In this exercise you will analyze the impacts on disclosure of a new policy announced by the US regulator, the Federal Trade Commission (FTC). The policy consisted on a "warning that undisclosed posts would start to be prosecuted. This policy was established mid 2017. Data available to study the question include about 400 macro and mega influencers from the UK and US. Macro influencers are those with a number of followers between 500,000 and one million, and mega influencers are those with more than a million followers. Our data includes the following variables: Userid: unique identifier number for each influencer Country: US or UK Week: indexes weeks of 2017 (1...,52). The FTC policy was established in week 30 (included) Followers: each influencer's number of followers (measured in week 1) number_posts_all: total number of posts made by each influencer on a given week number_posts_disclosed_ad: total number of disclosed posts made by each influencer on a given week For the following analysis, use data of UK influencers as a control. 1/ In R, estimate diff-in-diff specifications that address the following questions: a. By how much the FTC policy impact the total number of posts? b. By how much the FTC policy impact the number of disclosed posts? 2/ In R: a. Create a variable at the user/week level which reflects the probability of disclosure b. Estimate a diff-in-diff model to that addresses: by how much the FTC policy impact the probability of disclosure? c. Provide a possible explanation for the joint insights of your analyses in (ii) and (iii.b) Screenshot(Alt + A) 3/ In R, estimate diff-in-diff specifications to address the following questions: a. Was the impact of the FTC policy on the total number of posts different for macro and mega influencers? b. Was the impact of the FTC policy on the number of disclosed posts different for macro and mega influencers? c. Was the impact of the FTC policy on the probability of disclosure different for macro and mega influencers? d. Provide a possible explanation for the results obtained above