This doesn't always happen of course. In fact, sometimes it seems as though the larger the stakes, the smaller the influence of objective analysis. Election promises are a big contributor to this sad state of affairs. For example, the billions going into a new fibre-optic broadband network (UFB) started life as an election promise, as does the NZPower proposal. Both are huge upheavals without any cost-benefit analysis, so we just don't know whether they make sense.
Which brings me to the dreaded Roads of National Significance (RoNS) and to the recent paper by Michael Pickford analysing how wasteful road investment has become. Here is a summary chart from Michael's data, showing the fraction of road investment going to projects with high, medium & low benefit-cost ratios:
In 2005/06 none of the ca$h was spent on low value projects, though almost half went to medium value stuff. That has been getting worse and in 2009/10 almost 70% of the spend went on low value stuff. Michael picks 2003 as the year this madness began, when the criteria for investment changed from "pick the high benefit-cost ratios" to include two other, more subjective criteria. Then, once the RoNS arrived on the scene, "being a RoNS" was added to the criteria list, even though that status is a decision made entirely by politicians.
Oh and in case you're wondering, there is just over $1bn of public cash buried in those green bars on the chart.