Week 6 — Performance Pricing vs Performance Time Limits: What Shoup Got Inverted

This is the most substantive correction of Shoup we will make in this series. It builds on the externality argument we sketched in Week 2 — the one that says performance pricing optimises for the deepest pockets rather than the widest prosperity. Here we develop it into a positive proposition: time limits, not prices, are the primary curb-management lever for community welfare. Pricing is the secondary lever.

The argument is simple, but it requires noticing something most American cities already do — and have always done — at every part of the curb except the part Shoup wrote about.

Cities already manage every other curb function this way

Walk down any block of an American downtown and look at how the curb is divided. There is a fire hydrant at one end and a bus stop at the other. In between there is a loading zone, a taxi stand, a few 2-hour parking spaces, an ADA space, an electric-vehicle charging stall, possibly a passenger drop-off zone in front of the office tower, and on the side street a commercial-loading-only segment.

Six of these seven uses are managed by time and use. Only one is managed primarily by price.

• Fire hydrants aren’t auctioned to the highest bidder — they’re use-prohibited to all parking.
• Bus stops aren’t priced — they’re use-restricted to buses.
• Loading zones aren’t sold to the wealthiest delivery company — they’re time-limited (typically 30 minutes) so every business on the block can receive deliveries.
• Taxi stands aren’t auctioned to the cabbie willing to pay most for permanent positioning — they’re a rotating queue.
• ADA spaces aren’t priced higher — they’re use-restricted to permit holders.
• Drop-off zones at schools and offices aren’t sold to wealthy parents or executives — they’re “kiss-and-go” with 1-to-2-minute limits.
• EV charging stalls in busy areas aren’t priced to the customer willing to charge for the longest — they’re often time-limited (e.g., 1-hour maximum) so more EVs can use the public infrastructure.

In every one of these cases, the city has intuited correctly that pricing cannot do the work. The lever is time and use, not price. The pricing-economics framework — “let willingness to pay decide” — is unanimously rejected at the fire hydrant, at the bus stop, at the loading zone, at the taxi stand, at the ADA space, at the drop-off curb, and at the EV charger.

It’s only retail parking — the section of curb between the fire hydrant at one end and the bus stop at the other — that we still manage with pricing as the primary tool.

That inconsistency is not justified by anything in the structure of curb space itself. It is the inheritance of Shoup’s framework. Two decades after The High Cost of Free Parking convinced the parking-economics literature that demand-responsive pricing was the answer, retail parking became the one curb function we manage with auction logic instead of with the time-and-use logic the city has always applied to every other curb segment. The reason for the inconsistency is intellectual, not practical.

The question this post asks: if every other curb function works correctly under time-and-use rationing, what is special about retail parking that requires pricing instead?

The answer, on close inspection, is nothing. Retail parking is a scarce time-limited public asset where multiple users want access, the access enables commerce at adjacent businesses, and the community externality scales with rotation. So does the loading zone. So does the drop-off zone. So does the EV charger. So does the taxi stand. They are managed correctly. Retail parking is managed by an economist’s framework that asks the wrong question.

The arithmetic — pure pricing vs performance time limits

Consider a representative 15-space retail-heavy block face on a midsized-downtown corridor. 12 hours of operation. Roughly 180 motorists want to park there each day. Our research and operations-data analysis, calibrated against real-world outcomes from leading municipal parking programs, compares four candidate policies on the same physical curb.

Daily averages on a representative retail-heavy block:

B1 vs A2 (auto-enforced time limits vs realistic performance pricing — the comparison most cities actually face):

• 2.03× the curb-adjacent commerce
• 2.04× the city revenue
• 47% more customers served

B1 vs A1 (auto-enforced time limits vs steel-manned Shoup pricing):

• 1.85× the curb-adjacent commerce
• 1.42× the total city commerce
• 1.90× the city revenue
• 38% more customers served

Both comparisons are decisive in the same direction. The A1-vs-B1 number is what theorists argue about; the A2-vs-B1 number is what city managers actually see when they switch policy regimes.

Annualised across 330 operating days, the B1-over-A1 uplift per retail-heavy block face is approximately:

• ~$29K/space-year in curb-adjacent commerce
• ~$28K/space-year in multiplier-adjusted community commerce
• ~$18K/space-year in city revenue (meter + citation + sales tax)
• ~$46K/space-year in total community value uplift

For a midsized downtown of ~1,000 managed retail-heavy spaces, that’s ~$46M/year in community value uplift. For a larger downtown of ~5,000 managed spaces, the order of magnitude scales accordingly. Build the model with your own city’s inputs and the conclusion doesn’t move.

This is the gap Shoup’s framework produces by choosing pricing as the primary lever instead of time limits. It is not a small gap.

Three failures, all from the same root

Pricing as the primary lever fails for three reasons, and all three trace back to the same misconception — that the user willing to pay the most is the use the city most wants to encourage.

1. Pricing rations by purse, not by intended use. The 4-hour interviewee at the courthouse has higher willingness to pay than the 30-minute haircut customer because the interviewee has no alternative. They will outbid every retail shopper, every time. Pricing systematically delivers the curb to the user with the worst alternatives — and the lowest social-value contribution.

2. Pricing requires the price to be visible at the moment of decision. A driver in motion at 20 mph cannot read the rate at the corner sign 75 feet behind them. They commit to a space before they know what it costs. Demand-responsive pricing produces price differentials between blocks that are real on the city’s pricing dashboard but invisible to the motorist.

3. Pricing assumes an off-street alternative. Shoup’s framework relies on long-stayers self-routing to garages priced at market rate. In real downtowns, garages are often scarce, more expensive than the meter for 2 to 4 hour stays, time-taxed by entry/exit costs, or simply not nearby. The rate the city would have to charge to deter the long-stayer at the curb — pure pricing, no time limit — is on the order of $50/hour. That price is politically impossible. Time limits achieve the same allocation outcome at $2/hour.

Even weak enforcement of time limits still beats performance pricing

The argument so far has compared performance pricing against performance time limits with strong (automated, sensor-driven, 80% capture) enforcement. The harder question is what happens to cities that adopt time limits but can’t yet fund automated enforcement. Manual officer patrols cap out at roughly 7% capture on heavy beats — a long-stayer doing the math sees a $5 expected ticket cost against a $30 garage and rationally stays at the curb. The time limit becomes a paper tiger.

This is a real concern. But the answer the research returns is unambiguous: performance time limits, even at manual-enforcement capture rates, deliver more customers served and more curb-adjacent commerce than performance pricing in either form.

Specifically, manual-enforced time limits (B2) versus realistic performance pricing (A2) — the version of pricing most cities actually run — produces +18% on cars served, +30% on curb-adjacent commerce, +30% on total city commerce, with city revenue roughly tied. Against steel-manned Shoup pricing (A1), B2 still produces +11% on cars served, +18% on curb commerce, +10% on total commerce, with city revenue trailing by ~6%.

The asymmetry is exactly what the externality argument predicts. Community-value metrics — customers served, commerce on the block, sales tax to the city — favour time limits even without enforcement. City-revenue metrics — meter + citation + sales tax to the treasury — require enforcement to also favour time limits. The reason is that pricing extracts more meter revenue from each parker than time limits do; only when automated enforcement adds citation revenue does the treasury line also tilt to time limits.

The procurement implication is two-step. First, switch from performance pricing to performance time limits even if automated enforcement isn’t immediately affordable — the community-value gain is real at any enforcement level, and merchants on the block will feel it within a quarter. Second, graduate to automated enforcement when funding allows, doubling the uplift the policy switch already produced. Auto enforcement is not a precondition for adopting time limits. It is the multiplier that turns “still better than pricing” into “transformative.”

What performance time limits look like in practice

The same data feed Shoup proposed for pricing — sensor-based real-time per-space occupancy — can drive either of two symmetric policies aimed at his 85% occupancy target:

• Performance pricing: Block X at 95% occupancy → raise rate by 25¢. Block Y at 60% → drop rate by 25¢.
• Performance time limits: Block X at 95% occupancy → shorten time limit from 2 hours to 1 hour. Block Y at 60% → extend time limit from 1 hour to 2 hours.

Both hit 85% occupancy. They have radically different distributional consequences, and — as the arithmetic above shows — dramatically different community-welfare outcomes. Performance time limits is the symmetric proposal Shoup never explored, and it is structurally superior.

A city with sensor-equipped curbs can implement performance time limits today. The same dashboard that tells SFpark to raise rates can tell a forward-thinking city to shorten time limits. The procurement implications follow directly:

• Specify time-limit policy as the primary curb-management lever, pricing as the secondary lever.
• Use sensor data to drive time-limit calibration, not just rate calibration.
• Design decision-point hardware at the space to display time limit as the dominant information, with rate secondary — because the time limit is what determines whether the motorist can use the space at all.
• Enforce overstays of clearly-displayed time limits, not the random patrol-driven enforcement of complex rate variations.

The lineage

This is a substantive correction of Shoup, not just an extension. It does not undo what he built. It does say that the parking-reform conversation of the next decade has to broaden from “what is the right price?” to “what is the right use, and how do we allocate the curb to it?” The right price is a one-dimensional question. The right use is a multi-dimensional question that requires the city to think about its merchants, its workers, its residents, and its visitors as a coherent public, not as competing private demands.

Shoup is right that the curb is undervalued. He is right that pricing reform is part of the answer. He is wrong about which lever does the heavy lifting. Time limits do — and the data and technology to set them well now exist. The question for any city today is not whether to charge for the curb, but whether to charge by intended use (the same logic the city already applies to every other curb function) or by willingness to pay (auction logic, the inheritance of Shoup’s framework). Two decades after The High Cost of Free Parking, the empirical record makes clear which one the city — and the merchants on the block — should choose.

Next week: Forward from here — the curb that Shoup would recognise, plus the layers he could not see yet.