A Bridge to Sell

Unknown Recently at the Warren Centre ( there was some commentary around the cost of duplicating Brisbane’s Gateway Motorway Bridge.  Namely, between the original and the duplicate, costs increased fivefold.  On the surface that would be a stunning thing.  However, when we consider 24 intervening years (1986-2010) between the costings for the two projects, we are looking at an annualised increase in costs of 6.9%.  This is still significant when inflation only increased 3.4%p.a. over the same period.  I know there are construction price indices around as well.  However, I wonder whether design and materials were significantly different, and the connecting road infrastructure was more challenging. That might explain some of the residual increase.  Also, the duplicate would have been constructed in a much tighter market for engineering and construction services.  Should we have bought the extra lanes way back in 1986 or would the opportunity cost have been too large?


NSW Northern Beaches Hospital Deal

One of the largest public health projects in NSW has been slowly releasing more and more information about the total cost of the project. Like some other public private partnership arrangements this project is a combination of capital investment and operating service payments.

The Northern Beaches Hospital project was belatedly revealed to the public to cost $2.14 billion cost, up from “over $1 billion” in December 2014. See:

A lot of information was shrouded in commercial secrecy and the Health Minister has indicated she did not get involved in the financing.

So let’s have a look at what the good taxpayers of NSW may be getting.

From public information, the hospital development itself will cost $600 million. The NSW government plans to chip in some $400 million in adjacent transport improvements. A big assumption is that this is actually delivered on schedule and budget.

That basically handles the capital investment side of the equation and may leave $1.14 billion for the hospital services NSW is purchasing up to 2038 from the consortium.

Let’s say the hospital is operational in 2018, so there is a 20 year service period. According to some reports, 488 public beds will be provided.

This means the average daily cost per bed – occupied or not – is around $320 over this period.

Across the NSW health system the average cost per occupied bed was around $1,400 per night in 2012-13, based on information given to the NSW Auditor General.


So this looks like it might actually be a good deal even at low occupancy rates. But since we do not know the basis of the service payments we simply cannot be sure.  The consortium get to make some extra cash on the side by selling private services through the site.

At about this point I can hear a lot of health economists scream.

So let’s recognise there is a lot of detailed, good work being done by the Independent Hospital Pricing Authority (see: that is putting evidence-based price signals into the hospital funding system, allowing funding arrangements to move from block grants to activity-based funding payments.

Over time this should help improve the efficiency of our hospital systems.


What Price Data?

When we look at infrastructure services we should also consider the context in which they are provided.

Currently I am on assignment in PNG.  Telekom here in the past week has slashed mobile broadband top up prices by 50% to around A$25 per gigabyte (GB).  It sounds impressive and may have a meaningful impact.  However even that level it is still eyewateringly expensive compared to Australia.  But stop and think about it in the context of purchasing power.

Some data I looked at recently suggested GDP per person in PNG was probably A$2,500.  In Australia GDP per person is around $45,000, some eighteen times the level in PNG. Depending on the definition the actual figures can vary but consider that as a rough context for income.

If Australian’s faced the same share of GDP spent on broadband top ups as in PNG that would look iike a cost of $450 per GB.  How fast a rate of uptake would you expect if Australia faced that price?  Mind you it was not that long ago we were paying those prices for dial-up.

This also supports earlier analysis by the International Telecommunications Union:

“By early 2013, the price of an entry-level mobile-broadband plan represents between 1.2-2.2% of monthly GNI p.c. in developed countries and between 11.3- 24.7% in developing countries, depending on the type of service.”


Clearly further, lower prices are needed in PNG along with relative increases in GDP per person.  The digital divide remains huge and the context remains relevant whether you are looking at PNG or Australia.

Image source:


Infrastructure and Human Development

I am currently in Port Moresby on an assignment.  It is a three hour flight from Brisbane but a world away in many other respects.  The UN releases reports on human development.  Some contrasts are quite stark.

PNG       Australia    Measure

0.411      0.933         Human Development Index

157th      2nd            Human Development Index rank

7.25        23.3           Population (million)

17.07       950.65      GDP (US$billion 2011)

2,381       42,278      GDP per capital (US$ 2011)

0.617       0.113        Gender inequality (lower is better)

62.42       82.5         Life expectancy at birth

When we look at infrastructure it makes you think vey hard about what is important.  It really is all about context.



Accuracy and precision are different things, especially for infrastructure

The effort invested in ‘getting it right’ should be commensurate with the importance of the decision – Daniel Kahneman

Accuracy is the absence of error; precision is the level of detail. Often in infrastructure planning and analysis we see an unnecessary need for precision in the early stages with not much emphasis on accuracy.

Effective problem solving requires always being accurate but only being as precise as is helpful at any given stage of problem solving. This is about delivering analysis that is sufficient to proceed to the next stage in developing an infrastructure project. Many project approval gateway processes recognize this, but it is often poorly implemented.

Early in the problem solving process, accurate but imprecise methods, rather than very exact methods, will allow consideration of all reasonable approaches and minimize the tracking of needlessly detailed data. In this way, less apparent but potentially higher value options and scenarios can be considered and compared with the standard approaches.


Better Infrastructure Planning

Better infrastructure planning avoids presuming a solution.

Often we get locked into assumptions about the nature of the problem, its causes and desirable solutions. Wise infrastructure planners step back – examine what caused the problem, what caused the causes, and what caused those causes. This reveals possibilities very different from what end users envisage, but meet the true need most effectively.

A proper needs analysis is critical. If overlooked, huge value is often lost – for society, for the economy and for business.