Over sixty years ago, in 1952, Markowitz developed Modern Portfolio Theory MPT, and got a Nobel Prize for his pains. MPT remains a widely used tool for assemble investment portfolios of stocks, government bonds, and other investables, including by pension funds. Until MPT, investors generally made ad-hoc investments in areas where they fancied their chances. Some still do. MPT’s claim to fame is the idea that the greatest reliable level of benefits (i.e. low variation) comes not from trying to ‘pick winners’, but from systematically investing in a diverse portfolio of options whose value will not vary in the same way or direction (think of a global portfolio of umbrella manufacturers and ice cream makers). The logic goes one step further with the realisation that you can create portfolios with higher average levels of expected return than others, but the uncertainty — i.e. the variance or distribution around the average — will also be greater, with a greater proportion of catastrophic losses, leading to an alternative name, Mean Variance Portfolio Theory. So why would people not always seek the highest possible return? One reasons is if, as for a pension fund, the need to avoid catastrophic loss is important. But MPT is also used for business-to-business finance. There, if institutional investors are only trying to cover another financial risk, of a known size, they would search for the cheapest portfolio that adequately covers that risk, not the very highest possible return that will likely cost more to buy the assets.
Getting ESG analytics into MPT
Whether one can actually achieve greater returns than a passive market tracking fund is, depending on who you listen to, down to the skills of the investor, or an exercise in futility (see drop-down Multiple Paradigm Shifts). However, if you are happy to go with MPT, and given the huge global range of choices now available from which to construct a portfolio for any given level of projected risk/reward, it is possible to add additional layers of refinement to represent other concerns, for example the requirements or ‘behaviours’ of the investors. This can include environmental or social criteria, in much the same way, and using the same data, as done by adding Environmentally Extended satellite accounts to standard Multi Regional Input Output tables, creating what we call ‘Really Modern Portfolio Theory’ or rMPT.
In 2010 a still active Markowitz and colleges published one method of adding these additional layers, called Behavioural Portfolio Theory with Mental Accounts. As the name suggests, this tries to capture the different ‘mental accounting’ frameworks clients may want (their agent) to capture in their investment portfolios. So the investor may want part of their money to go into securing a certain level of income with high certainty and another layer of more risky investments to spend on luxuries that would be nice to have, but could be done without. Environmental or Social trend bending requirements could be another flavour, an additional layer included where the investor is not fully confident that they will produce the returns of conventional investments, and want to hedge their bets across a broad portfolio. It may not be perfect in the eyes of the purist, but may appeal to those who would otherwise not invest at all in bending the trends.
Turned another way, one may want to create an investment portfolio in a specific sector that targets a certain level of return on investment, but also optimises the reduction in greenhouse gas emissions for that level of return. So Z/Yen, among others, have explored the theory of assembling portfolios of energy investments with different levels of carbon dioxide emissions, in conjunction with other investments, such as forestry, that absorb CO2. One may know the (variation in the) cost of particular means of generating electricity without abatement, the (variation in the) various levels of CO2 reduction that could be achieved and the (variation in the) costs associated with these levels of abatement. So – given the power of modern computers – one can apply brute computational force to do masses of simulations of random combinations of investments, and outcomes for each, consistent with the underlying distribution of costs and returns and CO2 reductions expected for each. Out of the possibly tens of thousands of projected combinations and outcomes, one will find a few that financially perform exceptionally well — but are likely implausible because they requires a myriad of things to turn up trumps at the same time. If one comes down a couple of notches in one’s expectations, one can find many possible combinations that produce pretty good returns. The environmental twist is to then look amongst those options, all producing the same level of predicted financial returns , for the subset that produce no more than your limit for greenhouse gases emissions. In a further twist, one can also look at those factors affecting CO2 emissions and costs which, if uncertainty could be reduced, would allow investments with greater returns to be selected with the level of certainty required by the investor. In Z/Yen’s case they found significant uncertainty associated with forestry, which had a big overall effect on the overall returns.
Created 2014 MMG