Leo Dobes, a critic of the project, has written a paper for ANU journal Agenda, in which he said the approach of including "wider economic benefits" in assessments of major infrastructure projects risked "white elephant projects".

The government's cost-benefit analysis of the Gungahlin tram line, released in October 2014, said the project had a benefit to cost ratio of 1.2, meaning for every $1 in spending it would bring $1.20 in benefits.

At the time, economist and former senior Treasury official David Hughes criticised the figure as a "fantasy", saying three-fifths of the benefits were "unsubstantiated claims about increased land values and productivity".

Without them, benefits were just half the cost, he said.

Capital Metro Minister Simon Corbell defended the analysis as conservative, responsible and prudent.

But Professor Dobes, an adjunct associate professor in the ANU Crawford School of Public Policy, said governments were increasingly using cost-benefit analyses that went well beyond conventional calculations.

Conventional cost-benefit analyses for transport projects included savings in travel time and fuel costs, as well as impacts on the environment and crashes.

But wider impacts counted the benefit to firms of cheaper access to customers and suppliers, and closer contact with similar businesses through sharing ideas, equipment and labour, as well as the benefit to workers of more jobs. People spent less time commuting so were thought to work longer hours, firms might relocate to the higher-productivity area, and presumed lower transport costs were counted as a benefit to businesses more widely. Increased economic activity in turn meant more tax flowing to the government.

In the case of Canberra, these wider economic benefits amounted to a substantial 20 per cent of the benefits calculated for the project.

But the use of "agglomeration benefits" could be problematic if not misleading, Prof Dobes said in a paper jointly written with Joanne Leung from the New Zealand Ministry of Transport. The "Achilles heel" was "the need to demonstrate causality".

"In particular, it is necessary to establish that changes in productivity or GDP are due to an increase in effective density generated by a specific transport project," they said, a link that was rarely, if ever, established. The temptation was to include unwarranted "uplift" factors, using examples from large cities like London, whether or not the experience there was relevant.

As to the claim that more workers travelling to Civic by tram would bring an increase in productivity because they swapped ideas and knowledge, Professor Dobes said the strength of that effect for a Gungahlin tram to the city was open to question.

While it could be appropriate to include increased income tax, the calculation should also include the losses from higher taxes and borrowing to fund such large projects, along with the reduced economic activity from the depressive effect of increased tax.

"In other words, funding a large project, other than by reducing government expenditures, will reduce GDP," he said.

"Failure to include both the positive and negative GDP/taxation effects of an infrastructure project can only result in a less-than-impartial analytical approach that will bias upward any estimation of the benefit."

And in the Canberra case, income tax collected at the national level and GST increases should not be counted as a benefit to residents of Canberra.