Congestion pricing, III

Continuing with the theme . . .

It occurred to me that another benefit of congestion pricing is that it lets us know when we ought to add new lanes to a road.

Take a congested highway like MoPac.  Traffic is routinely bogged down on both AM and PM commutes.  Will adding lanes help?  Without a toll, perhaps not.  The traffic jams on MoPac discourage some drivers from taking MoPac; the inevitable wait in traffic is a higher cost than they're willing to pay.  If a lane is added, increasing the traffic capacity, some of these deterred drivers will rejoin the traffic.  Depending on how many such drivers there are, an eight-lane MoPac quickly could become just as congested as the six-lane MoPac. 

Congestion pricing tells us when to add another lane.  As demand for travel on a particular road increases, the congestion-priced toll will have to increase in order to keep traffic flowing.  Perhaps the congestion toll starts at $3 during peak travel time.  As population increases and more commuters want to use the road during peak periods, the toll goes up to $5, then $6, and so on.  Building a new lane will increase the "supply" of road, allowing the congestion price to fall.  This gives us a way to balance the costs and benefits of a new lane:  Build a new lane if the saving in congestion charges exceeds the cost of construction.  (The cost of construction, of course, should be passed along to drivers in the form of a revenue toll.)

This might seem paradoxical -- Won't this guarantee that the state loses money? -- but the point of the congestion-price toll is not to raise money, it's to ration access to a scarce good.   

An example.  Let's say a congestion-priced MoPac can carry a total of 20,000 cars during the afternoon commute at a congestion charge of $5/car.  Suppose a new lane would increase the total throughput to 25,000 while lowering the congestion charge to $3/car. 

Should the new lane be built?

We can calculate the benefits. I think we ignore the 5,000 new drivers -- we don't know whether driving on MoPac is worth $3.01 or $4.99 to them.  (If they valued it at $5, presumably they'd already be taking MoPac.)  But we do count the benefit to the 20,000 drivers who were already paying the congestion price.  They valued driving on MoPac at a minimum of $5.  The new lane (and lower congestion price) is worth at least $40,000/day to them.

Assume an identical benefit to morning commuters.

If the pro rata cost per day of building and maintaining the new lane is less than $80,000, it should be built.  If it is more than $80,000, it should not be built. 

OK, this is over-simplified.  But I think it focuses on the right question.  I'm not sure how you even begin to make a cost/benefit analysis without congestion pricing.  What's the cost of congestion on a "free" road?  Someone's making that cost/benefit analysis, I realize -- roads are being widened all the time.  I just don't see how they have any confidence that the benefits exceed the costs. 

NB:  A draft version of this post was inadvertently published this morning.  My apologies if you waded through that gibberish before I de-published it.