as a model implied illiquidity discount. Jacob also looked at the discount that must be applied to an asset price in order to achieve a high probability of sale. He found that a fire sale discount of 21% relative to the expected transaction price is required in order to reach 66% probability of success in selling the property.
Equity market illiquidity
Patrick Tuijp, from the University of Amsterdam, investigated two possible ways in which equity market liquidity can dry up: a general market wide deterioration versus one that only hits the segment of the less liquid stocks. His statistical factor analysis for the US equity market shows that the two different effects are indeed present in historical data. These explain together 66% of the observed variation in illiquidity. The general component, with 57%, explains the lion share of this liquidity dry ups and is especially important in times when market returns are bad. The less dominant slope factor, i.e. the factor that deals with the less liquid segment only, is driven most by high turnover in the liquid segment. The framework can be used to classify liquidity crises. Patrick used a liquidity enhanced version of the CAPM to find that the economic impact of the general component carried a premium of about 1.5% per annum, while the slope channel did not had any significant price impact.
Allocation to illiquid assets
Raman Uppal from the EDHEC Business School discussed how investors allocate to illiquid assets as a function of their experience. Due to short history and infrequent observations, returns to illiquid assets are opaque and difficult to estimate precisely, even in the past. Moreover, illiquid assets come with substantial trading costs of up to 10% for small stocks, can be above 10% for private equity and above 3% for institutional real estate. Raman claimed that investors with experience in illiquid assets are likely to hold more precise estimates of their expected returns, modelling how inexperienced investors should allocate wealth to new illiquid assets, how they should revise their investments as they gain experience, and what pricing implications are there for both new and existing assets.
[starbox id=erik.kroon][starbox id=raul.leotedecarvalho]