In: Statistics and Probability
Metro Transit seeks to understand the cross price elasticity of demand for bus transit. This dataset includes deseasonalized ridership data for one of the Metro Transit commuter markets from January 2009 to May of 2016 along with gas price and unemployment data in the twin cities region. The data are all in natural log form and we will assume constant elasticity/log-linear functional form for this analysis. Metro Transit has not changed the price during this period, so we will not consider this in the analysis. Please perform a regression to identify the cross price elasticity of demand for gas prices.
1. Run a regression using OLS
2. Provide the regression output
3. Discuss the effect of gas prices and unemployment on ridership
4. Are these effects elastic? Inelastic?
5. What variables might we be excluding?
6. Are there any policy implications for your findings?
Month | LN_Deseasonalized_Ridership | LN_Gas | Unemployment |
Jan-09 | 13.37310641 | 0.60976557 | 1.90822905 |
Feb-09 | 13.32106787 | 0.63127178 | 1.94090292 |
Mar-09 | 13.4810664 | 0.66782937 | 1.97254304 |
Apr-09 | 13.46502645 | 0.69813472 | 1.99616889 |
May-09 | 13.26538987 | 0.81977983 | 2.0192493 |
Jun-09 | 13.37107386 | 0.96698385 | 2.03232635 |
Jul-09 | 13.31737585 | 0.89199804 | 2.04523473 |
Aug-09 | 13.29812228 | 0.9282193 | 2.04904754 |
Sep-09 | 13.40551694 | 0.88376754 | 2.05284587 |
Oct-09 | 13.48313729 | 0.90421815 | 2.04997282 |
Nov-09 | 13.22331027 | 0.94000726 | 2.04709163 |
Dec-09 | 13.2114525 | 0.91629073 | 2.0399689 |
Jan-10 | 13.29923244 | 0.96698385 | 2.03279493 |
Feb-10 | 13.30863827 | 0.9282193 | 2.02433493 |
Mar-10 | 13.50917425 | 0.98581679 | 2.01580275 |
Apr-10 | 13.42724516 | 1.01884732 | 2.00796192 |
May-10 | 13.24494832 | 1.00063188 | 2.00005913 |
Jun-10 | 13.36003719 | 0.97077892 | 1.99195164 |
Jul-10 | 13.22552564 | 0.97455964 | 1.98377789 |
Aug-10 | 13.34415553 | 0.96698385 | 1.97365218 |
Sep-10 | 13.39940096 | 0.97832612 | 1.9634229 |
Oct-10 | 13.45701554 | 1.00063188 | 1.95311315 |
Nov-10 | 13.30133828 | 1.01523068 | 1.94269585 |
Dec-10 | 13.21605909 | 1.05431203 | 1.93389381 |
Jan-11 | 13.38636673 | 1.0919233 | 1.92501362 |
Feb-11 | 13.30487091 | 1.11841492 | 1.91436936 |
Mar-11 | 13.55652222 | 1.21491274 | 1.90361058 |
Apr-11 | 13.45702078 | 1.28647403 | 1.88962759 |
May-11 | 13.39155446 | 1.30833282 | 1.8754463 |
Jun-11 | 13.43485923 | 1.24415459 | 1.86031211 |
Jul-11 | 13.22750392 | 1.24126859 | 1.84494551 |
Aug-11 | 13.49816051 | 1.23547147 | 1.83016944 |
Sep-11 | 13.44749317 | 1.22082992 | 1.81517191 |
Oct-11 | 13.51930471 | 1.15373159 | 1.79949266 |
Nov-11 | 13.38385594 | 1.13140211 | 1.78356347 |
Dec-11 | 13.23804989 | 1.10525683 | 1.76837773 |
Jan-12 | 13.43538754 | 1.137833 | 1.75295765 |
Feb-12 | 13.408108 | 1.1817272 | 1.74092747 |
Mar-12 | 13.51266369 | 1.26412673 | 1.7287508 |
Apr-12 | 13.47070802 | 1.26412673 | 1.72037231 |
May-12 | 13.45759496 | 1.22082992 | 1.71192284 |
Jun-12 | 13.36429221 | 1.18478998 | 1.70470555 |
Jul-12 | 13.31210014 | 1.16627094 | 1.69743597 |
Aug-12 | 13.48748903 | 1.24990174 | 1.68871885 |
Sep-12 | 13.38541267 | 1.26976054 | 1.67992527 |
Oct-12 | 13.6355304 | 1.20896035 | 1.66897507 |
Nov-12 | 13.36439471 | 1.13140211 | 1.65790364 |
Dec-12 | 13.12661336 | 1.09527339 | 1.6450336 |
Jan-13 | 13.4380618 | 1.08518927 | 1.63199596 |
Feb-13 | 13.3397626 | 1.20597081 | 1.61841906 |
Mar-13 | 13.42558241 | 1.21194097 | 1.60465549 |
Apr-13 | 13.52591974 | 1.17557333 | 1.59127618 |
May-13 | 13.39815126 | 1.22964055 | 1.57771564 |
Jun-13 | 13.2768924 | 1.23256026 | 1.56521661 |
Jul-13 | 13.37759076 | 1.17248214 | 1.55255937 |
Aug-13 | 13.36846794 | 1.17248214 | 1.54134385 |
Sep-13 | 13.45266352 | 1.16627094 | 1.53000112 |
Oct-13 | 13.64705169 | 1.09861229 | 1.51727745 |
Nov-13 | 13.30320038 | 1.0612565 | 1.50438957 |
Dec-13 | 13.18893326 | 1.05779029 | 1.48812076 |
Jan-14 | 13.34526174 | 1.07840958 | 1.47158267 |
Feb-14 | 13.31717668 | 1.10856262 | 1.45346925 |
Mar-14 | 13.45446108 | 1.16938136 | 1.43502167 |
Apr-14 | 13.51276241 | 1.18478998 | 1.41652842 |
May-14 | 13.36842706 | 1.1817272 | 1.39768647 |
Jun-14 | 13.34653835 | 1.2029723 | 1.37977868 |
Jul-14 | 13.38627655 | 1.15373159 | 1.36154434 |
Aug-14 | 13.32658434 | 1.12817109 | 1.34477916 |
Sep-14 | 13.41585552 | 1.10856262 | 1.3277281 |
Oct-14 | 13.55875286 | 1.0260416 | 1.3144981 |
Nov-14 | 13.13673318 | 0.95935022 | 1.30109071 |
Dec-14 | 13.13530136 | 0.78845736 | 1.29054446 |
Jan-15 | 13.27057192 | 0.58778666 | 1.27988607 |
Feb-15 | 13.2181603 | 0.69314718 | 1.26989059 |
Mar-15 | 13.37712701 | 0.76546784 | 1.25979391 |
Apr-15 | 13.3900275 | 0.76546784 | 1.25140576 |
May-15 | 13.11402112 | 0.84156719 | 1.24294666 |
Jun-15 | 13.25916856 | 0.90016135 | 1.23700489 |
Jul-15 | 13.27298664 | 0.88376754 | 1.23102731 |
Aug-15 | 13.25256961 | 0.85015093 | 1.22921619 |
Sep-15 | 13.35563596 | 0.74193734 | 1.22740208 |
Oct-15 | 13.47766232 | 0.75141609 | 1.22745219 |
Nov-15 | 13.13445939 | 0.64710324 | 1.2275023 |
Dec-15 | 13.12019576 | 0.54812141 | 1.22782692 |
Jan-16 | 13.21435484 | 0.48242615 | 1.22815114 |
Feb-16 | 13.24498229 | 0.39877612 | 1.22894265 |
Mar-16 | 13.37820262 | 0.56531381 | 1.22973324 |
Apr-16 | 13.32100748 | 0.61518564 | 1.23095431 |
May-16 | 13.20547256 | 0.70309751 | 1.23217418 |
As per the given data from - Jan09 to May16 then solution is;
Part 1:
We ran the regression using R.
code used :-
> reg<- lm(ln_ridership ~ ln_GAS + ln_unemployment , data
= datafile name)
> summary(reg)
The result is as follows:
Residuals:
Min 1Q Median 3Q Max
-0.249822 -0.059711 -0.005834 0.066461 0.275685
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 13.07126 0.07388 176.937 <2e-16 ***
ln_GAS 0.21077 0.05424 3.885 0.0002 ***
ln_unemployment 0.04518 0.04048 1.116 0.2674
---
Part 2
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Residual standard error: 0.106 on 86 degrees of freedom
Multiple R-squared: 0.1926, Adjusted R-squared: 0.1739
F-statistic: 10.26 on 2 and 86 DF, p-value: 0.0001009
Part 3:
effect of gas price = If gas prices changes by 1 % then on an average ridership changes by 0.21 % keeping all other variables constant.
effect of unemployment = If unemployment changes by 1% then on an average ridership changes by 0.04% keeping all other variables constant.
Part 4 :-
both of these effects are inelastic as elasticity is less than 1. The regression coefficients are elasticities.
Part 5:
we might exclude unemployment as it is statistically insignificant since the p value is less than 2( also p value is less than alpha which we assume to be 5%)