Why Congestion Calls for a User Charge
When more people use a congestible facility, each user may face additional delay. A driver entering a busy road, for example, adds time not only to their own trip but also to trips made by other drivers. The entrant considers the delay they personally experience, but may not account for the extra delay imposed on everyone else. That gap is a marginal external congestion cost.
An efficient access charge makes users face this additional cost when deciding whether to use the facility. The aim is not to eliminate use: it is to align users’ private decisions with the total costs of their use. The analysis below uses a simple model of usage and delay. The treatment is aligned with the SOA VEE Economics (Validation by Educational Experience) syllabus.
The definition assumes that all users experience the same delay, value time at the same rate, and treat the overall usage level as fixed when making an individual decision. The model also assumes \(t'(q)\geq 0\): delay does not decrease as use increases. The toll is considered in addition to any resource cost of providing or operating the facility. It does not apply unchanged when there are other external costs, such as pollution, or when users have different delay costs; those require additional terms or a more detailed model.
Private Cost, Social Cost, and the Efficient Charge
Let \(B(q)\) be the total benefit from \(q\) uses, so that \(B'(q)\) is the marginal benefit of one additional use. Let \(c\) be a constant marginal resource cost per use, such as an operating cost. Without a toll, an individual user considers the direct resource cost \(c\) and their own delay cost \(wt(q)\). Society must also account for the delay imposed on other users.
The social marginal cost includes both the additional user’s own delay and the delay their entry causes others. It is the relevant cost for deciding the efficient level of use. The access charge should equal the marginal external congestion cost at that efficient level—not the whole social marginal cost, since users already experience and account for their own delay.
The condition is an interior solution: it assumes the efficient quantity is positive and not constrained by a capacity limit or another boundary. If use is restricted to a range, the efficient choice may instead lie at a boundary. The toll must also be calculated using delay and the value of time in consistent units.
Why the Formula Works
Total surplus is total benefit minus real resource and delay costs. A toll payment is not itself a resource cost in this model: it transfers money from users to the facility operator or public authority. It can affect the decision to use the facility, but it should not be subtracted as though it destroyed resources.
For an interior efficient choice, marginal surplus is zero. Differentiating total surplus gives the following condition:
Now consider an individual choosing whether to make one more use while taking the overall delay level \(t(q)\) as given. With a toll \(\tau\), the user compares marginal benefit with private marginal cost \(c+wt(q)+\tau\). Setting the toll to \(wqt'(q)\) makes that private cost equal to \(c+w(t(q)+qt'(q))\), exactly the social marginal cost in the efficient condition. This is why the charge corrects the incentive to use the congestible facility.
Multiply delay per use \(t(q)\) by the value of time \(w\).
Multiply the delay cost per user by the number of users: \(D(q)=wqt(q)\).
Equate marginal benefit with marginal resource cost plus the derivative of total delay cost.
Evaluate \(wqt'(q)\) at the efficient usage level and check that the user’s decision condition holds.
Worked Applications
Worked Example: A Linear Delay Function
A hypothetical access road has inverse marginal benefit \(B'(q)=120-2q\) dollars per trip and marginal resource cost \(c=\$20\) per trip. Delay per trip is \(t(q)=0.01q\) hours. Users value time at \(w=\$100\) per hour. Find the efficient number of trips and the efficient charge per trip.
Method. First find the delay cost each user bears and the marginal external cost, then solve the social efficiency condition. This avoids treating one user’s own delay as an external cost.
A user’s own delay cost is \(wt(q)=100(0.01q)=q\) dollars per trip. Total delay cost is \(D(q)=wqt(q)=100q(0.01q)=q^2\) dollars. Therefore \(D'(q)=2q\). Equivalently, the external component is \(wqt'(q)=100q(0.01)=q\), in addition to the user’s own delay cost \(q\).
The efficient condition is \(120-2q=20+2q\). Rearranging gives \(100=4q\), so \(q^*=25\) trips. The efficient charge is \(\tau^*=wq^*t'(q^*)=100(25)(0.01)=\$25.00\) per trip.
Check. At 25 trips, marginal benefit is \(120-2(25)=\$70\). A user facing the charge pays a marginal resource cost of \(\$20\), their own delay cost of \(\$25\), and a toll of \(\$25\), for a total of \(\$70\). The private decision condition matches the efficient one.
Worked Example: Converting Delay Units into a Toll
A hypothetical recreation route has inverse marginal benefit \(B'(q)=90-q\) dollars per visit and marginal resource cost \(c=\$10\) per visit. Each visit adds to delay according to \(t(q)=0.005q\) hours per visit. The value of time is \(w=\$40\) per hour. Find efficient usage and the access charge. Report the quantity to four decimals and the charge to the nearest cent.
Method. Convert hours of delay into dollars first, then distinguish the delay cost paid by the user from the cost imposed on other users. This avoids a time-unit or double-counting error.
A user’s own delay cost is \(wt(q)=40(0.005q)=0.2q\) dollars. Total delay cost is \(D(q)=40q(0.005q)=0.2q^2\) dollars, so its marginal cost is \(D'(q)=0.4q\). Of this, \(0.2q\) is the user’s own delay cost and \(0.2q=wqt'(q)\) is the external cost.
Set marginal benefit equal to social marginal cost: \(90-q=10+0.4q\). Thus \(80=1.4q\), giving \(q^*=400/7=57.1429\) visits to four decimals. The charge is \(\tau^*=0.2q^*=80/7=\$11.43\) per visit to the nearest cent.
Check. At \(q^*=400/7\), marginal benefit is \(90-400/7=230/7\), or about \(\$32.86\). The charged user’s marginal cost is \(10+80/7+80/7=230/7\): resource cost, own delay cost, and toll, respectively. The rounded charge is consistent with the unrounded calculation.
Worked Example: Nonlinear Congestion
A hypothetical facility has inverse marginal benefit \(B'(q)=200-2q\) dollars per use and marginal resource cost \(c=\$40\) per use. Delay is \(t(q)=0.001q^2\) hours per use, and the value of time is \(w=\$100\) per hour. Determine efficient use and the efficient charge.
Method. Differentiate total delay cost, or equivalently calculate the user’s own delay and the external delay separately. For a nonlinear delay function, the marginal external cost changes with usage.
The user’s own delay cost is \(wt(q)=100(0.001q^2)=0.1q^2\) dollars. Total delay cost is \(D(q)=100q(0.001q^2)=0.1q^3\), so \(D'(q)=0.3q^2\). The external component is \(wqt'(q)=100q(0.002q)=0.2q^2\); the remaining \(0.1q^2\) is the user’s own delay.
The efficiency condition becomes \(200-2q=40+0.3q^2\). At \(q=20\), the left side is \(200-40=160\), and the right side is \(40+0.3(400)=160\). Thus \(q^*=20\) uses. The positive solution is unique because marginal benefit falls with \(q\), while social marginal cost rises with \(q\). The efficient charge is \(\tau^*=100(20)(0.002(20))=\$80.00\) per use.
Check. At 20 uses, the user’s own delay cost is \(0.1(400)=\$40\). Resource cost, own delay, and toll sum to \(40+40+80=\$160\), equal to marginal benefit \(200-2(20)=\$160\).
Common Mistake: Charging the Wrong Cost
A frequent error is to charge each user the entire social marginal congestion cost, or the total delay cost, instead of the marginal external cost. The first approach charges users for their own delay again; the second confuses the total accumulated cost with the cost caused by one additional user.
In the linear-road example, at the efficient quantity \(q=25\), own delay cost is \(\$25\) per trip and external marginal cost is \(\$25\) per trip. The incorrect charge of the entire congestion component of social marginal cost would be \(\$50\); it counts the user’s own \(\$25\) delay a second time. The total delay cost at that usage is \(D(25)=25^2=\$625\), which is not the appropriate per-trip toll either. The correct charge is \(wqt'(q)=\$25\) per trip.
Check Your Understanding
For each problem, identify the private delay cost where needed, find the relevant marginal external cost, and use consistent units.
- A hypothetical trail has \(q=30\) users, delay \(t(q)=0.004q\) hours per user, and a value of time of \(\$60\) per hour. Calculate the marginal external congestion charge at this usage. (Answer: \(\$7.20\) per user.)
- A facility has inverse marginal benefit \(100-q\), resource cost \(\$10\) per use, delay \(t(q)=0.01q\) hours, and value of time \(\$50\) per hour. Find efficient use and the charge per use. (Answer: \(q^*=45\) uses; charge \(=\$22.50\).)
- A hypothetical crossing has inverse marginal benefit \(135-1.5q\), resource cost \(\$15\) per use, delay \(t(q)=0.002q\) hours, and value of time \(\$100\) per hour. Find efficient use and the charge, giving both to four decimals. (Answer: \(q^*=63.1579\) uses; charge \(=\$12.6316\) per use.)
- A facility has inverse marginal benefit \(100-q\), resource cost \(\$20\) per use, delay \(t(q)=0.001q^2\) hours, and value of time \(\$100\) per hour. Find efficient use and the nonlinear congestion charge. (Answer: \(q^*=10\) uses; charge \(=\$20.00\) per use.)
- A hypothetical route has inverse marginal benefit \(150-2q\), resource cost \(\$30\) per trip, delay \(t(q)=0.002q\) hours, and value of time \(\$50\) per hour. Find efficient use and the efficient charge to four decimals. (Answer: \(q^*=600/11=54.5455\) trips; charge \(=\$5.4545\) per trip.)