ITM 448 Homework Week 12

Task 1

We are studying the effect of smart contract GitHub activities on the underlying cryptocurrency price. We propose that high GitHub activities (lets only focus on the number of pull requests) indicate active innovation and debugging efforts, which in turn signify high-quality service and maintenance, thus resulting in a higher price of the underlying cryptocurrency.

We are studying the effect of smart contract GitHub activities on the underlying cryptocurrency price. We propose that high GitHub activities (let's only focus on the number of pull requests) indicate active innovation and debugging efforts, which in turn signify high-quality service and maintenance, thus resulting in a higher price of the underlying cryptocurrency. We collected data for 200 cryptocurrencies over a period of 6 months. For each cryptocurrency, during each month, we collect its monthly cumulative return (rt), and the monthly total number of GitHub pull requests (pt) generated by the contributors for the focal cryptocurrency community. We also capture the monthly average market cap (ct), and monthly average trading volume (vt) as additional control variables. Note that t indexes the month. Please complete the following steps: 1. If we directly run a regression rt=a+bpt+cct+dvt+, and find that the estimated b is positive, can we conclude that higher number of pull requests causes increase in the returns of the underlying cryptocurrency? Why? 2. To obtain causal inference, we plan to use IV method. We found an instrumental variable, Pull_OtherProject t, which is the monthly total number of pull requests generated by the GitHub contributors of the focal cryptocurrency community on other projects. Please justify why this could be a valid instrumental variable? 3. The dataset is provided in "cryptocurrency_github.txt", please complete the 2SLS IV estimation, and state your conclusion. Is higher number of pull requests causes increase in the cryptocurrency returns. Note that in the second stage, please also control for the cryptocurrency cross-section fixed-effect and the month fixed-effect. We are studying the effect of smart contract GitHub activities on the underlying cryptocurrency price. We propose that high GitHub activities (let's only focus on the number of pull requests) indicate active innovation and debugging efforts, which in turn signify high-quality service and maintenance, thus resulting in a higher price of the underlying cryptocurrency. We collected data for 200 cryptocurrencies over a period of 6 months. For each cryptocurrency, during each month, we collect its monthly cumulative return (rt), and the monthly total number of GitHub pull requests (pt) generated by the contributors for the focal cryptocurrency community. We also capture the monthly average market cap (ct), and monthly average trading volume (vt) as additional control variables. Note that t indexes the month. Please complete the following steps: 1. If we directly run a regression rt=a+bpt+cct+dvt+, and find that the estimated b is positive, can we conclude that higher number of pull requests causes increase in the returns of the underlying cryptocurrency? Why? 2. To obtain causal inference, we plan to use IV method. We found an instrumental variable, Pull_OtherProject t, which is the monthly total number of pull requests generated by the GitHub contributors of the focal cryptocurrency community on other projects. Please justify why this could be a valid instrumental variable? 3. The dataset is provided in "cryptocurrency_github.txt", please complete the 2SLS IV estimation, and state your conclusion. Is higher number of pull requests causes increase in the cryptocurrency returns. Note that in the second stage, please also control for the cryptocurrency cross-section fixed-effect and the month fixed-effect