Why flat wages growth in Australia tells us nothing about the impact of migration on the labour market
One of the important concepts one learns in studying the way firms work with respect to pricing and markups is the distinction between quantity and price adjustment over the course of an economic cycle. When economists talk of supply and demand, they usually refer to price adjustment, where prices adjust up or down when there is an imbalance between these aggregates. Orthodox economics presumes that prices adjust, for example, to eliminate an excess supply, which they apply to the labour market and conclude that the cure for mass unemployment is wage cutting. The problem is that in many circumstances, firms use quantity adjustments long before they contemplate price adjustments, because the former involves lower costs. The Australian Broadcasting Commission (ABC) ran a story from its business reporters today (November 16, 2021) – As migration restarts, will it hold down wages for everyone? – which has also become a feature news segment on its television coverage today. The analysis presented is seriously misleading. It not only fails to characterise the problem properly but buys into a highly contentious debate about whether migration has negative impacts on the labour market prospects for local workers. It behoves analysts to actually construct the problem correctly before they start taking sides in this debate. The ABC article fails in that regard which is disappointing. Their failure also reflects the lack of diversity in opinion they seek these days. They chose to simply rehearse the arguments presented by one pro-migration analysis as if it was definitive rather than seek expert opinion from neutral analysis. But it also demonstrates why understanding the difference between quantity and price adjustment is crucial to getting the conclusions right.
The immigration debate
The debate in Australia about the impacts of migration on employment and wages in Australia is long-standing and won’t go away.
The business lobby all claim that migration is a positive factor because it adds more to demand than supply, but their members also use immigrant labour to undermine trade unions and suppress wages growth.
The trade unions claim the opposite.
Conservative governments support the business lobby, while the Labor Party wax and wane (as usual).
Most mainstream economists side with the business lobby.
However, in a Speech to the Economic Society of Australia (Queensland branch) – The Labour Market and Monetary Policy (July 8, 2021) – the Reserve Bank of Australia governor Phililip Lowe said. the ability of local firms to access cheap foreign labour:
… dilutes the upward pressure on wages in these hotspots and it is possible that there are spillovers to the rest of the labour market. This hiring can also dilute the incentive for businesses to train workers to do the required job.
This is a point I have been making in my own work for some decades now counter to what most of my academic colleagues argue.
It is not a case against migration. Rather it is a case for ensuring there are enough jobs, training and transition for school-to-work opportunities available for all.
Too often, governments have allowed short-term visas to foreign workers to be granted when the local labour is enduring mass unemployment and the governments have been cutting funding to local apprenticeships and vocational training institutions.
That sort of migration definitely damages the prospects of the more disadvantaged local workers.
I wrote about that work in this recent blog post – Restricting population growth is good for local workers (June 22, 2021).
I update some of the graphs from that post below.
The ABC article gets it wrong
The ABC article cited in the introduction is an example of rather crude mainstream analysis, which ignores the proverbial elephant in the room.
It poses the question:
Does mass migration push wages growth lower for everyone?
And in opening its analysis, it gets a first-year microeconomic textbook out and tells the readers:
It seems a simple case of supply and demand: having more of something compared to how much is wanted keeps prices low.
And it uses example about ice-cream shops by the beach competing for trade selling a largely homogenous good at the same ‘competitive’ price.
They then claimed that:
… when COVID-19 forced Australia to shut international borders, it started a test that might kill off the applicability of that basic economic concept when it comes to people.
They get this idea about a ‘natural experiment’ from an economist working for a mainstream economics lobby group – the Committee for Economic Development of Australia (CEDA), which has been strongly pro-migration for decades and firmly neoclassical in its economics (mainstream that is).
The problem for mainstream economists is constructed by the ABC article in this way:
1. The external border closures since March 2020 as the Australia government worked out how to deal with the pandemic has meant that some “500,000 temporary migrants have left the country” and “population growth fell for the first time in a century”.
2. “With the same amount of work to be done, the laws of supply and demand mean wages should have rocketed as employers use higher wages to compete for a much smaller pool of workers”.
3. But wages growth is flat.
4. So, what gives? Why isn’t demand and supply working?
The ABC article quotes CEDA research which concludes that:
We’ve seen the experiment of that in COVID, and in shutting down of our borders, and what impact that has on the labour market, which has been quite negligible.
Wages growth has remained flat but migrants have disappeared, which is the juxtaposition that motivates their conclusion.
Crude to say the least.
Quantity rather than price adjustment
Last Friday (November 12, 2021), the Australian Bureau of Statistics (ABS) published the latest – Overseas Arrivals and Departures, Australia – and we get this stunning graph from the data.
It shows overseas arrivals and departures from September 2011 to October 2021 in millions.
That is the definition of a ‘shock’.
The average arrivals before the pandemic was 1.53 million persons per month, departures 1.51 million per month.
Since April 2020 (after the border closure effect began) the averages are 0.04 million and 0.03 million, respectively for arrivals and departures.
I have also published this graph before in my monthly labour force updates.
It shows Australia’s working age population (Over 15 year olds) from January 2015 to October 2021.
With the external border still largely closed, immigration has shrunk to virtually zero and this has forced employers to work harder to get workers and is one of the reasons unemployment is falling quite quickly, given the circumstances.
And that is the clue to the mystery.
Quantity rather than price adjustment.
In the blog post – Restricting population growth is good for local workers (June 22, 2021) – cited above, I presented the evidence to date on why firms were using quantity adjustment over price adjustment to deal with the vanishing access to external migrants.
I updated the graphs today.
Labour markets exhibit varying responsiveness to demand and supply changes, depending on the context.
In some cases, they will be price adjusting.
In many cases, they will be quantity adjusting.
To understand the difference between these two adjustments, we see that firms typically react to changes in aggregate spending by adjusting only the quantity of output rather than the price and quantity.
First, firms use mark up pricing principles, whereby they add a profit mark up to their unit costs and face roughly constant unit costs over the output range within which they normally produce.
Typically, they maintain some excess capacity and can thus increase output relatively easily without further investment in productive capacity, which would take time. If they face insufficient capacity relative to demand, firms are likely to raise prices to ration demand which inevitably leads to the loss of customers to competitors.
Second, firms face various costs when adjusting prices and thus only periodically make such adjustments.
It has been said that firms use ‘catalogue pricing’, whereby they make their prices known to their prospective customers through advertising and other means and then are prepared to sell goods and services at those prices irrespective of demand (up to their full productive capacity).
At the end of the current catalogue period, they will then make any necessary adjustments to prices based on expected future demand and any recent and expected movements in unit costs.
When there is no excess capacity left to respond to nominal spending growth, then firms will shift to predominantly price adjustment as a way of rationing the spending growth.
In the real world, these two adjustments are going on all the time but the bias is towards quantity adjusting when there is excess capacity and vice versa.
As the economy approaches full employment, more sectors will be closer to capacity than others, so the different adjustments are spread unevenly across sectors.
How does this help us understand why the ABC article was misleading?
The fact is that when the pandemic struck the economy was nowhere near full employment.
There was massive excess capacity.
The aggregates as at March 2020 were:
1. Unemployment rate – 5.3 per cent (723.2 thousand)
2. Underemployment rate – 8.8 per cent (1205.5 thousand)
3. Broad labour underutilisation rate (unemployment plus underemployment) – 14.1 per cent (1,928.8 thousand).
4. Participation rate – 65.9 per cent compared to August 2019 peak of 66.2.
GDP growth was slowing and the labour market was deteriorating quickly even before the pandemic hit.
So 14.1 per cent of the available local labour supply was not working in one way or another.
Which means that there was a massive pool of labour that local firms could draw on as the external borders shut and the pool of cheap imported labour dried up.
And, all they had to do was offer a job at existing wage rates, for this excess local labour to accept employment.
No wage bidding was necessary.
That is, firms quantity adjusted rather than price adjusted to the shifting conditions.
They covered the ‘shortage’ by accessing the local underutilised worker pool.
The following graphs show definitively the impact that the slowing population growth has had on the ability of the labour market to absorb unemployed workers back into productive employment following a recession.
I went back to the three major cyclical downturns since the modern labour force data series began (February 1978) for Australia.
I then found the employment troughs in those recessions and examined the period of recovery afterwards for 20 months in each instance.
Those periods were:
1. April 1983 to December 1984.
2. July 1991 to March 1993.
3. July 2009 to March 2011.
4. December 2020 to September 2021 – the current period of recovery is shorter obviously. I excluded the October 2021 observation, where the unemployment and rate shot up signalling the end of that period of recovery.
The following graph shows the relationship between monthly employment growth (horizontal axis) and the change in unemployment (vertical axis) for each of these periods, denoted separately by different coloured and shaped markers.
The thick line is a simple linear trend and the simple linear regression is displayed next to each line.
The number next to the x measures the responsiveness of the change in unemployment to the employment growth.
A larger number means the responsiveness is higher.
In each case, the slopes of the lines are negative as one would expect. Rising employment growth reduces unemployment.
But the respective slopes (measured by the number next to the x in each equation) show that the 1982 and 1991 responsiveness was very low.
It took a long time to reduce the unemployment rates (and they never came back down to the full employment levels).
The GFC downturn was different because the fiscal stimulus really prevented Australia from enduring a recession. There was a slowdown and unemployment did rise (as has never returned to the pre-GFC level).
However, the current recovery is very different – the slope of the blue line (and number on the x variable) is several times larger than the 1982 and 1991 recessions.
The major explanatory factor is the record low growth in the working age population as a result of the border closures.
The next graph shows the same data except that I replaced the change in unemployment with the change in the unemployment rate.
A similar tale, although with the added complexity of cyclical participation adjustment, which I won’t go into here.
The ABC article didn’t take into account the state of the labour market as the pandemic hit.
By suggesting that supply and demand analysis would lead to the conclusion that wages should rise as supply fell, the article assumed that the labour market was ‘on both curves’.
That is technical terminology to depict the neoclassical equilibrium where workers can get as much work as they want at the current wage (that is, they are operating on the ‘supply curve’) and firms can get as many workers as they want at the current wage (that is, they are operating on the ‘demand curve’) and the two curves are intersecting at that current wage.
But that assumes there is no excess labour at the current wage.
Clearly, if there is excess labour than workers are ‘off the supply curve’ because they cannot get the hours of work they desire (the supply curve reflects their desires).
As a consequence the labour market can adjust to a shortage of workers in one area by absorbing the excess of workers elsewhere.
Things might have been different if the economy was already at full employment when the borders closed.
Then price adjustment would have been the only option available to firms.
I hope it gets to that soon and wages rise.
That is enough for today!
(c) Copyright 2021 William Mitchell. All Rights Reserved.