1. Carry out the analysis in sub-section 5.4 (which is above) that compares land use policy and fuel tax policy. Report your findings clearly but briefly.  
2. Offer some interesting thoughts on your findings in your own words to conclude.

Among the two competing policy options, namely the policy of land use versus fuel tax, opinions with regard to the policy choice differ. Unfortunately, there are pitfalls in using tax as an instrument. For example, tax as an instrument is not politically expedient as it leads to higher gasoline price, which may have immediate adverse spill-over effect in other markets leading to general price rise. Also, since travel behavior is habit-persistent it requires a long-term approach to urban and transportation planning, whereas tax policies can be short-lived. A comparison of the two policies involves a cost-benefit analysis of a change in the two policies. However, data limitation to monetize costs and benefits prevents us from undertaking such a comparison and thus, instead of comparing the two policy choices, we will present only an analysis of the quantitative impact of a change under each policy option. In the full sample, the annual average vehicle miles traveled per household is 22,182 miles. A 10 percent increase in residential density, jobs per-capita, and per-capita public transit operating expenses amounts to an increase of 234 people per square miles, 0.045 jobs per person, and $17,000 more transit operating expenses per person, respectively. If indeed there was a 10 percent increase in the urban form values as above, then, all else equal, the independent impact computed from the estimated 3SLS model would be an annual reduction in VMT by 1,826 miles associated with residential density, 1,431 miles associated with jobs percapita, and 1,143 miles associated with per-capita public transit expenses. Using the 3SLS model, a combined impact of a 10 percent change in each of the three urban-form measures on vehicle miles traveled would be a maximum of 4,400 miles - a reduction of about 20 percent, which is quite substantial. If on the other hand, a price-based instrument was used, such as a fuel tax, that would impact the dollar cost per mile in our model directly. In our sample, average dollar cost per mile is roughly 6.7 cent. Based on our 3SLS model, a 10 percent increase in dollar cost per mile would result in a 3,975 miles reduction in annual vehicle miles traveled – a reduction of about 18 percent. We note that the quantitative impacts of the two policy options estimated above from the models are strictly not comparable, since the costs associated with the change under the two scenarios may differ. Moreover, the time frame to achieve the target in the two strategies may differ because the speed of the adjustment process may vary. For example, fuel tax policy may lead to a faster but painful adjustment process, whereas a comprehensive land use policy may result in a slower adjustment process giving households time to adapt to the new urban design and at the same time reduce auto dependence. Broadly, the latter policy to reduce auto dependence involves revisiting the current zoning policies for targeted changes, which may include 

(i) redevelopment plans to build denser neighborhoods and the necessary infrastructure that supports such neighborhoods; 

(ii) building communities with a wide variety of education, skills, and income levels; 

(iii) incentivizing employers to locate in specific regions to help achieve a desired job-housing balance, and finally, 

(iv) targeted investments in more public transportation infrastructure.