## Description

This question will ask you to estimate the price elasticity of demand for fresh sardines across 56

ports located in 4 European countries with monthly data from 2013 to 2019. The data are

contained in the file EU_sardines.csv, which is available on Gauchospace.

Each row in the data file is a combination of port location (where the fish is landed and sold) in a

given year and month. You can ignore the fact that the sample is not balanced (the number of

monthly observations varies across ports).

For the assignment, you will need the following variables: year, month, country, port (port where

sardines are landed and sold), price_euro_kg (price per kg in €), and volume_sold_kg (quantity

of sardines sold in kg). In the questions below, I use log() to denote the natural logarithm.

(a) Estimate a bivariate regression of log(volume_sold_kg) on log(price euro_kg). What is the

price elasticity of demand for sardines? Test the null hypothesis that the price elasticity is equal

to -1.

(b) Like in Lecture 8 (see the IV.R script), we will use wind_m_s as an instrument for

log(price_euro_kg). To begin, estimate the first-stage regression relating log(price_euro_kg) to

wind_m_s. Interpret the estimated coefficient on wind speed. Does it have the expected sign?

Also test for the relevance of the instrument and whether it is a “weak” instrument by reporting

the proper F-statistic.

(c) Estimate the TSLS estimator of the price elasticity of demand for sardines using wind_m_s

as an instrument for log(price_euro_kg). What is the estimated price elasticity of demand for

sardines?

(d) Repeat the exercise in (c), but include fixed effects for each year, month, and country. [Hint:

you can use the command “as.factor(country) + as.factor(year) +as.factor(month)” to the ivreg

function in R]. Report the estimated price elasticity of demand and the F-statistic testing for

relevant and non-weak instruments.