Markups charged by Australian retailers have climbed from $1.3 billion in 1994 to $14.5 billion in 2022, according to new research presented at a UNSW Sydney workshop.

Kevin Fox, a Professor in the School of Economics at UNSW Business School, set out a method for measuring markups without relying on the restrictive economic models used in much of the existing literature. Based on findings from his working paper with Professor Erwin Diewert on markups and productivity in the Australian retail sector, Prof. Fox also noted that the maximum markup ratio in the sector (a distinct measure from the dollar figure above) reached 10.9 in 2020, its highest point on record.

Speaking at the 4th Firm and Industry Dynamics Workshop, jointly hosted by UNSW Sydney and the University of Hong Kong, Prof. Fox told the audience that the rise in markups did not begin during COVID-19, despite common assumptions, and traced the pattern back decades rather than to any sudden change.

"Interestingly, markups have gone up, but they have been going up for a long time. It's not like something suddenly changed,” said Prof. Fox, who highlighted a jump in 2019, ahead of the pandemic. "In 2019, markups jumped up, and then again in 2020, but at a lower growth rate. Many people think the pandemic was when price gouging started. In general, there seems to be an increasing trend from the late 90s,” he observed.

Retail sector markup

Large supermarket chains have faced recent court action from the Australian Competition and Consumer Commission over pricing conduct, and one major chain was found to have engaged in misleading advertising. Prof. Fox said that the retail sector's background made it a useful case study for this new methodology, given its concentration and the public interest in how prices are set.

Falling input costs, not price rises, drove 2020 increase

The research traced the spike in markups around 2020 to falling costs faced by retailers rather than sharp increases in prices charged to consumers. Prof. Fox walked the audience through the data behind that conclusion.

If retailers had been price gouging, output prices would have surged sharply in 2020. Instead, Prof. Fox found output price growth was only slightly higher than in 2019, while input prices actually fell. He said that the decline in input costs, rather than any jump in prices charged to consumers, accounted for much of the increase in markups that year.

Decomposition of output growth into explanatory factors

He noted that input growth fell in 2019, likely reflecting the drought and bushfires, while output held up. In 2020, both inputs and outputs rose, but outputs grew by more. Combined with falling input prices, he said, this pushed average costs down, which in turn helped drive markups higher.

Asked by an audience member about mergers and subsidies as possible drivers, Prof. Fox said his data pointed elsewhere. "There have been significant disruptions in the industry, but there's no evidence that the supermarkets suddenly started increasing their prices, leading to very large markups. Rather, recent increases are part of a longer-term trend."

New method avoids production function estimation

The working paper sets out a way to estimate markups using only observed combinations of inputs and outputs over time, rather than estimating a production function (an approach Prof. Fox said carries known problems for analysts). He said the aim was to derive an easily replicable method that statistical agencies could adopt.

"Our approach is the following: we look at past vectors of inputs and outputs for a production unit, whether that's a firm or a sector, and they're used to build up an approximation to the true technology. In other words, what's been done in the past should be feasible in the future,” he explained.

"You shouldn't lose blueprints in a modern economy, so any decline in productivity shouldn't be due to technological forgetfulness"

KEVIN FOX

Prof. Fox said the method breaks markup growth down into three parts: changes in efficiency, changes in prices, and a combined measure of technological progress and returns to scale. He said the team deliberately avoided attributing any decline in output to forgotten know-how. "Our view is that you shouldn't lose blueprints in a modern economy, so any decline in productivity shouldn't be due to technological forgetfulness,” Prof. Fox said.

He cited the 2019-20 downturn as an example: "When the economy goes into recession and productivity falls, is it right to assume that you’ve forgotten how to do things as you could previously? That would be technological regress. We're calling it inefficiency, because firms can still know how to do things as before, but for whatever reason, they're just not able to do it."

Prof. Fox extended this observation by pointing to farming during drought and bushfire periods. "It's not that farmers have forgotten how to farm as productively as previously, it's just they don't have, say, the water to farm,” he said. “Their labour and their capital are left idle, so that's better thought of as inefficiency rather than a sudden fall in technological ability."

Official data understates productivity

Prof. Fox said the national accounts approach used by the Australian Bureau of Statistics rules out markups by construction, because it sets an internal rate of return on capital that forces revenue and cost to match. "So, there can't be any markups in this case. This means that using only official ABS data, we're not going to find anything,” he explained.

To get around that, his team built its own estimates of the cost of capital using ABS data alongside unpublished data supplied by the agency. Prof. Fox said this also affected measured productivity, because the ABS excludes land and inventories from its inputs for this sector. "Why are we showing so much more productivity than the ABS? Because the ABS did not include land and inventories as inputs into production,” he said.

Prof. Fox also pushed back on the widely held view that Australia has experienced no productivity growth at all. He said the picture varies sharply by industry sector. "The performance of the aggregate market sector leads people to say ‘There's been no productivity growth in Australia,’ but there has been plenty of productivity growth, but it varies dramatically across industries,” he affirmed.

Subscribe to BusinessThink for the latest research, analysis and insights from UNSW Business School

“Some sectors have had spectacular improvements in productivity, especially agriculture, forestry, and fishing, except for drought years, bushfires and floods."

He said retail sat close to the aggregate market sector figure, which was dragged down by weaker performers. "Retail trade is just above the market sector, which is dragged down by industries like electricity, gas, water, waste services, construction and mining."

Rising markup share raises concern

Prof. Fox closed by flagging the broader economic cost of a rising share of income going to markups, an issue he said should concern policymakers and business audiences alike, given the parallel decline in multifactor productivity growth over the same period.

Income shares for six types of capital services, markup profits and total shares of capital services

"The growth of markups as an income share is worrying, because markups are like a tax on society. We don't want that,” he said.

The paper forms part of a wider body of work by Profs. Diewert and Fox on measuring markups and productivity, building on a nonparametric approach first outlined at the same workshop in Hong Kong the previous year. The retail sector analysis presented at UNSW represents the first full test of the method against real Australian data.