Public-private partnerships in the US

A public-private partnership, or P3, goes further than any other delivery method in handing responsibility to the private sector. Instead of buying construction, the public owner buys a service: a hospital that is available and maintained to a stated standard for thirty years, a highway lane that is open and in good condition, a water plant that treats a given flow to a given quality. A private consortium designs and builds the asset, usually finances much of it, and often maintains or operates it for decades. The public owner pays mainly through performance-based payments once the asset is in service, and deducts money when it falls short.

The United States has used P3s more selectively than Canada, mostly for toll roads, managed lanes, airport terminals, transit lines and a growing number of social infrastructure projects such as university campuses, courthouses and civic buildings. According to the Federal Highway Administration, 40 states plus the District of Columbia and Puerto Rico had enabling statutes for transportation P3s as of October 2025, but the scope of those laws varies widely, and much US infrastructure is financed instead through tax-exempt municipal bonds.

How a P3 is structured

The private side forms a special purpose company, often called Project Co, owned by equity investors, which may include the construction company, the facility manager and infrastructure funds. Project Co signs the project agreement with the public authority and immediately passes most of its obligations down: design and construction to a design-build joint venture under a fixed-price, date-certain contract, and maintenance or operations to a facility management or operations company under a long-term services contract. Project Co borrows senior debt from banks or bond investors to fund construction, together with the equity, and repays both from the payments it receives from the authority over the concession.

How the private partner is paid

Most social infrastructure P3s, and many highway and transit P3s, use availability payments. The authority pays nothing, or only milestone payments, during construction. From substantial completion it pays a monthly or annual service payment for the life of the concession, which covers repayment of debt and equity, maintenance and life-cycle costs, and the private partner's return. The payment is reduced by deductions if parts of the asset are unavailable, such as a closed operating room or lane, or if services fall below standard, such as slow repairs or poor cleaning. User-pay concessions instead let the private partner collect tolls or fees, and it bears the risk that demand falls short of forecast.

# Monthly availability payment with deductions for a hospital P3 (illustrative)
monthly_payment = 4_200_000       # full monthly service payment
areas = {  # area: (weight in payment, hours unavailable this month, hours in month)
    "Operating suites":   (0.20, 18, 720),
    "Inpatient units":    (0.35, 0, 720),
    "Emergency":          (0.15, 6, 720),
    "Clinics and offices": (0.30, 40, 720),
}
availability_deduction = 0
for area, (w, hrs, total) in areas.items():
    d = monthly_payment * w * hrs / total * 2.0      # unavailability deducted at twice the pro-rata rate
    availability_deduction += d
    print(f"{area:20} unavailable {hrs:3} h  deduction {d:>10,.0f}")
service_failures = 14 * 1_500                        # 14 late responses at $1,500 each
total_ded = availability_deduction + service_failures
print(f"Service failure deductions {service_failures:,}")
print(f"Payment this month {monthly_payment - total_ded:,.0f} (deductions {total_ded / monthly_payment:.2%})")

Operating suites     unavailable  18 h  deduction     42,000
Inpatient units      unavailable   0 h  deduction          0
Emergency            unavailable   6 h  deduction     10,500
Clinics and offices  unavailable  40 h  deduction    140,000
Service failure deductions 21,000
Payment this month 3,986,500 (deductions 5.08%)

This month the deductions take about 5 % of the payment, mostly for the operating suites and clinics. Deductions are calibrated so that poor performance costs the private partner more than it would save by cutting corners, and repeated failures can lead to default. Because the payment repays the lenders, lenders watch performance closely and can step in to replace a failing contractor or facility manager. That lender oversight is one of the main reasons owners use private finance: the people whose money is at risk have strong reasons to make sure the project is built on time and maintained.

Value for money

A P3 almost always costs more to finance than public borrowing, because private debt and equity carry higher interest and returns than government bonds. It also has higher transaction costs for lawyers, advisers and bid preparation. P3s are justified, if at all, by transferring risks the public sector would otherwise bear and price poorly: construction cost overruns, delay, poor maintenance and life-cycle neglect. Public agencies test this with a value-for-money assessment that compares two estimates in today's dollars: a public sector comparator, which is the risk-adjusted cost of delivering the project conventionally, and the risk-adjusted cost of the P3.

# Simplified value-for-money comparison in present value (illustrative, $ millions)
def pv(cashflows, r):
    return sum(cf / (1 + r) ** t for t, cf in enumerate(cashflows))
r = 0.045                     # public discount rate used for both options
years = 33                    # 3 years of construction + 30 years of operation
# Public sector comparator: construction paid as built, maintenance paid annually, plus risks retained by the public owner
psc = [350, 350, 300] + [9.0] * 30
psc_risk = [60, 60, 50] + [3.0] * 30      # expected overruns, delay and life-cycle neglect
# P3: nothing during construction; annual service payment covers capital, finance, maintenance and return
p3 = [0, 0, 0] + [80.0] * 30
p3_retained = [5, 5, 5] + [0.8] * 30      # risks the owner keeps under the P3 (changes, some site risks)
p3_transaction = [12, 4, 2] + [0.3] * 30  # owner advisers and contract management
a = pv(psc, r) + pv(psc_risk, r)
b = pv(p3, r) + pv(p3_retained, r) + pv(p3_transaction, r)
print(f"Public sector comparator, risk-adjusted PV: {a:,.0f}")
print(f"P3, risk-adjusted PV:                       {b:,.0f}")
print(f"Value for money: {a - b:,.0f} ({(a - b) / a:.1%} of PSC)")

Public sector comparator, risk-adjusted PV: 1,302
P3, risk-adjusted PV:                       1,242
Value for money: 60 (4.6% of PSC)

In this invented example the P3 shows value for money of about 4.6 %, less than the transferred risk in the comparator: without that risk adjustment, conventional delivery would look cheaper. That is typical of the method, and also its main point of criticism: the result depends heavily on how the retained risks in the comparator are estimated and on the discount rate chosen. Small changes in either can reverse the conclusion.

Value-for-money analysis is less standardized in the United States. Some states and agencies publish value-for-money reports for P3s, and the Federal Highway Administration offers guidance and tools for P3 assessment, but there is no single national method. Critics argue that tax-exempt public borrowing is so cheap in the US that P3 finance rarely makes sense unless risk transfer is large or the project could not otherwise be funded.

Risks transferred and retained

P3 contracts are long and detailed because they must anticipate decades of events. The authority keeps the right to change the asset or the services through a variation procedure, with compensation; it can terminate for convenience at a defined price; and Project Co can be terminated for default, with lenders given time to fix problems first. At the end of the concession the asset is handed back in a defined condition, often verified by surveys in the last years of the term.

The procurement process

P3 procurement is long and expensive. A typical process includes a business case and value-for-money assessment; a request for qualifications to shortlist three teams; a request for proposals lasting six to twelve months, with commercially confidential meetings in which bidders comment on the draft project agreement; evaluation of technical and financial proposals; selection of a preferred proponent; and financial close, when the project agreement and financing are signed and the interest rate is fixed. Owners often pay unsuccessful shortlisted bidders a fee, because bids can cost each team several million dollars.

US P3 procurements are often supported by federal credit programs. The Transportation Infrastructure Finance and Innovation Act (TIFIA) program provides low-cost federal loans and loan guarantees for transportation projects, including P3s, and private activity bonds give P3 projects access to tax-exempt debt within a national volume cap. The current surface transportation authorization under the Infrastructure Investment and Jobs Act expired on 30 September 2026; a short-term extension to 11 December 2026 was moving through Congress in August 2026, so check the current status of federal transportation programs before relying on them.

Lessons from troubled P3s

P3s promise date-certain, fixed-price delivery, but some of the most visible project failures in both countries have been P3s, which shows the limits of risk transfer.

In Maryland, the Purple Line light rail, an availability-payment DBFOM, faced delays and disputes that led its original design-build contractor to withdraw in 2020; the state and concessionaire negotiated a settlement and brought in a new contractor, with large increases in cost and a delayed opening. Other US P3s, such as some early toll-road concessions, ran into financial trouble when traffic fell short of forecasts. Successful projects, such as several availability-payment highways and university campus P3s, show that the model can work when risks are well understood and the market can price them.

The common lessons are that complex integration risk, especially for transit systems, is hard for any one party to carry; that bidders price large risks heavily or misprice them; and that when a private partner fails on an essential public service, the public owner cannot simply walk away. Lesson 8 looks at the progressive and collaborative models that grew out of these experiences.