The newly invigorated Australian Council of Trade Unions is demanding a $45 a week increase in the federal minimum wage, a rise of 6.7 per cent, which has shocked and appalled the employer groups and the Turnbull government.
If I was on the minimum wage, however, I wouldn't start spending the increase yet. It's all a bit ritualistic, with the unions demanding far more than they expect to get, while the employers cry poor and predict huge job losses should anything more than the tiniest increase be imposed on them by the Fair Work Commission.
Not that many years ago, most economists would have shared the employers' doubts about the wisdom of even a modest increase in the minimum wage.
Indeed, conventional economic analysis – using the "neo-classical" model of markets – told them that government intervention in the labour market to set a "binding" minimum wage – that is, one higher than would be set by the unfettered interaction of supply and demand – might benefit those workers who managed to retain their jobs, but must inevitably mean many unskilled workers would be prevented from getting jobs.
Just how many people were unemployed as a consequence of holding the minimum wage above its "market-clearing" level would be determined by the "elasticity" – the degree of sensitivity to price changes – of employers' demand for unskilled labour.
There are probably plenty of economists who still believe all that, particularly those who don't make a study of the economics of the labour market and rely on elementary analysis of any and every market.
After all, such analysis is completely logical, given the assumptions on which the simple model rests.
Trouble is, it's long been obvious to those who cared to look that the conventional model isn't much good at predicting what will happen to employment and unemployment.
For instance, those economists who use the neo-classical model – as opposed to a Keynesian approach – to explain the behaviour of the macro-economy are obliged to argue that the jump in unemployment during recessions is voluntary rather than involuntary.
It's just a lot of workers choosing that moment to take an unpaid holiday.
But the big challenge to economists' conventional wisdom that minimum wages cause unemployment came in 1995, when two American economists, David Card and Alan Krueger, published empirical evidence showing that a 19 per cent rise in New Jersey's minimum wage actually saw a small rise in employment.
Many studies since then have come up with similar findings.
This suggests the conventional model of markets doesn't offer a useful description of how the labour market works. Either the model's many assumptions don't hold, or there are key factors affecting labour markets that the model doesn't capture.
This is no radical idea. A father of neo-classical economics, Alfred Marshall, argued as long ago as 1920 that the market for labour differed from two other "factor markets" – markets for the factors of production - land and capital.
Why? Because, according to Marshall, workers retain ownership of their human capital (skills) – they're free agents – and because workers must be present in the workplace for the delivery of their skills.
The first characteristic means that anything workers learn on the job, or are trained to do, remains their property, not their employer's, thus giving them some control over the use of those skills.
The second characteristic – that every unit of labour an employer purchases comes with a human being attached – means workers can't live very far from the workplace.
Since moving homes involves cost and inconvenience – especially if the worker has a family – this may give employers some ability to exploit their workers.
Remember this and the notion that a model for the buying and selling of land, or machines, or for the borrowing and lending of dollars, would work just as well in explaining the buying and selling of labour, is fanciful.
So what other, better models of the labour market are there? Labour economists are working on many. A favourite of Professor Alison Booth, of the Research School of Economics at the Australian National University, is the "oligopsony" model.
Huh? Monopoly means there's just one seller of a product. Monopsony means one buyer of a product or, in this case, input. Oligopsony means just a few buyers – by no means uncommon in a modern economy where a few big companies dominate many product markets.
The oligopsony model assumes that even if workers have identical skills and abilities, they have differing preferences on which employer they want to work for, influenced by such things as how far the firm is from where they live, the hours they want to work, or whether they like the boss and their fellow workers.
It takes time and effort (that is, cost) for workers to find alternative employers they like at least as much as their present one and, similarly, it's expensive for employers to find a worker they like as much as the one they could lose.
This makes many workers reluctant to change jobs and many bosses reluctant to change workers. And because these preferences are private information – the other side can't be sure how strong they are – there's scope for "economic rents": for workers to be paid less, or more, than the value of their work. Less is more likely (except for me).
Booth says the attraction of the oligopsony model is its ability to show how a minimum wage can actually increase employment, as well as why employers provide general training to workers who could leave and take the training with them.
Trouble is, these alternative models of the labour market may be more realistic, but they're also more complicated and harder to reduce to a set of equations.
Keynes once said it was better to be roughly right than precisely wrong. A lot of economists disagree.
