Discussing how to determine the optimal subsidy for public transport

The commentary piece “Determining the optimal subsidy for public transport” (The Straits Times, 20 July 2018) is a useful exercise for critical thinking.

First, a logical principle: “To determine optimal subsidies for a transport system, we need to identify the justifications for subsidising the system.” This statement is in line with the key principle in critical thinking: that every position must be supported by a good argument.

A “commonly invoked argument” is described: “Public transport subsidies benefit the poor since the poor are more likely than the rich to use public transport.” But transport is a “small portion” of the poor’s household expenditure. Hence, this argument is rejected.

The “main argument” for public transport subsidies is an “economy of scale” argument. Subsidies increases ridership, which reduces the average cost of providing public transport services (total operation cost divided by more users). More ridership also means more frequent services, which reduces commuters’ waiting time, which encourages more commuters, and so on. This is known as the Mohring Effect. Economists Ian Parry and Kenneth Small say this argument is used to justify subsidies in London, Washington and Los Angeles – especially during off-peak times.

Public transport subsidies encourage people to switch from cars to public transport. This reduces road congestion, pollution and accident costs. Professor Stef Proost and Dr Kurt Van Dender find this “congestion externality” consideration to be “quantitatively more important” than the “economy of scale” factors in justifying subsidies in Brussels and London.

On congestion costs: Singapore imposes a congestion tax via the Electronic Road Pricing system. Associate Professor Leonardo Basso and Assistant Professor Hugo Silva find that this lessens the congestion reduction benefit of public transport subsidy. The writer here adds: “Nevertheless, the benefit of subsidies on reduced congestion on Singapore’s roads is likely to be a non-negligible amount”. [No argument is presented to support the subsidiary claim of “non-negligible amount”.]

On pollution costs: Economists found that “the opening of a new subway network decreases particulate concentrations by about five percent in the 10km disk around the city centre”.

So subsidies provide the following benefits: lower average cost to service provider, lower waiting time for service, less road congestion, less pollution, less accident costs.

We are next introduced to some “arguments against public transport subsidies”.

Public transport subsidies increase ridership – which increases crowding, and hence increases discomfort and extends boarding and alighting times. These are the main costs of subsidies: discomfort, longer boarding and alighting times. Since boarding and alighting times are independent of distance travelled, subsidies should be a function of distance travelled less a fixed amount for boarding and alighting delays. This should result in “commuters traveling short distances with a smaller fraction of their fares subsidised”. This is a subsidiary point – relating to the quantum of individual subsidy, not the fact of general subsidy.

Money spent on public transport subsidies means money not spent on “other purposes, such as education, healthcare, or welfare.” This is the “opportunity cost” of public transport subsidies. Professor Proost and Dr Van Dender find that for Brussels, this opportunity cost “wipes out” the economy of scale benefits.

“Combined, these reasons determine the optimal subsidy for public transport operations.”

This is the key sentence. It tells us that the considerations relevant to public transport subsidy are: average cost to service provider, waiting time for service, road congestion, pollution, accident costs, passenger discomfort, boarding and alighting times, money available for other purposes. It is interesting that the actual money value of subsidy is not mentioned.

This list of considerations suggest that the justification argument is utilitarian in nature. However, this is not explicitly stated, so we must not presume so.

Dr Parry and Professor Small find optimal subsidies “to be very large” (mostly over 90 percent of operation costs) in Washington, London and Lost Angeles. Prof Proost and Dr Van Dender “find the optimal subsidy for Brussels during peak periods should be close to zero.” Prof Basso and Prof Silva find “optimal subsidies in Santiago to be about 55 percent of operation costs”.

The Brussels finding is clearly a recommendation, since it says “should be”. The statements regarding the other cities use the phrase “to be”, which precedes a fact (meaning that recommendations were implemented). The question that interests us is: How were these optimal subsidies determined?

The differences “stem from the various cities’ characteristics, including congestion levels, commuters’ preferences for different transport modes, and commuter valuation of travel time”. This is just a summary (as we have also done above) of the relevant considerations.

The essay concludes: “There is no hard and fast rule for optimal subsidies that Singapore can easily adopt. Instead, the Government will have to determine optimal levels by assessing the merits of each consideration in the Singapore context.”

Here is the remaining problem: The “considerations” may be as listed above, but we have not been told how to use them to determine the optimal (value of the) subsidy. So, referring to the opening logical principle: the “justifications for subsidising the system” have not been fully identified. Worse, what does the unequivocal declaration that “there is no hard and fast rule for optimal subsidies” portend for any further discussion of this matter?

END