Reading time
4 mins
August 23, 2020
Dr Ganesh Nana (Former Research Director)

If not benefit-cost ratios, then what?

Is it time to put the CBAx model out to pasture?

It’s Saturday 14 March 2020 and you, a key policy advisor in the Treasury, have just been sent an ‘all hands on deck’ call-to-arms from your manager. This isn’t that much of a surprise, as you’ve been monitoring the increasingly large death tolls from COVID-19 in China, Iran, and Italy and there are a range of draft economic scenarios for the May Budget in front of you.

But this call didn’t sound like a drill – with tomorrow’s national memorial event postponed, and a $12.1bn package with business and wage subsidies set to be announced in the coming week. A new four level alert framework is also ready to be announced. Within that framework is the potential for an unprecedented lockdown of large swathes of business and economic activity if the top level were ever activated.

To assess and advise on the lockdown policy decision, you turn to the go-to tool of policy analysis: benefit-cost ratios.

Your job in the immediate next few hours is to assess the options implementing such restrictions.  As per normal, you turn to the go-to policy analysis tool that has guided you since your first day in the job. The benefit-cost framework has always been the bulwark of analysis for public policy choices. It ranks options and enables the robust and rigorous identification of the best option.  

However, this time around you don’t have the time for a deep dive into all the costs and benefits. But you endeavour to progress some first-cut, admittedly crude, calculations in order to provide a flavour of the consequential advice you have to deliver.

“Go hard, go early” could lose us more than $79 billion of GDP.

A “go hard, go early” Level 4 lockdown strategy indicated (according to the draft Budget Economic and Fiscal Update (BEFU) scenarios) a GDP loss of the order to $78.7 billion (to 2024). There is also increased government spending in the form of a response and recovery fund of the order of $62.1bn. That’s $140 billion in costs so far. So, what are the benefits? With the value of a statistical life (VOSL) from the Treasury CBAx model at $4.9m, benefits would approach $140 billion as long as a Level 4 lockdown decision were to save about 28,700 lives.

You know the above crude calculations would need to be finessed for the counter-factual, as the alternative scenario should also factor in some GDP loss due to global turmoil. So, erring on the side of caution, you slash the $140 billion cost to $70 billion. In addition, there are savings from cases avoided (as well as deaths avoided), and there is the impact on the health system to be allowed for too. Again, your cautious approach leads you to an early estimate of needing to save about 10,000 lives to get a balancing benefit.

Consequently, your policy advice takes shape:

“Prime Minister, as long as the number of deaths expected to be avoided is more than 10,000 we advise a ‘go hard, go early’ strategy. However, if avoided deaths are expected to number less than 10,000 then it would be best to hold off on the lockdown and avoid a draconian loss of GDP.”

Admittedly, the above is a fiction.

And I must record that the behind the scenes work at this time of crisis is to the credit of the New Zealand public service. That the implementation of a widespread lockdown across much of New Zealand activity took place relatively smoothly is testament to its professionalism. Especially, when considering the speed at which these measures – with relatively little notice - were instigated.

But, is it more than academic that when faced with an unprecedented one-in-a-hundred-year event, our standard go-to tool for policy decisions was left on the side-line? In our time of crisis, desperate need, and pending chaos, why did our standard benefit-cost approach for public policy decisions fail us?

One of the most tumultuous policy decisions taken in recent memory was not guided by any formal (even cursory) benefit-cost assessment.

Moreover, it is indeed telling that the overwhelming majority of the populace agreed to and complied with the decision made. That suggests that our go-to benefit-cost assessment framework probably doesn’t reflect the values, choices, and trade-offs that are preferred by the communities of Aotearoa.

So, this leaves us as policy analysts and advisors in a quandary. What do we (should we?) now use as a basis for our policy analysis, advice, and recommendations? Not to mention policy decisions.

Perhaps now is the time to revisit the dominance of benefit-cost analyses, and more openly embrace multi-criteria analysis (MCA). Rather than converting all benefits and costs into monetary valuations, an MCA framework is better placed to incorporate the nuances of inherently qualitative preferences and trade-offs.

Such a framework can also be aligned to wellbeing perspectives – and ideally could be designed around the four wellbeings purpose of local government written into section 10 of the Local Government Act 2002.

This suggestion aims to capture the inherently qualitative nature of many desired wellbeing outcomes, as well as allay the discomfort many feel in monetising all values.

I know some (if not many) economists may baulk at this suggestion – especially those that are inherently uncomfortable with preferences that are not quantitative or converted into monetary values. But this suggestion is not intended to escape from choices and trade-offs inherent in any decision-making process. However, it does aim to capture the inherently qualitative nature of many desired wellbeing outcomes, as well as allay the discomfort that many in our communities feel when monetising all the values and taonga we hold dear.

In shifting more explicitly towards an MCA framework, policy analysis may then better reflect the values of the communities of Aotearoa.