--- title: "Exercise template for 'Causal mechanisms'" author: Your name format: html bibliography: book.bib --- ## Discussion questions ### @Li:2018tj 1. Let's think about the causal mechanism in @Li:2018tj. Suppose you are a manager involved in the process of preparing information about significant customers for your firm's 10-K. What do you think your main area of expertise would be? If a court upheld or rejected the IDD while you are in this role, would you expect to learn about that in your role? If so, how would you learn this information? If not, how might the court ruling nonetheless have an effect on your decisions regarding what to include in the filing? Assume that the relevant managers are aware of relevant court rulings on IDD, how would you expect this to affect disclosure decisions? When would this effect occur? How long would it last? 2. Note that PepsiCo is headquartered in the state of New York. Reading the discussion of the *PepsiCo Inc. v. Redmond* case [@Li:2018tj, pp. 272-3], does this fact have implications for their research design. How might @Li:2018tj need to do things differently? 3. Look at the plots in @Li:2018tj [p. 284]. What do we expect to see here? Try to imagine the plots without the fitted curves and red vertical dashed line. How compelling is this evidence? Why do you think the authors included this? How might one depict the results reported in Table 2 on the figures in @Li:2018tj [p. 284] (pay attention to magnitudes)? 4. In [Section 3.2](https://iangow.github.io/far_book/reg-basics.html#sec-regs-r), we ran regressions with **two-way fixed effects** (i.e., fixed effects for `grade` and for `id`, the student identifier). What variables are analogous to `grade` and for `id` in regressions (3) and (4) in Table 2 of @Li:2018tj? Why is there no $\mathit{POST} \times \mathit{TREAT}$ coefficient reported in Table 2? How would you code $\mathit{POST}$ for a firm in a state that never has IDD? 5. Compare Tables 2 and 3 of @Li:2018tj. Which regressions are most comparable across the two tables? What insight does Table 3 provide over Table 2? ### @Burks:2018vb 1. Suppose you are a senior executive at a bank. What opportunities might implementation of IBBEA by a state other than your own provide to you? How would you evaluate such opportunities? How long would it take to evaluate them? 2. Suppose you are a senior executive at a bank in state that is about to implement IBBEA. What actions might you take if you anticipate increased interest in your state by interstate banks? Would the actions you take depend on how out-of-state banks enter? 3. @Burks:2018vb do not explain what a "deposit cap" is. Rice and Strahan (2010) describe it thus: "IBBEA specifies a statewide deposit concentration limitation of 30% with respect to interstate mergers that constitute an initial entry of a bank into a state." What might drive differences in the deposit cap for different states? (See Table 1.) Would this restriction affect behaviour of all banks in an IBBEA-implementing state in the same way? 4. "The results are consistent with disclosure significantly increasing upon introduction of the IBBEA (the main effect) ... we find that the restriction imposing a deposit cap of less than 30% has a significant mitigating effect on disclosure." How many states implemented a cap of less than 30%? (See Table 1.) We checked one of these states (Colorado) and it had a cap of 25%. In 1997, the bank with the largest market share had 17.87%. With these facts, what is the practical difference between 25% and 30% in this case? ### @Christensen:2017uw; @Glaeser:2017vk 1. @Christensen:2017uw claim to find a causal effect of requiring disclosure in SEC filings on mine safety. Why is it significant that the information was already available on the MSHA website? How persuasive do you find the evidence of a causal effect to be? What elements do you find particularly persuasive (or problematic)? 2. What is the causal mechanism through which @Christensen:2017uw argue (or conjecture) this effect occurs? 3. @Glaeser:2017vk suggest that use of securities laws to enforce safety regulations is another mechanism through which the causal effect occurs? Does this undermine the validity of the estimate of the causal effect?