-
- QUESTION
EC 226 Homework #7
- Use the date in DB1B which is the Airline Origin and Destination Survey. It contains 10% sample of airline tickets collected by the Bureau of Transportation Statistics.
- To download data go to https://www.transtats.bts.gov/tables.asp?DB_ID=125&DB_Name=, then click ‘download’ in DB1BMarket. In the next page, you choose one geography, year and period. For example, I use Massachusetts in 2018 Q1.
- You can download all variables (shown as field name) which makes your data too big, so only choose the following variables; ItinID, MktID, Origin, Dest, OpCarrier, MktFare, MktMilesFlown, NonStopMiles.
- The name of the downloaded file is ‘Origin_and_Destination_Survey_DB1BMarket_year_quarter’ and change its name to ‘Airline sample’.
- The next step is to change your ‘Airline sample. cvs’ to ‘Airline sample. dta’ file. Start Stata program and import text files (comma- or tab-separated files) by going to File > Import > Text data. Then, choose your data and click ok.
- Start ‘Airline sample. do’ file downloaded from the Moodle.
- Find the average values of market fare and market miles flown, along with standard deviations.
- Estimate the regression model
by OLS and OLS with clusters. Interpret the coefficients.
- Find the p-value for the test against . Do you reject at 5% level?
- Find the p-value for the test against . Do you reject at 5% level?
- Estimate the following simple regression of P on HHI: . Is the estimated coefficient on HHI much different from the estimate in part b? Why or Why not?
| Subject | Statistics | Pages | 5 | Style | APA |
|---|
Answer
- Find the average values of market fare and market miles flown, along with standard deviations.
On the analysis, the market fares indicate a mean of $301.74 with a standard deviation of 238.07. The maximum fare viewed at $9,000 with a minimum of $0. The market miles covered presented a mean of 2309.68 with a standard deviation of 1170.22, having a maximum of 9700 miles with a minimum of no distance covered.
- Estimate the regression model
by OLS and OLS with clusters. Interpret the coefficients.
On the regression model, the intercept indicates a positive effect on the market prices of the flights (β=176.84). The parameter indicates a significant effect on the market prices of the airlines, hence required in the model (t=45.76, p=0.000). Jet blue indicates a negative effect on the model (β=-0.022) with an insignificant effect on the market prices. This males the Jet blue parameter not required by the model (t=-1.30, p=0.193). South West on the other hand also indicated a positive effect on the market prices (β=0.008) with no significant effect on the market prices (t=0.49, p=0.624).
Spirit indicated a positive effect on the market prices charged (β=0.031) with a significant effect on the market prices at 10% level of significance. This ensures the inclusion of the parameter in predicting market prices charged. HHI indicated a positive effect on the market prices charged (β=0.004) with an insignificant effect on the market prices at 5% level of significance (t=0.25, p=0.80). This ensures the inclusion of the HHI parameter in predicting market prices charged. Spirit indicated a positive effect on the market prices charged (β=0.031) with a significant effect on the market prices at 10% level of significance. This ensures the inclusion of the parameter in predicting market prices charged. Market miles covered indicated a positive effect on the market prices charged (β=0.088) with a significant effect on the market prices at 5% level of significance. This ensures the inclusion of the parameter in predicting market prices charged. Non-stop miles indicated a negative effect on the market prices charged (β=-0.037) with a significant effect on the market prices at 5% level of significance.
- Find the p-value for the test against . Do you reject at 5% level?
P=0.624. This shows evidence to fail to reject the null hypothesis at 5% level of significance since the p-value greater than 0.05.
- Find the p-value for the test against . Do you reject at 5% level?
P=0.800. This shows evidence to fail to reject the null hypothesis at 5% level of significance since the p-value greater than 0.05.
- Estimate the following simple regression of P on HHI: . Is the estimated coefficient on HHI much different from the estimate in part b? Why or Why not?
On the estimated model, we have the equation:
The equation shows no much differences from part (b), with the evidence of a non-statistical implication of the HHI on the market prices (t=-0.21, p=0.833).