The above publication -

a)      undertook in 2009 a mathematical cost-benefit analysis ("CBA") of the National Broadband Network; and

b)      opined on the scope for significant economic mismanagement across a range of project types when the disciplines of CBA are not rigidly observed.

The above publication (35 pgs) draws upon a longer publication (73 pgs) written by the same authors titled "The Social Losses from Inefficient Infrastructure Projects: Recent Australian Experience" - Henry Ergas and Alex Robson  -  Productivity Commission Round Table - Strengthening Evidence-Based Policy in the Australian Federation - August 2009 that additionally presented a cost-benefit analysis of the East-West rail (EWR) project in Victoria, which broadly aimed at improving the rail links between Melbourne, Geelong and the regions to Melbourne’s west, that was recommended by Infrastructure Australia for immediate funding.

Below is a second extract from Evaluating major infrastructure projects:  how robust are our processes?

"Cost–benefit analysis can be viewed from four complementary perspectives.

First, cost–benefit analysis is related to (though not identical with) the basic equi-marginal condition for overall efficiency in resource allocation. Thus, given a cardinally measurable objective function and perfect knowledge of the effect on welfare of any decision, it is a condition of an optimal set of decisions that the marginal dollar of public expenditure has a benefit equal to that of the marginal dollar of private expenditure (thus assuring that the overall level of public expenditure is optimal) and that the benefit of a marginal dollar of public expenditure is equalised across programs, projects and project elements. Because cost–benefit analysis aggregates willingness to pay across agents with different marginal valuations of income, it is not a perfect measure of underlying utility (and hence cannot be treated as an ideal social welfare function); nonetheless, taking that important caveat as given, one would at least question whether a set of public decisions was optimal if it did not maximise the aggregate benefits obtainable for given aggregate costs or minimise the aggregate costs required to obtain a given aggregate benefit, in each case measured using cost–benefit analysis.

Second, set against the backdrop of a given portfolio of projects, cost–benefit analysis can be used to evaluate whether one or more public projects should be added to or removed from that portfolio. In other words, cost–benefit analysis is a tool that can be used to assess whether wealth (the difference between the aggregate valuation of outcomes and the cost of obtaining those outcomes) would be increased by the decision to, say, proceed with a particular project, compared to the relevant alternatives (which may involve doing nothing, deferring or otherwise varying the project, or proceeding with an alternative project). 

Third, cost–benefit analysis is an instrument that the principals in public sector governance can use to improve the decisions taken by their agents, and to enhance their supervision of those agents (see, for example, Adler and Posner 2006, Posner 2001 and Spence and Cross 2000). As a result, the requirement to carefully assess and report the costs and benefits of decisions can improve the quality of decision-making and reduce the information asymmetry between principals and agents. In doing so, it can:

•         help reduce the risk of ‘capture’, in which the agent’s decisions, rather than reflecting the interests of the principal, come to be determined either by the agenda of self-interested third parties or by the agent’s own interests and aspirations

•         help correct ‘policy bias’, which is a situation in which those working in an agency have policy commitments that differ from (and may undermine) those of the public

•         help overcome ‘shirking’, in which agents do not exercise as much diligence in taking decisions as would be warranted

•         help disclose and correct the cognitive biases that affect decision-making

•         increase consistency in decision-making, both by standardising the information base on which decisions are taken and by highlighting anomalies, such as differences between project appraisals in the valuation of common elements

•         improve performance auditing and accountability by providing a standardised ex ante statement of key expected values for costs and benefits.

Ultimately, all of these effects mean that cost–benefit analysis is never merely an analytical tool; rather, as Aaron Wildavsky (1966) emphasised many years ago, it is inevitably an instrument in shaping bureaucratic structure and process, both within each public sector body and between that body, the other elements of the public sector with which it interacts, and the wider political system.

Fourth and last, cost–benefit analysis can be an anchoring device that reduces undesirable policy instability."

•         that all cost–benefit analyses be disclosed to the public well in advance of Financial Close which would -
            *        enable community consultation; and
            *        highlight projects that had not been subjected to economic project evaluation; and

•         that obligate comprehensive systematic auditing of cost–benefit analyses, both at the stage of the financing decision and post-project completion, because there is generally little audit currently, and in many instances, cost–benefit analyses have not been updated or properly archived after the initial ‘go/no go’ decision is taken.

"6.4  Conclusions

Infrastructure investment is a cost, not a benefit; a means, not an end. This proposition, which is obvious to economists, is as utterly alien to contemporary Australian politicians as the notion of comparative advantage was to their predecessors.

That matters should be so is by no means a new phenomenon. Thus, in Hancock’s magnificent Australia (1930), now sadly out of print, the great historian famously said that it was a failing of democracies, and especially of Australian democracy, to constantly confuse ends and means, and to show too much reluctance ‘to refuse favours, to count the costs, to discipline the policies they have launched’. ‘[The] policies therefore yield diminishing returns, until at last, they may become a positive danger to the national purpose that called them into existence.’ Nowhere was this more marked, Hancock noted, than with public involvement in infrastructure ventures such as rail, where Australian government was ‘particularly slow to confess it has got into a bad business, for its mere entry … has created vested interests which immediately express themselves in politics … So … it throws good money after bad, and hopes that something will turn up. In this way, losses accumulate in a lump, and the crisis, when it comes, is likely to be prolonged and severe.’

The costs and risks of this approach to infrastructure have also been known for many years. There are surely many echoes in current telecommunications decisions of the tendency, identified by Butlin, Barnard and Pincus (1982, p. 294) in their analysis of the development of the Post-Master General’s Department, for Australian public enterprise to provide ‘services that were too large, too quickly supplied and too cheap’. That so little should have changed is not encouraging.

Set against that background, how great a contribution can improved project appraisal make to securing better outcomes? Little and Mirrlees (1994, pp. 225–7) develop a simple model of the value of information in which good project appraisal yields benefits that, in expected value terms, are in the order of 10 per cent of project value.29 For an economy investing over $10 billion per year on its transport and communications infrastructure, 10 per cent of project value would seem like a saving well worth seeking. That said, the Little–Mirrlees model assumes unbiased estimates and a decision-maker who, as a benevolent social planner, maximises social welfare; it is hardly contentious that those assumptions do not hold — if they did, central planning would be a far better system than it has ever proved to be.

To recognise this, however, is not to imply that no value should be placed on good appraisal; on the contrary, it is one of the protections taxpayers deserve to have.  Testimonials of commitment to ‘evidence-based policy’ notwithstanding, shaping an environment in which project appraisal can effectively discharge this task remains as great a challenge as it has ever been.

Overall, our review suggests the following conclusions:

•      Insufficient attention is paid in the evaluation process to options that would avoid investment, or, more broadly, that would focus on securing greater efficiency from the existing capital stock. Simply put, infrastructure investment appears to be viewed as a benefit, rather than a cost.

•      The distortions arising from this undesirable narrowing of the range of options considered are then compounded by evaluations that are too vulnerable to ‘fudge factors’. In a Gresham’s law of evaluation, bad evaluations (often by consultants) can drive out good, given that they trade at equal values.

In our view, these outcomes are driven by governments that see little real value in major project evaluation. They may see merit in evaluation of essentially routine decisions (such as the decision to place a new roundabout or improve a road surface) or in cost-effectiveness analysis of the options available for meeting predetermined goals (such as improving bus transit in a congested area), but not in the full analysis of objectives and options (including the option of not spending taxpayers’ money). This, we argue, reflects the impact of a perception (initially due to strong economic growth, and then to a belief that the global financial crisis justifies greatly increased outlays) that public funds have a negligible opportunity cost. This perception has been accentuated by the growing blurring of accountability in the Australian federation, which reduces the budget disciplines on the States, and the blurring also of responsibility for financing infrastructure as between the public and private sectors (which, whatever its other merits, increases the return to rent-seeking deals between governments and private infrastructure developers). Together, these trends risk making cost–benefit analysis merely a box to be ticked, rather than an exercise that has real value, not least to government itself.

We are not optimistic that changes to cost–benefit analysis processes alone can counteract these powerful trends. Nonetheless, we think three changes would have merit:

•         a requirement for all cost–benefit analyses to be disclosed that would also highlight which projects had not been subjected to economic project evaluation

•         far greater and systematic auditing of cost–benefit analyses, both at the stage of the financing decision and post-project completion. In contrast, there is little or no such audit currently, and in many instances, cost–benefit analyses are not even -
*        updated,
*        maintained; or
*        properly archived after the initial ‘go/no go’ decision is taken.

•         the establishment of a centre of excellence or reference for cost–benefit analysis within the Australian Government, preferably in an independent entity, such as the Productivity Commission."