Weekly Collection of Interesting Things (4/16/23)
Discover more from A Causal Affair
Weekly Collection of Interesting Things (4/16/23)
A weekly attempt to collate interesting ('Metrics + Economics + Finance) things I've found
Links
If you are a Stata user and want to either start using R, or at least feel more comfortable, this webpage from Kyle Butts, Grant McDermott and others is fabulous:
https://stata2r.github.io/
Welcome. This website is for Stata users who are interested in learning R. But it could also be useful for those going the other way around. We provide side-by-side code snippets for common tasks in both Stata and R, so that users have a dictionary for navigating across the two languages.
There are two main pages (✌️📄) on the site:
Data wrangling (🗄🧹) with data.table
Regression analysis (💻📈) with fixes
Nice thread on Optimal Experimental Design for Staggered Rollouts by Xiong, Athey, Bayati and Imbens (forthcoming in Management Science). This paper’s results are quite intuitive, and nicely link the idea of adaptive bandits to more standard econometric settings!
The multi-armed bandit problem models an agent that simultaneously attempts to acquire new knowledge (called "exploration") and optimize their decisions based on existing knowledge (called "exploitation"). The agent attempts to balance these competing tasks in order to maximize their total value over the period of time considered.
Easy to see how this trade-off is intuitive when rolling out treatments to different groups where the effect can take a while to kick in!
CSMGEP is sponsoring a dissertation session at the AEAs next year. If you would like to nominate a grad student (or self-nominate yourself), email Rodney Andrews by May 19th, 2023.
Call for papers from the Philly Fed’s Consumer Credit conference
A Chrome add-in to open any ArXiv working paper directly in Overleaf (source code and all!): https://github.com/amitness/open-in-overleaf
Code
Neat paper doing high dimensional logit choice modeling in PyTorch: https://arxiv.org/pdf/2304.01906.pdf Code here with some nice examples. In case you’re unfamiliar with PyTorch, you can use it within Python (or R, if you use Reticulate)
Papers
Some interesting papers for this week
Algorithmic Fairness
Unpacking the Black Box: Regulating Algorithmic Decisions by Laura Blattner, Scott Nelson and Jann Spiess
We characterize optimal oversight of algorithms in a world where an agent designs a complex prediction function but a principal is limited in the amount of information she can learn about the prediction function. We show that limiting agents to prediction functions that are simple enough to be fully transparent is inefficient as long as the bias induced by misalignment between principal's and agent's preferences is small relative to the uncertainty about the true state of the world. Algorithmic audits can improve welfare, but the gains depend on the design of the audit tools. Tools that focus on minimizing overall information loss, the focus of many post-hoc explainer tools, will generally be inefficient since they focus on explaining the average behavior of the prediction function rather than sources of mis-prediction, which matter for welfare-relevant outcomes. Targeted tools that focus on the source of incentive misalignment, e.g., excess false positives or racial disparities, can provide first-best solutions. We provide empirical support for our theoretical findings using an application in consumer lending.
Modeling advances create credit scores that predict default better overall, but raise concerns about their effect on protected groups. Focusing on low- and moderate-income (LMI) areas, we use an approach from the Fairness in Machine Learning literature — fairness constraints via group-specific prediction thresholds — and show that gaps in true positive rates (% of non-defaulters identified by the model as such) can be significantly reduced if separate thresholds can be chosen for non-LMI and LMI tracts. However, the reduction isn’t free as more defaulters are classified as good risks, potentially affecting both consumers’ welfare and lenders’ profits. The trade-offs become more favorable if the introduction of fairness constraints is paired with the introduction of more sophisticated models, suggesting a way forward. Overall, our results highlight the potential benefits of explicitly considering sensitive attributes in the design of loan approval policies and the potential benefits of output-based approaches to fairness in lending.
ESG in Finance
Four Facts About ESG Beliefs and Investor Portfolio by Stefano Giglio, Matteo Maggiori, Johannes Stroebel, Zhenhao Tan, Stephen Utkus & Xiao Xu
We analyze survey data on ESG beliefs and preferences in a large panel of retail investors linked to administrative data on their investment portfolios. The survey elicits investors’ expectations of long-term ESG equity returns and asks about their motivations, if any, to invest in ESG assets. We document four facts. First, investors generally expected ESG investments to underperform the market. Between mid-2021 and late-2022, the average expected 10-year annualized return of ESG investments relative to the overall stock market was –1.4%. Second, there is substantial heterogeneity across investors in their ESG return expectations and their motives for ESG investing: 45% of survey respondents do not see any reason to invest in ESG, 25% are primarily motivated by ethical considerations, 22% are driven by climate hedging motives, and 7% are motivated by return expectations. Third, there is a link between individuals’ reported ESG investment motives and their actual investment behaviors, with the highest ESG portfolio holdings among individuals who report ethics-driven investment motives. Fourth, financial considerations matter independently of other investment motives: we find meaningful ESG holdings only for investors who expect these investments to outperform the market, even among those investors who reported that their most important ESG investment motives were ethical or hedging reasons.
Counterproductive Sustainable Investing: The Impact Elasticity of Brown and Green Firms by Sam Hartzmark and Kelly Shue
We develop a new measure of impact elasticity, defined as a firm's change in environmental impact due to a change in its cost of capital. We show empirically that a reduction in financing costs for firms that are already green leads to small improvements in impact at best. In contrast, increasing financing costs for brown firms leads to large negative changes in firm impact. Thus, sustainable investing that directs capital away from brown firms and toward green firms may be counterproductive, in that it makes brown firms more brown without making green firms more green. We further show that brown firms face very weak incentives to become more green. Due to a mistaken focus on percentage reductions in emissions, the sustainable investing movement primarily rewards green firms for economically trivial reductions in their already low levels of emissions.