Road Recovery Ratio
Let’s perform an experiment. Some people like to judge public transportation investments based on their farebox recovery ratio or the percentage of operating costs paid by passengers as opposed to subsidies or other revenue. The naive assumption is if a public transportation system has a high ratio, it is a good system. However, fairbox recovery ratios may be low because of other priorities such as providing essential transportation services to low income people or maintaining service late into the night when ridership is lower. In this experiment, let’s do the same for roads. (A couple caveats: 1. This is unfair because fairbox recovery ratios only include operating costs and we will look at road construction costs and 2. Similar to other priorities for public transportation, roads may be "unprofitable" but still provide other benefits like a new carpool lane for express buses.)
To find our road recovery ratio, we will use the following below. Because “roads are paid for by gas taxes” (gas taxes do not come close to covering the costs of roads), we will use federal and state gas taxes as our source of revenue. Therefore, the amount of revenue from one mile of road in one year is
where the units of each quantity are
Caltrans maintains a Traffic Census Program where we can find the Annual Average Daily Traffic or average number of vehicles that drive along a segment of highway per day. The US Department of Energy states the the average car fuel economy is 37.6 miles/gallon. (Another caveat: yes, there are older, less efficient cars on the road.) Wikipedia maintains a table of federal and state gas taxes.
Now let’s take a look at a few recent highway projects and see how they did... According to this article, adding one lane to Interstate 405 for 10 miles ended up costing $1.61 billion, or $161 million/mile. In 2017, the average annual daily traffic through the Sepulveda Pass was about 280,000 vehicles/day. The freeway is 10 lanes wide which gives 28,000 vehicles/lane/day. Plugging these numbers into our above formula gives the following
Given the construction cost of $161 million/mile, it will take 745 years for the road to pay for itself with gas taxes. (Last caveat: this assumes maintenance cost is zero and all numbers stay the same for the next 745 years…) The above article also notes that lane “has not improved traffic.” How about another project? Spending $425 million to add two lanes for 10 miles to highway 101 in Santa Barbara County? The project description states the AADT is 90,000 vehicles/day for the current 4 lanes which gives 22,500 vehicles/lane/day.
In this case it will take 123 years to pay back. Better than the 405, but still not great. What did we learn with this experiment? Adding lanes to highways is incredibly expensive: roads are not profitable. The road in front of your house is going to have a much worse road recovery ratio because of lower traffic volumes. But just like public transportation, roads are not supposed to be profitable. Rather, they serve an important transportation function.