By LARRY SWEDROE
At the beginning of 2021, asset owners and investors representing over $100 trillion had signed on to the UN Principles for Responsible Investment. In addition, according to the 2020 Report on U.S. Sustainable and Impact Investing Trends, ESG investing now accounts for more than one-third of total assets under management in the U.S., or about $17 trillion, a 42% increase since 2018.
Joop Huij, Dries Laurs, Philip Stork and Remco Zwinkels contribute to the sustainable investing literature with their November 2021 study, Carbon Beta: A Market-Based Measure of Climate Risk, in which they examined the evidence of the impact of a climate risk factor on equity portfolios. To help investors address their concerns about climate risk in their portfolios, they proposed a proxy for a climate risk factor, the pollutive-minus-clean (PMC) portfolio. PMC captures the differences in returns to firms that have high versus low corporate emissions. By regressing individual stock returns on the PMC factor, they obtained estimates of asset-level climate risk exposure, what they called “carbon beta”. Their greenhouse gas emissions data (scope 1 and scope 2) was from S&P Global Trucost, a leading provider of corporate emissions data.
They explained the economic rationale behind their carbon risk factor: “If concerns regarding the climate unexpectedly rise, consumer demand will shift from brown products and services to green ones. Producers of these products and services will benefit accordingly, which pushes up their valuation. Simultaneously, investors who care about the climate will substitute their brown asset holdings for greener alternatives, either because they derive more utility from holding green assets, because they are publicly pressured to do so, or because they anticipate the introduction of stricter environmental policies.”
Their proxy for the carbon risk factor was intended to capture relative changes in the valuation of brown compared to green firms caused by unexpected changes in consumers’ and investors’ concerns about the climate. They constructed their carbon risk factor by forming a long-short portfolio based on companies’ relative emissions. The long leg of the portfolio contained the stocks of the heaviest-emitting 30% of firms offset by a short position in stocks of the least-emitting 30% of firms. To determine stock return sensitivities to carbon risk, they performed time-series regressions of equity returns on the carbon risk factor while controlling for additional factors known to drive returns. Their data sample covered the period 2004-2020.
Huij, Laurs, Stork and Zwinkels also constructed a climate policy uncertainty (CPU) index that was determined by the textual similarity between daily news articles published in the Wall Street Journal and a corpus of texts on climate change collected from official reports and Wikipedia. News data began in 1998. They hypothesised that periods in which climate change is frequently reported in the news tend to coincide with episodes of heightened uncertainty around future climate policies. Following is a summary of their findings:
The highest carbon betas are in the energy, materials and utilities sectors, while firms in the information technology, financial and health care sectors tend to exhibit negative carbon betas.
Carbon betas in South America, South Africa and Australia are relatively high. European countries tend to have low to negative carbon betas, reflecting that the European Union is at the forefront of regulating climate change.
All else equal, firms with high carbon betas tend to have lower market capitalisations; higher property, plant and equipment, and investment as a fraction of assets; lower research and development expenses compared to assets; lower profitability; and higher greenhouse gas emissions.
Comparisons of carbon betas with alternative measures of climate risk reveal robust associations between carbon betas and emissions, emission intensities, climate change exposures and MSCI Climate-Value-at-Risk (CVaR) scores (firm emissions and green patent innovations are the main determinants of CVaR).
Assets with negative carbon beta have a tendency to appreciate in times when investors are concerned about the climate. Such assets can be regarded as “climate hedge” assets because they deliver high returns when climate change concerns increase.
Returns to stocks with high carbon betas were lower during months in which climate change was more frequently discussed in the news, during months in which temperatures were abnormally high (in months with temperature anomalies above the 90th percentile, a standard deviation increase in carbon beta tended to be associated with a 36 basis points lower monthly excess return, all else equal) and during exceptionally dry months. (The data sample spanned the period 2007-2020.)
In months when the CPU index increased (decreased), stocks with higher carbon betas tended to have economically significant lower (higher) returns—for two firms that differed by only one standard deviation in carbon beta, the firm with the higher carbon beta tended to underperform the other firm by an annualized 1.32 percentage points for each standard deviation with which the CPU index increased. (The data spanned the period 2007-2020.)
Asset prices adjusted much more in response to increases in the CPU index than in response to reductions in the CPU index.
When two stocks differed in carbon beta by one standard deviation and were otherwise equivalent, the return on the asset with the higher carbon beta was about 49 basis points lower for each standard deviation with which the CPU index increased.
Variation in carbon betas correlates with green patent issuance and forward-looking measures of climate risk.
The peak in the CPU was reached in December 2009, when the 15th Conference of the Parties (COP15) was held in Copenhagen. COP15 was one of the earliest international conferences to bring climate change to the highest political level. As a result of the conference, the Copenhagen Accord was signed. The accord expressed clear political intent to limit carbon emissions and respond to climate change. The CPU index reached its second-highest level in November and December 2015, during COP21 in Paris. At that conference, the Paris Agreement was negotiated. At the end of 2020, the index remained at elevated levels. The period marked a series of mass protests to demand action on climate change. Those strikes coincided with the Climate Action Summit in New York.
The underperformance of high carbon beta firms was primarily an industry effect: Firms with high (low) carbon betas tended to operate in sectors that had low (high) realized returns over the sample period. However, controlling for additional factors increased the carbon risk premium to about 1% per year for each standard deviation increase in carbon beta.
Their findings led Huij, Laurs, Stork and Zwinkels to conclude: “If investors dislike states of the world in which climate concerns increase, they should be willing to accept lower expected returns for climate hedge assets, in return for their ability to insure climate risk. To the contrary, in times of increasing concern about the environment, assets with high carbon betas have a tendency to depreciate in conjunction with the PMC portfolio, because these assets are expected to be negatively affected by a low-carbon transition. As such assets are riskier and shunned by investors, they should trade at discounts, and offer higher expected returns as a result. Our results indeed indicate this to be the case.”
The researchers added: “Selling pressure and higher discount rates induced by heightened climate concerns cause pollutive firms to depreciate, and clean firms to appreciate, in value. As the PMC portfolio holds a net long (short) position in brown (green) stocks, both channels lead to a reduction in the PMC portfolio’s value. The opposite occurs when concerns regarding climate change unexpectedly lessen, so that the return on the PMC portfolio increases.”
Investor takeaway
Investors can use the approach of Huij, Laurs, Stork and Zwinkels as a screening tool, flagging the companies whose carbon beta indicates a significant exposure to climate risks. Carbon beta can also be employed as an indicator of climate hedge potential, used for the construction of hedge portfolios with high returns in periods of climate stress.