Essays in Climate Finance

Rahul Bhattacharjee, PhD
Sentiment Analysis, Climate Risk, News Articles, Lexicons, Positive/negative Shock, UNPRI Signatories, Climate Risk Factors, Physical Climate Risk

Abstract :

Study the effect of climate related news sentiments: This paper develops a method to study the impact of climate related news sentiments on daily stock prices. First the paper uses two natural language processing algorithms, namely the VADER and TextBlob lexicons. With the proper  implementation of the lexicons the sentiments from the news articles are extracted. The polarity scores generated as an output quantify the sentiments to be either positive or neutral or negative. The sentiment analysis is done on daily climate related news for the FTSE 100 stocks. The study period considered for this paper is from 2006 to 2020. It should also be noted that since climate related news might not be published daily so in this paper I only consider those days when the news was published. After the first step, the paper analyzes the stock return movements on the relevant dates for the respective stock names that had climate related news published. To study the stock return movement the paper bifurcates a return on a particular day into two types, expected normal return and excess over normal or shock component. To do this the paper utilizes two models, the Constant mean model and the Market model. Further the paper approaches the method of calculation the normal and shock returns from two time windows, one been short term 5 days window and the other been a bit longer 22 days window. The study analyzes the relation between the sentiment obtained from the news articles and the stock return behavior for the relevant days. It also studies the evolution of the shock return over the study period for the FTSE 100 stocks. Further it analyzes the contribution of the different industries year wise to the overall positive or negative shocks. I find that the technique based on the lexicons to extract the sentiments and then relate it to the historical stock returns show around 50% accuracy. The study been a simple setup hence the accuracy could be improved by analyzing the sensitivity of the different parameters and that remains a further area of research. It also shows that over the study period particularly during the significant COP events like 2009, 2015-2016, etc. the positive or negative premium of the stock returns increase and then fades away as the event ends and time progresses. And that primarily the industries that have exposure to climate risk related events like Banks, Mining, Oil & Gas are the ones whose companies are more responsive when there are positive or negative climate related news. 

Study on the approaches adopted by responsible investors to incorporate climate considerations in their operations: This paper studies the adoption of approaches to incorporate climate-related consideration in investment decisions by the global signatories of UN Principles for
Responsible Investment (UNPRI). The study uses a unique data-set from the (UNPRI) database reported in the year 2018. It provides the magnitude of Assets under Management of 50 largest economies, associated to the adoption of various climate-related metrics and tools, by the different asset managers/signatories reported. The 14 approaches that were reported in the database could be divided into two categories - Activities undertaken by investors to respond to climate  change risks and tools used by the investors to manage emission risk. As part of the analysis in this paper I study the relationship between the approaches and the responsible investment for each region. And then do an inter region comparison of the results. In this paper I also study the concentration of responsible investment across the different approaches. From the study I find that across the regions the approaches are significant to responsible investment. For example, Europe region show sensitivity on activities that seek changes from policy makers related to address climate risk. Investors use tools to manage risk like adopting formal contracts to integrate climate risk in external investment and fund managers actively monitoring the risk related to investments in entities that contribute to emissions. One interesting find of the study is that approaches related to physical climate risk like scenario testing is particularly significant for Oceania countries that are more prone to experience severe physical risk events like sea level rising, cyclones, etc. I also found that overall investments are more sensitive to use tools like carbon foot printing and encourage the investors to monitoring emissions related risks in their portfolio choice, etc. In terms of activities undertaken for climate risk, I found that investors across regions are more prone to use emissions data analysis in their portfolio decisions or seek that climate change related risks should be integrated into the decision making process of the companies. As part of the study I observe that larger investors tend to adhere to these climate related approaches. 

Publication date of the thesis

Thesis committee

Supervisor:  Gianfranco Gianfrate, EDHEC Business School 

External reviewer: Marco Grotteria, London Business School 

Other committee members: Emmanuel Jurczenko, and Enrique Schroth, EDHEC Business School