Risk factors help to identify the fundamental value drivers of real assets and explain differences in the reported returns of public and private equity investments that hold substantially similar assets. By combining the fundamentals of real asset valuations with the statistical tools required to unlock the component risk factors of asset classes, it is possible to replicate the returns of private real asset investments using liquid publicly traded instruments. We believe risk factor replication is a more effective funding mechanism for private equity capital calls. It is intended to capture the underlying value drivers of private market investments during the typical commitment period while reducing liquidity risk.
Real assets, also referred to as “hard” or “tangible” assets, remain one of the fastest-growing asset classes across institutional portfolios as investors look to mitigate inflation risk, seek stable cash flows and achieve additional diversification in their portfolios.
Investments in real assets generally take the form of direct or indirect exposure to physical assets. According to research firm Preqin, commitments to real asset private equity funds have exceeded $670 billion in aggregate from 2003 to 2012 with over $200 billion yet to be called. Some of the larger pension and sovereign wealth funds also engage in direct investments and joint ventures with operating companies. In either case, a time lag often exists between when the capital allocation decision is made and the actual deployment of capital. Under a typical four-year commitment period, for example, even an institution with a mature private markets portfolio may have the equivalent of one-quarter of its real assets allocation in unfunded capital commitments. During that time, prices, asset valuations and relative values across sectors change, while investors remain underexposed to the value drivers – the risk factors – of the real asset strategies relative to their desired targeted asset allocation.
Up until now, investors had three basic choices in managing the unfunded capital commitments in real asset portfolios: 1) reserving in cash; 2) reserving across the liquid asset classes, such as equities and bonds; and 3) using a just-in-time approach where private equity realizations are expected to fund capital calls. Each of these choices has flaws. To fully reserve these commitments in low-risk cash and liquid securities would create a significant cash drag on the portfolio. Funding from equities and fixed income may result in unintended fluctuations in the portfolio’s allocation across assets classes and unstable risk factor exposures.
Finally, as many endowments learned in 2008, the timing of private equity capital flows can be very difficult to estimate, making the “just-in-time” approach a capital budgeting nightmare during extreme market conditions. Some investors have set aside a separate liquidity pool to fund illiquid capital commitments, but few have considered the potential tracking error of these reserves to their intended portfolio.
We believe there is a better solution that addresses the issue from a portfolio risk management perspective and that may allow investors to capture the factor exposures of real assets while they wait for capital calls to fund their real asset investments. The solution is to create a liquid portfolio of investments that matches the risk factor exposure of the private real asset investments. The composition of the portfolio should be based on 1) a fundamental economic analysis of the real asset investments that identifies the appropriate risk factors and 2) an econometric analysis that estimates the appropriate loadings on each of these factors.
The replication portfolio
The keys to real asset replication are to isolate the risk factors that drive investment returns of the illiquid portfolio and to find suitable liquid proxies that match those risk factors. Several fundamental tenets underpin this strategy:
- Public and private investments in fundamentally identical assets should have similar returns over an investment cycle. A substantial fraction of the differences between public and reported private index returns can be explained by differences in valuation methodologies and appraisal smoothing as opposed to actual disparities in investment performance.
- Portfolio analytics can be effective in isolating risk factors of asset classes when used in concert with fundamental valuation models. Real asset returns can largely be explained through exposures to market risk factors such as equity beta, real and nominal duration, credit spreads, currency, commodity and industry factors.
- Liquid proxies intended to match the risk factors of illiquid investments should be designed to track the reported returns of illiquid investments over an investment cycle but will likely demonstrate tracking errors in the short term due to appraisal smoothing and differences in the underlying portfolio vs. the index proxy.
- A risk factor replication portfolio seeks to capture only the beta of the investment return of the target portfolio. It will not capture the various sources of liquidity premium and alpha that may be associated with investments in private real asset markets. Investors with slightly higher volatility tolerance may be able to achieve similar expected returns of an illiquid real asset portfolio through leverage of the underlying risk factor replication proxies.
The effective implementation of a risk factor replication strategy requires a fundamental understanding of the return drivers of privately held real assets, a rigorous econometric framework and the capability to cost-effectively implement the replicating portfolio at the portfolio management level.
Analysis of real asset returns
By definition, the total return to any asset or investment has two components: income and capital appreciation. Capital appreciation in a given period can be decomposed into two components: earnings growth, and changes in multiple expansion (duration multiplied by the capitalization rate, i.e., the unlevered discount rate). The formula below illustrates these components of total investment returns:
Each real asset is exposed to changes in specific factors that affect price, volume and operating costs for those assets. For some real assets, earnings dynamics are driven mostly by price and input cost changes, while other real assets exhibit greater variability in productive volume. Volume risk tends to be more relevant in exploration and production assets and less so for fixed asset investments involving real estate, infrastructure, farmland and timber, among others. The return dynamics for the most stable cash-flow-generating assets will tend to be dominated by changes in discount factors (“capitalization rates”).
In our replication approach we include liquid investments whose returns are linked to the same factors that drive earnings potential in each of the private asset classes, as well as investments that capture the factors that affect discount rates of private asset classes.
As an illustration of this approach, Figure 1 shows estimated market risk factor exposures for U.S. farmland as well as the risk contribution from each factor. Our analysis indicates that farmland returns are predominantly driven by a couple of market risk factors, namely, commodity prices and real interest rates. Specifically, we estimate the empirical real duration of farmland is approximately 7.2 years (accounting for approximately 50% of return volatility), whereas farmland exposure to commodity prices, a combination of corn and soybean prices, is 0.4 (accounting for roughly 40% of the return volatility).
Consequently, the portfolio of liquid investments that, in our belief, best replicates these underlying factor exposures of farmland mostly consists of commodity futures and U.S. TIPS (Treasury Inflation-Protected Securities). Treasuries are also included to further improve the liquidity profile of the replicating portfolio. From a fundamental valuation perspective, the commodity futures exposures are a proxy for the earnings growth component of returns whereas the TIPS and Treasury exposures capture capitalization rate exposures.
Risk factor replication methodology
For a complete real asset portfolio, the appropriate liquid replication portfolio is determined in a three-step process.
First, the client defines the targeted portfolio composition of real asset classes. This is often based on a directive from a client’s investment committee or derived directly from their investment policy statement.
The second step consists of a formal econometric analysis of the targeted portfolio’s risk factors, which is validated by a fundamental analysis as described above. The asset class factor mappings are based on the private market indexes where available or based on a close liquid market proxy for asset classes without a recognized private market benchmark index. In some cases we are also able to test our results against a customized private market benchmark using transaction- or portfolio-level performance data.
Together, these two approaches help determine the component risk factors for each sector. When aggregated, these risk factors form the portfolio-level risk factor matrix that we look to replicate.
Finally, we identify the set of investable liquid proxies with similar risk factor profiles, and optimize their composition to bring the liquid replication portfolio into proper alignment with the illiquid target portfolio, while ensuring the most efficient implementation from a portfolio management perspective.
The three steps involved in our risk factor mapping and replication approach are summarized in Figure 3 for a diversified real asset portfolio that includes real estate, infrastructure, energy, metals and mining, farmland and timber.
Conclusion
Risk-factor-based portfolio analytics, when combined with a thorough understanding of asset valuation and operating fundamentals, can be effective in isolating the fundamental value drivers of real assets and in reconciling the differences in the reported returns of public and private investments. These tools form the building blocks in a risk-factor-based replication strategy for illiquid asset classes.
We believe a risk factor replication portfolio may be an effective means of capturing the intended market factor exposures of an illiquid real asset portfolio. Investors should consider evaluating and quantifying their liquidity risk and developing a dedicated funding mechanism for their illiquid capital commitments. Risk factor replication should be considered as a risk management solution for these liabilities.
The authors would like to thank Bob Greer, Bruce Brittain and Ashish Tiwari for their comments and guidance, and Fei He whose analytical support was instrumental in both the development and proof of the investment thesis.