Global Divestment Day was February 13th, so the Wall Street Journal trotted out an opinion piece a few days in advance by the author of a new study arguing that “fossil-fuel divestment could significantly harm an investment portfolio”. As holiday-themed advice goes, it is about as useful as that provided by Punxutawney Phil, but a sight less entertaining.
To find out what financial analysis the study by Daniel Fischel actually conducted requires reading through 60 numbered paragraphs that include somewhat disjointed references to virtually every concept found in a basic investment textbook (indeed, no fewer than five of the paper’s references are to the estimable “Investments” textbook by Bodie, Kane & Marcus, with two references to Burton Malkiel’s “A Random Walk Down Wall Street” added for good measure). Those paragraphs also mention and summarily dismiss many other arguments, financial and non-financial, being made for divestment. All that information is intertwined enough to suggest to the uninitiated reader that this is obviously a mess best left alone, especially since the author has made clear that doing anything will be a significant financial hit to the investor’s portfolio.
What is that financial hit? Here’s the author’s fundamental claim from the opinion piece – “What we found is that the optimal portfolio, which included energy stocks, generated average returns 0.7 percentage points greater than portfolios that excluded them on an absolute basis.” That sounds pretty definitive and quite stark. What are these two portfolios?
The author assembled his portfolios from all the securities in the CRSP (Center for the Research in Security Prices) database of the major US exchanges. Those securities were divided into 10 sectors based on SIC (Standard Industry Classification) code. Under that definition, the energy sector consists of only companies in oil, gas, and coal extraction and products. The “divested” portfolio is everything in the economy other than those companies.
While this may match what some investors mean by divestment, there are many other approaches, so this “divested” portfolio offers no useful information about divestment strategies as a whole. It is also unrealistic to believe that an investment manager would remove an entire sector from a portfolio without making any adjustments to the allocation among the other sectors. Indeed, that is what almost every manager who creates a divested portfolio does.
The divested portfolio is the least of the problems in the analysis. The author then argues that this divested portfolio should be compared not to the market portfolio (as was suggested early in the paper), but to an “optimal” portfolio that will maximize the investor’s risk-weighted return (calculated as a Sharpe ratio) by combining the non-energy portfolio and the “energy” portfolio (the companies that were divested). That optimal portfolio “contains an allocation of 56.3% to the non-energy index and 43.7% to the energy index”. Read that sentence a second time. To project that level of performance, the authors propose investing over 40% of an endowment portfolio in a small, highly volatile industry sub-sector. That allocation, by the way, is projected to have been maintained over 50 years.
Based on those portfolios, you can stop reading the study right there. Find a endowment manager who suggests such an unrealistic allocation and I suggest you’ll find an unemployed endowment manager. Basing the analysis on such unrealistic portfolios completely undermines the fundamental claim the author has made for the cost of divestiture.
Other claims of cost impact from divestiture in the study are consistent in standing up equally poorly to analysis. If anything, transaction costs for the optimal portfolio are likely to be higher than for the divested portfolio. The divested portfolio, once established, will undergo almost no changes. The optimal portfolio will require periodic rebalancing – incurring costs – to maintain the required ratio of non-energy stocks to energy stocks. All portfolios have compliance costs, and the simple exclusion list of the divested portfolio is not among the most expensive things to monitor.
A bid-ask spread is paid every time an investor buys or sells a security. How can that be included as a cost of divestment unless no stocks are ever bought or sold in the non-divested portfolio?
The author further proposes that price impact would result in an additional $304 million cost of divestiture, another very precise number that is almost certainly completely inaccurate. Price impact is a term used to describe a loss of value experienced when a large block of stock is bought or sold in a very short time. Yet at the same time, the author is arguing that divestiture will not affect the equity valuations of companies because the divesting institutions “hold a very small share of stock in the targeted companies”, and that small share is spread over hundreds of endowments.
The $304 million cost assumes that all fossil fuel companies are sold by all endowments in a short enough time that all the transactions experience the price impact effect, and are then replaced with other stocks in an equally short time, with the much higher price impact of buying different stocks also completely credited to divestiture. The percentages used to calculate the price impact (0.35% for sales and 1.00% for purchases) are an average taken from a single study, one in which the authors exhaustively describe a wide range of values depending on the size of the trade, the size of the financial institution making the trade, the length of time it takes to make the trade and other characteristics. Did I mention that the stock trades studied took place between 1986 and 1988? Securities markets look very different thirty years later.
Even these are not all the flaws plaguing this study. But added together, they make clear that this study has no business claiming to have properly used the tools of investment analysis to demonstrate any cost to divestment, let alone the very large and specific numbers it cites.
Divestment is not a one-size-fits-all solution and it may not be right for a given investor. But don’t look to this study for accurate data produced by thorough analysis.