An essential though incomplete way to answer the question is by compar- ing the proposed bridge’s costs and benefits. For now, we mean internal costs and internal benefits. These accrue only to the bridge project itself and leave out, say, the noise and exhaust the bridge creates or the possibility that quicker mobility will spur investment in new office buildings downtown. In studying just the internal effects, we confine ourselves to the narrow world in which all we care about is cost to the bridge owner-operator (that’s the city government) and benefit to the bridge users. We get only later in this
chapter to the problem of external costs and benefits, which affect the city, the region, or society as a whole.
Even for the seemingly straightforward question “How much will it cost to build?” the answer is, during early estimation, far from certain. Con- struction costing is difficult. There is site work with excavation, backfill, and the driving of piles; tons of concrete for the deck and arches plus steel for reinforcement; and various kinds of labor needs, not to mention the construction equipment, from concrete mixers to truck-trailers, cranes, and crawler tractors. And there is the problem of labor and equipment productiv- ity: just how much can they accomplish per hour in the river mud and in the fog that often envelops the downtown? The answers are, furthermore, subject to price changes after the estimate is made, and not just because of overall inflation, but also because costs will fluctuate item by item, depend- ing on supply and demand.
For a shortcut, we can just estimate costs of a bridge by looking at the state’s average construction costs per square foot of deck area. These are quite reliable for short highway overpasses and can give a reasonable range with which to start, but are less reliable for predicting the unique specifics at a longer river crossing. Having to make an early case on whether the bridge makes sense or not, the initial estimator has the further problem that the engineering and environmental studies, which are still ongoing, may conclude that the structure has to be different than first expected. They may show that rocks meant to support the feet of the arches are less reli- able than expected, requiring deeper drilling, hence greater cost. Or seismic study can reveal a greater-than-expected chance of earthquakes, forcing a redesign that makes the bridge more resilient during shaking but also makes it more expensive.
Maintenance cost estimates offer a particular opportunity to make the bridge seem cheaper than it is. The authorities making the decision can easily be lulled into letting that happen. They are under pressure to solve downtown traffic problems at the river’s edge, but also under pressure to reduce city and state borrowing. They want to borrow just enough to solve the problem. When built, the bridge will be new; deferred maintenance won’t be noticed for years. Under pressure for results now, the authori- ties may not want to be bothered with the price a decade from now of repaving the deck that weathered too quickly or fixing a pier that was not built strongly enough to withstand minor boat collisions. By constructing the bridge with less durable materials now, it is possible to shift costs to those who will later be responsible for maintenance. To prevent such tricks, infrastructure decision makers must remember that the correct desideratum is the life-cycle cost, which includes the original project cost plus all costs to accrue over the years of service.
Measuring the costs is difficult enough. The measurement of benefits is even more so, because the methods are more susceptible to error and there is much legitimate disagreement over what a correct measure would be. Somehow the correct method would capture the benefit from a new convenient route across the river: the benefit of avoided travel time. As we shall see in the next chapter, it is possible to project how travel patterns would change with the new bridge, perhaps reducing travel delays. Among the many difficulties of such models is that a new, less congested road in itself gives an incentive to more people to drive—so the very existence of the bridge (not counting any increase in business activity, population, and employment) might put more vehicles on the city’s roads than were there before. However difficult the effort may be, the model should be able to measure the avoided travel time as the bridge’s primary benefit.
What adds greatly to the travel modeler’s burden is the second step in which she wants to figure out how much this avoided travel time is worth.
The answer includes the operating cost of the vehicle: gasoline, insurance, and depreciation. It also includes the cost of the travelers’ time, perhaps at their average salary rate, though of course such rates vary greatly. To answer, the analyst may well rely on standardized data, like that for Min- nesota in table 7.1.
Table 7.1. Recommended Standard Values for Vehicle Operation, State of Minnesota
Operating Costs $/mile
Auto $0.23
Truck $0.62
Value of Time $/person hour
Auto $10.46
Truck $19.39
Crash Values $/crash
Fatal $3,600,000
Injury Type A only $280,000
Injury Type B only $61,000
Injury Type C only $30,000
Property damage only $4,400
Source: Minnesota Department of Transportation, Benefit Cost Analysis Guidance, June 2005
Since shorter travel times reduce accidents, the conscientious mod- eler will also take a third step—that of estimating the value of avoided accidents. In view of current state accident rates, the analyst can estimate the reductions the new bridge will bring, and estimate averages for avoided vehicle damage (repair or replacement cost) and injuries (health care and lost income costs). Minnesota’s transportation department recommends that accidents be broken up into three classes of severity, each of which has a cost attributed to it (table 7.1). Accidents also cause deaths, to which it is con- ventional to attribute a life cost, in Minnesota’s case $3.6 million in 2005, though perhaps inflation has since increased Minnesotans’ nominal value.
The idea of attributing dollar value is likely to alarm those who first encounter it. Should we let the Minnesotan die if the cost of saving him is $4 million? Should children be valued less than adults, who are income earners, or more than old people, who have less left to live? Despite first impressions, however, modelers we have known have not been ghoulish.
They’re concerned, after all, about an abstract calculating standard called a statistical life. And they are, let credit be given, confronting a difficult problem in transportation planning: roads do generate accidents and deaths, while they also provide access and mobility, which are essential to life. By building a stretch of roads to a certain construction standard, which will inevitably pose a given statistical risk of future fatalities, and not spending the additional million for the finest paving and roadside barriers, the planner does implicitly give a monetary valuation to lives of future users, whether he admits it or not.
Now let us recap on the measurement of bridge benefit. We obtain, first, the number of travel hours that the new bridge allows vehicles to avoid; second, the average vehicle operating costs and traveler time value per avoided hour; and third, the average dollar value of avoided accidents per hour. Their product is a measure of bridge benefit.
The first step is clearly essential, but the second and third steps are worth questioning. While they get closer to the full benefit against which to weigh cost of the bridge, they pile ever more measurement assumptions on top of data that is already shaky. New inaccuracies compound old inaccura- cies, possibly making the complex answer less reliable than the basic answer.
It is even questionable whether accident data should be included at all. Adult travelers implicitly accept risk when they travel; that risk is incorporated into their valuation of whether they should travel or not. In America, highways and streets are subject to laws, liability requirements, and professional standards. If a road exposes travelers to a significantly greater risk than is standard, then the authorities have the obligation to fix the problem, by repairing it, slowing traffic, or closing it. Unless a road is being planned specifically to replace an unsafe road (greater road safety is a spe-
cific objective), it makes sense to assume, in accounting for benefits, that the new road’s safety benefits are simply proportional to travel time. So for a preliminary decision, it can be adequate just to complete the first of the three steps—to find only the cost per avoided travel hour.
For the sake of illustration, here are two simplified decision rules. First option: divide total bridge cost by the number of avoided travel hours. If cost per avoided travel hour is low compared to other, similar bridge projects, then the new bridge appears worthwhile. Second simple option: compare total bridge cost with total bridge revenues that travelers would reasonably be expected to pay through tolls. If tolls will pay for bridge engineering, construction, maintenance, and upkeep over the bridge lifetime, we have good grounds for saying it is worthwhile. But remember that these must be preliminary results, because we are so far considering only internal costs and benefits.