This article is sponsored by Nuveen.
Few investors realise that traditional portfolio construction tools can be adapted to include social and environmental goals. Standard mean variance optimisation models map the most efficient portfolio for desired risk and return targets. Adding impact metrics to the mix allows investors with social and environmental objectives to create efficient portfolios that serve risk, return and impact outcomes. It helps them consider the trade-offs, if any, between these three goals across different investment types and different portfolios.
To do this, investors must be clear on the impact they are aiming to achieve and be able to measure it. Perhaps their impact goals align with the United Nations’ Sustainable Development Goals, as they do at Nuveen. Leveraging the UN’s framework, we have developed our own metric – a net impact score for specific investments and portfolios – that can be used as an input to standard mean variance optimisation modelling.
Other metrics for different impact goals can also be used. Net CO2 emissions data, for example, can be used for investors with environmental objectives such as net zero carbon targets.
Solving for impact
The impact portfolio optimisation model begins in the same way as the standard mean variance optimisation model, which is defining the investable universe.
Inputs for the traditional model are the unique investment profile for each investment opportunity, which includes expected return, return variance and its covariance with other opportunities in the universe. For the impact model, the investment opportunity set includes a broad range of strategies along the responsible investing spectrum, from conventional to more impact-focused strategies. The profile of every investment opportunity includes its impact value (in this case, the net impact score), along with the expected risk-and-return metrics of the standard model.
The solution to the optimisation problem is the efficient frontier, which describes the trade-offs between risk and return that are possible given the set of opportunities. For the traditional model, every point along the frontier is an optimal portfolio investment, maximising return for a given level of risk.
For the impact model, the solution is also an efficient frontier, but now we are optimising over three variables instead of two. Every point along the efficient impact frontier is an optimal portfolio investment that maximises return for a given level of risk and impact.
Nuveen piloted the approach on a subset of our private real asset portfolios. The solution (illustrated in Figure 1) is a set of efficient frontiers representing portfolios that maximise expected returns across the relevant range of risk levels and for every possible level of net impact in our example. Each efficient frontier is constrained by the portfolio-level net impact score and gives the set of best possible portfolios for the specified level of impact.
Frontiers on the right have the highest portfolio net impact score. Those to the left have lower scores. Moving from right to left improves risk-adjusted return, but at a certain point (beyond a portfolio average net impact score of seven), allowing for more net impact does not provide any additional return or risk benefit.
Pursuing additional net impact up to a score of seven does not sacrifice return or incur more risk – two objections often levelled at impact investing. This tells us that portfolios with certain levels of positive net impact perform just as well as, on a risk-return return basis, less impactful portfolios. Over this range of net impact, investors can make greater contributions to the UN SDGs without sacrificing a single basis point of risk-adjusted returns.
Evaluating the trade-offs
Digging into the portfolio details, investors can determine which trade-offs, if any, they are willing to make. We analysed real asset portfolios, which could invest in timberland, farmland and infrastructure, with a net impact score of nine.
As shown in Figure 2, the optimal impact portfolio which achieved the target net impact score of nine and maximised the Sharpe ratio outperformed all single-asset allocations on a risk-adjusted return basis. It also outperformed the global Sharpe-ratio-maximising portfolio in terms of impact, although the global Sharpe-ratio-maximising portfolio outperformed the optimal impact portfolio on a risk-adjusted return basis.
Achieving impact goals and financial returns
By quantifying risk-return-impact trade-offs, we are able to do two things. The first is to identify the range of portfolio impact that is achievable without affecting risk-return efficiency. The second is to support the design of portfolios that minimise the reduction in risk-return efficiency required to achieve a targeted level of impact. The important takeaway is that impact can be, and needs to be, evaluated and considered in the investment process with the same rigour and prudence as financial performance.
To find out more, read Nuveen’s research paper Optimizing for impact with private equity and real assets: A guide to integrating impact into financial analysis of alternative assets, available at Nuveen.com