Skip to Content
 Close search

The COVID-19 recovery: A view from the supply side


Chair Michael Brennan provided an online briefing to the Australian Business Economists on 23 September 2020.

Download the speech

Read the speech

In 1977, James Tobin quipped that: ‘It takes a heap of Harberger triangles to fill an Okun gap’. What did he mean by this?

To take you back to introductory micro, a Harberger triangle is the deadweight loss in a partial equilibrium diagram — the efficiency loss due to a resource misallocation, usually due to a distorting policy (say a tax or regulation).

An Okun gap — to take you back to introductory macro — is the inefficiency due to the gap between actual and potential output — a consequence of inadequate aggregate demand in the Keynesian tradition.

Tobin’s point was that the inefficiencies from resource misallocation are pretty minor when you compare them to the consequences of a lack of aggregate demand.

Especially when you are in a recession.

As Paul Krugman put it:

… even bad microeconomic policies, which lead to substantial distortions in the use of resources, have a hard time doing remotely as much damage as a severe economic slump, which doesn’t misallocate resources – it simply wastes them.1

What does this imply for policy?

One interpretation is that when times get tough, macro dominates micro: the policy focus should be on getting demand stimulus out there, and you should worry less about whether you are causing inefficiencies or distortions in the economy along the way.

A variation on the theme would be: now is not the time to worry about distortions and inefficiency: deal with the aggregate demand crisis now, and save micro-economic reform for another day — once the recovery has been secured and life returns to some semblance of normality.

I am going to put to you an alternative view. Not because I think aggregate demand is unimportant – it’s fundamental; but it’s a necessary not sufficient condition for recovery.

My contention is that supply side policy is an important enabler of the recovery, without which demand-side stimulus is incomplete or compromised in its effectiveness. We need to take the supply side seriously.

It’s not so much about Harberger triangles — it’s more a dynamic view about the speed with which the economy can move from one state to another and how we minimise the frictions along the way. Think of it as the micro foundations of the macro recovery.

There are three main reasons why we should focus on microeconomic policy even in the midst of a recession.

  • First, because the COVID pandemic is, among other things, a reallocation shock
  • Second, we face a particular challenge around new business formation and risk appetite, and
  • Third, because a strong microeconomic lens can help us build resilience and protect against bad policy.

The reallocation shock

It’s natural to track the recession and recovery by following the trajectory of GDP(E) and its components.

But as we know, even as GDP(E) starts its climb back toward its March quarter levels, the underlying composition of the economy will have changed.

GDP(P) will look different, and beneath that, so will the constellation of firms, employees, suppliers and customers; so will the deployment of physical and financial capital across the economy, the use of land and even the use of that other finite resource — time.

Because COVID is a reallocation shock, as well as a demand shock. It is a reallocation shock partly because of the inevitably staggered lifting of restrictions. Some industries won’t be back for some time. And there is also the likelihood of enduring changes in consumer and business preferences, possibly in sectors like retail, travel, office accommodation and aviation.

Of course reallocation occurs all the time — even in times of relative macroeconomic stability but there is some evidence that it could be more salient in a recession like this one.

Economists Nicholas Bloom, Steve Davis and Jose Barrero2 tried to quantify the extent of possible reallocation through the pandemic in the US, by analysing forward-looking expectations data from firms. They calculated a metric that combined the expected gross increase in hiring by growing firms, plus the expected gross job shedding by shrinking firms (the majority) and comparing it to the absolute value of the net change.

This provides a sense of the gross flows relative to the amounts required to give effect to the overall net change in employment. They found that this metric more than doubled in the early months of the pandemic, compared to its pre-COVID average.

In other words, firms were collectively anticipating much greater reallocations of labour across the economy compared with the pre-COVID period.

Such a reallocation process is never frictionless. When looking at past shocks, they noted that there can be a lag between the destruction side of the reallocation process and the creation side. As they put it:

Reasons for the delayed creation response include the time needed to plan new enterprises and business activities, the time required to navigate regulatory hurdles and permitting processes to start or expand businesses, time to build in capital formation…and search and matching frictions in forming new relationships with suppliers, employees, distributors and customers.

These practical hurdles are an important reminder of how recovery actually occurs in practice.

Beneath the macro aggregates lie countless daily individual acts of risk taking and entrepreneurship: the decision to start a business, borrow, invest, expand, hire a new worker, or just change jobs. They are shaped by incentives.

These acts are the very essence of the creative process that underpins economic recovery, and it makes a material difference as to whether or not policy supports them.

While empirical work on this is always imprecise — not least because it is based on historical data — there is at least one study of which I am aware, that regresses the severity of past downturns and the speed of recovery against elements of regulatory policy.3

It found that economies with efficient, credible, transparent regulation experienced milder downturns and faster recoveries than those with heavier regulatory burdens. As the study notes:

[The] speed existing firms and new entrants can reallocate resources depends on the regulatory framework and the efficiency and transparency of its enforcement. Licensing requirements and similar business regulations constitute entry barriers that prevent entrepreneurs from seizing legal opportunities and thereby limiting the economic and social losses during crises.

So we need to address rigidities that prevent labour and other inputs from moving between firms, sectors and occupations.

Here’s one example: In August the NSW Productivity Commission released its Green Paper on productivity reform,4 which is an excellent piece of work, and I really commend it to you.

One of the areas they looked at was the ability to attract mid-career professionals into teaching. It’s a very intuitive idea — indeed there might be people in today’s audience who would be interested in teaching secondary maths and economics, and what a brilliant outcome that would be for preparing and inspiring the next generation to make a difference in the world.

The question is how you bring that about without requiring people to go back and do the equivalent of a 4-year teaching qualification designed for school leavers.

Teaching isn’t alone in that respect. There are similar rigidities across a range of occupations, brought about by registration, licensing and qualification requirements. In general, our skills system (or at least the funded element) is still very focused on the initial acquisition of a full qualification.

But for workers displaced by the current recession, is that always realistic? For many, the path into alternative occupations is through targeted skills acquisition (at best, components of qualifications) and recognition of existing knowledge and capability, perhaps through a system of independent assessment.

This is a complex area because context matters, so it requires a detailed look at the specifics of individual sectors to work out where and how we could remove barriers to lateral or mid-career movement.

One more obvious area is occupational licensing, where our 2010 report on mutual recognition illustrated the simplicity that could come from a system of automatic recognition of licenses across states, which the Federal and State Treasurers are now working on.

Stamp duty can also be a barrier to the movement of labour to different locations, as we noted in Shifting the Dial in 2017. It can also be a barrier to using land for new and more productive economic purposes.

So too, planning and zoning regulations — also noted in Shifting the Dial — which can slow the reallocation process by being prescriptive about where specific types of economic activity can and can’t take place.

Of course, one policy where this comes into sharp relief is Jobkeeper. In the initial stages of the downturn, Jobkeeper has been very valuable in supporting businesses, and maintaining the connection of employees to the workplace. And it provided a much needed income boost as the economy weakened. But it also creates a rigidity by tying existing workers to existing firms, and a time inevitably comes when that holds back the process of recovery.

One additional point about where efficiency considerations become relevant to the recovery.

Once a reallocation process is underway, we want to make sure that labour and capital (and land use) are gravitating towards sectors and firms where they will be more productively employed. We don’t want them gravitating to the beguiling security of sectors which benefit from inefficient and unproductive rents. If reallocation led to increased concentration and market power, then this would be a bad long term outcome.

Some of the Productivity Commission’s work explores areas where more competition or appropriate regulation could help minimise those rents, hence this work remains important even (perhaps especially) in a recession.

Risk and new business formation

My second point about the importance of a micro-economic perspective concerns risk appetite and new business formation. Again, abstracting from aggregate investment and production, we should think hard about the supply side.

In a normal year, there are around 300 to 350 thousand new businesses created in the Australian economy. A similar number exit — most of them not due to business failure.

As at June, we had over 900 thousand businesses on Job Keeper. Many of them have also benefited from interest deferral and (depending on the jurisdiction) some deferral of commercial rent. They may also have been given breathing space by measures like temporary relief from insolvent trading and higher thresholds for bankruptcy proceedings. Many will make it through, but logically some will not.

So the question arises, how can we improve the ease with which new business can get up and running? This is all the more urgent because one possible effect of the pandemic and recession is a reappraisal of risk.

A recent paper by Kozlowski, Veldkamp and Venkateswaran5 puts some framework around this. To over-simplify: lots of economic models incorporate risk — with economic agents facing a probability distribution over the value of some parameter or other.

When the parameter takes on a particular value, that affects the short term outcome in the model, but it doesn’t change the underlying probability distribution for subsequent periods. That’s basically consistent with a rational expectations assumption.

But how well does that assumption fit with extreme tail events?

As the paper notes: Isn’t it plausible that real world decision makers, when they observe a tail event, actually update their subjective beliefs about the likelihood of it being repeated? And if they do, then there is a form of long term scarring that occurs specifically because of changed beliefs: that is, a revision to their perception of future risk.

That needn’t be just a case of businesses deciding that pandemics are now more likely than they previously thought. It could apply to the whole range of disruptions that have gone along with the pandemic — border closures, lockdowns, supply chain disruptions, rent waivers etc.

Moving forward, it is possible that decision makers will face not just generalised economic uncertainty, and not just a reduced expectation about the future central scenario, but also a heightened sense of tail risk — a belief that extreme scenarios are just that little bit more likely, more believable, than they would have thought in 2019. And that could feed into their behaviour and their investment decisions.

The authors of the paper find that this ‘belief scarring’ could be bigger in its long term impact than direct impact of the pandemic shock itself.

What to do about it?

It’s easier to diagnose the problem than prescribe the medicine. (And I hear the macro-economists in my ear saying: get aggregate demand up asap — heal the wound fast and you’ll minimise the scarring). And that is no doubt true.

But supply side policy can also play an important role.

Some States have made significant efforts to streamline the approvals processes for new businesses — such as the Services NSW Business Concierge, and the Better Approvals program in Victorian local councils.

Again, the structure of the planning system plays a part, since much of the delay associated with starting new businesses can be getting planning approval.

Regulation reform more broadly is important — particularly the behaviour and culture of regulators (as distinct from the rules on the statute book). This is the subject of our current inquiry into regulation in the resources sector, which looks at best practice by individual regulators across the Federation.

Insolvency rules could also be an important policy lever, which we examined in our 2015 report on Business Set up, Transfer and Closure.

Workplace relations is another. Giving firms and employees the space to work out arrangements that achieve their shared goals while protecting minimum conditions.

One other area of policy — often neglected in this context — is the social safety net.

The safety net goes beyond income support. It includes things like rent assistance and family tax benefit, but also instruments like income protection through life insurance, access to health insurance, workers compensation, and as we have seen in the current pandemic, superannuation. It can also be extended to regulatory policy in areas like workplace relations or consumer protection.

The safety net serves a redistributive goal. But that is not its only role. It can also be an important determinant of risk appetite, as the New Zealand Productivity Commission found in its work on Technological change and the Future of Work.6

They looked at a range of designs, including the ‘flexicurity’ approach in some northern European countries. Just one example by way of observation: the US approach of linking health insurance to the employment relationship might sharpen work incentives, but could also dull the incentive to quit your job and start a business.

I am not calling for any particular reform; just noting that the issue warrants closer attention in Australia as part of the policy] mix to encourage efficient risk taking.

Our work has brought us into contact with many aspects of the safety net in the:

  • Veterans inquiry — observing a system designed around poor incentives
  • Superannuation inquiry — where we found that life insurance is, for many people, an accidental and poor value product
  • Shifting the Dial report — where we noted the limitations on private health insurance and the lack of incentive for funds to invest in keeping their members well
  • Mental Health inquiry — where we looked at the role of workers compensation arrangements.

The system is multi-faceted if somewhat fragmented.

In the current pandemic, we have seen a temporary mix of risk pooling through a higher short term payment, and a form of self-insurance through access to super. This will be a big policy question: how do we bring together the various strands of the safety net to balance the right level of redistribution with work incentives, but also to create the right incentives for efficient risk taking.


My final observation concerns resilience.

A recent paper by the Queensland Productivity Commission7 noted the importance of past micro economic reforms in building greater adaptability and resilience in the Australian and Queensland economies, especially in the face of shocks and disruptions.

It is a reminder that in economies, as in individual businesses, risk management is not a predictive science — it’s not a question of who has the best crystal ball. It’s about having the adaptive capability to deal with unforeseen events as and when they arise.

Once the dust settles, we can assess how well Australia has responded and adapted to this pandemic. Calls for greater national self-sufficiency should be assessed in that light. My instinct is that there is a lot more value in building general resilience than there is in identifying and preparing for specific future contingencies. I accept there could be a bit of both.

Another likely conclusion is that Australia’s relatively poor productivity growth in the period leading up to the pandemic (which has been a focus of the PC’s work) could reflect a decline in the dynamism of the economy, potentially exposing one source of vulnerability when the crisis hit.

The main point is: you can’t prevent a shock. But you can prevent bad policy, and a sound micro-economic framework is one of the best protections we have.

Nowhere is this starker than in infrastructure policy. If you build things solely for demand-side stimulus, you run the risk of wasteful spending. If you do careful project assessment, you can boost productive capacity and aggregate demand.

And there is an important subtlety here: a project assessment — say a cost benefit analysis — is a microeconomic tool. It isn’t perfect, because concepts like utility, consumer surplus, welfare, willingness to pay are abstract and they are hard to measure. But they at least try to get at sources of value that aren’t captured in macro aggregates like per capita GDP.

It is this approach that has allowed economists to apply a policy framework to a broader and more complex set of questions, which happens to be a staple of the Productivity Commission’s work.


Recessions in Australia have been rare of late. When they do come, they can lead to policy discontinuity. After the 1890s recession we pursued a policy of protection and regulation. After the early 1980s and early 1990s recessions, we took a different path, combining sound macro-economic frameworks with vigorous micro economic reform.

Today is different in the particulars. The reform priorities have to be calibrated to the times. But the principles are enduring. One such principle is that we should avoid any crude misinterpretation of the Krugman formulation — that is, to think that somehow in the midst of recession, we can ignore efficiency considerations, so long as we are adding to aggregate demand.

I don’t think the question is so much about Harberger triangles and Okun gaps, as it is about harnessing the supply side of the economy to hasten the recovery, build resilience, speed productivity growth and tackle broader policy questions.

In achieving all that, micro still matters.


1 Return to footnote 1

2 Jose Maria Barrero, Nicholas Bloom and Steven J. Davis, COVID-19 Is Also a Reallocation Shock , NBER Working Paper No. 27137, May 2020 Return to footnote 2

3 Christan Bjørnskov, Economic Freedom and Economic Crisis , IFN Working Paper No. 1056, 2015 Return to footnote 3

4 Return to footnote 4

5 Julian Kozlowski, Laura Veldkamp and Venky Venkateswaran, Scarring Body and Mind: The Long-Term Belief-Scarring Effects of COVID-19 , NBER Working Paper No. 27439, June 2020 Return to footnote 5

6 Return to footnote 6

7 Return to footnote 7