Almost every public-private partnership in Indian infrastructure begins with a projected rate of return. For urban rail, the figure is fourteen per cent. Since the Metro Rail Policy of 2017, a project that cannot show an economic internal rate of return of at least fourteen per cent will not be cleared. The shift from the earlier eight per cent financial test was sensible, since metros offer benefits such as shorter travel times, lower fuel consumption, and fewer road collisions, none of which are captured by fare-box revenues.
What follows is a long administrative process, and its length is important. A project report is prepared at the apex level, ridership is projected, and costs are assessed against revenues. If the numbers meet the required threshold, the project competes for a place in the fiscal budget. Expressions of interest are then issued, bids are processed, and a concessionaire is selected. This is followed by land acquisition and years of construction.
By the time the first fare is collected, the forecast that justified the project may be eight or ten years old. Yet it continues to remain in place as the standard against which the project is judged for the next three decades. That is what goes wrong after financial close. The problem is not a shortage of monthly returns, but the continued use of an old feasibility forecast as the benchmark for a project that is already in operation.
The distance between appraisal and operation
Consider Janmarg in Ahmedabad. The corridor opened in October 2009, while the feasibility work behind it dates back to the early years of that decade. Those projections described a city that has since changed significantly in terms of vehicle ownership, land use, and travel options. The issue is that, at the time of appraisal, both the project and the market it is expected to serve are still in the future. The forecasts are therefore based on assumptions about how the city and its travel patterns will develop over time.
A second difficulty is less often discussed. Feasibility numbers are calculated for the completed network, while the project itself is delivered in stages. A metro conceived as a two hundred kilometre network may initially open within just fifteen kilometres. Those fifteen kilometres cannot be expected to carry a proportionate share of the projected ridership because the usefulness of a metro system depends on the range of destinations it connects. Patronage does not increase in direct proportion to network size, yet the concessionaire is often held to the full-network ridership figure from the beginning.
What the Indian record shows
The evidence is difficult to ignore. The Delhi Airport Metro Express, the country’s first metro project under a concession, was expected to carry around 42,000 passengers a day by 2011. Ridership never crossed 20,000, and in June 2013, the concessionaire told the Delhi Metro Rail Corporation (DMRC) that it could no longer continue. Ridership later crossed 50,000 after fares were reduced, suggesting that the earlier shortfall was not simply a lack of demand, but was also linked to the pricing assumption.
Hyderabad Metro, which was briefly the largest metro concession in the world, is now being taken over by the Telangana government after years of losses. It is easy to attribute such outcomes to overly optimistic forecasts, and in some cases that may be justified. The more serious problem, however, arises after the project begins operations, when the original benchmark remains unchanged even though actual performance provides new evidence about demand, pricing, costs, and other operating conditions.
Counting what has been displaced
Ridership was never the only objective. These corridors are built to take cars and two-wheelers off the road, shorten journey times, and reduce collisions and emissions. After commissioning, the more useful question is not simply whether the projected ridership has materialised, but how much private vehicle use the corridor has displaced within its catchment area at its current stage of completion.
Answering this requires establishing a baseline before the service opens. This means measuring vehicle registrations, traffic volumes, modal shares, and journey times, and then measuring them again as each phase of the network is added. Janmarg’s early evaluations came close to doing this. About a fifth of its riders had shifted from two-wheelers, resulting in an estimated 200,000 vehicle-kilometres being avoided each day. These benefits remain relevant even when fare-box performance is weak in a particular year.
Reading a network through separate balance sheets
The complication that is least well handled is competition from within the public transport system itself. When Janmarg was planned Ahmedabad had no metro. It has one now, and some Janmarg passengers have shifted to it. Under the concession agreement, this is revenue lost. But from the perspective of the city’s wider transport objectives, the same shift can be seen as a success if the two systems together keep more vehicles off the road than either could have managed alone. The BRTS is not necessarily performing badly. Its numbers may simply be being measured against the wrong denominator.
Mumbai shows how far this can go. Suburban rail, the bus undertaking, several metro lines, and a monorail each operate under a separate board, with separate accounts and reporting. Four metro lines operated by three entities carry well under seven lakh passengers a day, against estimates approaching nineteen lakh. Each can report a shortfall against its own projection while the region as a whole may, in fact, be better served than any projection anticipated. The reverse should worry us more. All could meet their targets while private vehicle registrations continue to rise, and nobody within any one of these systems would necessarily notice.
Five changes worth making
Retire the forecast at commissioning -Keep it on record as the justification for the investment, but set fresh targets for each phase based on observed data from the project’s catchment area.
Report displacement alongside patronage -Measure trips avoided, vehicle-kilometres removed, time saved, casualties avoided, and emissions reduced. Establish the baseline before services begin. A counterfactual constructed afterwards is unlikely to be convincing.
Evaluate the network, not the entity - A metropolitan authority should consolidate ridership, modal share, and vehicle registrations across all operators and publish a single set of figures for which the city is accountable.
Distinguish between the causes of a shortfall -Operator underperformance, incomplete delivery, and an unanticipated externality are three different situations. They call respectively for a penalty, more time, and renegotiation. This is broadly in line with the approach pressed by the Kelkar Committee in 2015.
Review quarterly, with a named owner -Agreed thresholds should trigger an investigation while a problem can still be addressed, rather than waiting until it becomes an insolvency.
Feasibility appraisal answers one question: does this project deserve to be built? Evaluation after the project becomes operational has to answer another: what has changed in the city because of the project?
The problem with infrastructure PPPs is that the second question is often judged against the answer to the first. This is why sound projects can be written off as failures, while weak ones can go unexamined for years.
References
Ministry of Housing and Urban Affairs, Government of India, Metro Rail Policy, 2017 (revision of the approval criterion from an 8 per cent FIRR to a 14 per cent EIRR).
UNFCCC, Momentum for Change case study: Ahmedabad Bus Rapid Transit System, Janmarg - modal shift from two-wheelers and daily vehicle-kilometres avoided.
Business Standard, ‘What is wrong with PPP in India?’, 6 July 2013; and ‘Making PPPs work for metro-rails’, 23 April 2025 -Delhi Airport Metro Express patronage against projection.
Business Standard, ‘Funding woes of Metros: expansion continues, PPP model needs recalibration’, 12 February 2026 -Hyderabad Metro ridership, losses and transfer to the state government.
Scroll.in, ‘Urban immobility: why the Mumbai metro is failing to take off’, August 2025 - patronage against estimate, operator fragmentation and absence of common ticketing.
Department of Economic Affairs, Ministry of Finance, Report of the Committee on Revisiting and Revitalising the Public Private Partnership Model of Infrastructure Development (Chair: Dr V. Kelkar), November 2015.
