Monday, November 19, 2007

Ross Gittins: the economic trickery used to support WorkChoices

From this morning's Sydney Morning Herald:

IT'S probably not possible to estimate with any accuracy the effect that Work Choices has had on employment, productivity, economic growth, inflation or anything else over the 20 months of its existence.

But that hasn't stopped a few sympathetic economists from making dodgy calculations, nor employer groups and Liberal politicians from using and misusing those guesses.

I've written before about the dubious nature of the econometric modelling exercise prepared by Econtec economic consultants for the Australian Chamber of Commerce and Industry...

...and subsequently quoted with great approval by the Government and used in the business coalition's TV ads.

Now John Howard is quoting a professor of economics, Richard Blandy, as having estimated that Labor's plan to abolish Work Choices would destroy between 200,000 and 400,000 jobs.

As is all too usual, Mr Howard is misquoting Professor Blandy's findings. His econometrics estimated that Work Choices was responsible for about 290,000 of the new jobs created since it was introduced in March last year.

He further calculated there was a 95 per cent likelihood that the true figure lay between 194,000 and 390,000. (Even the most carefully done econometrics is so approximate that rounding this out to between 200,000 and 400,000 is fair enough.)

See the misrepresentation? Saying that Work Choices has been responsible for creating that many jobs isn't the same as saying its abolition would destroy that many jobs. To give him his due, Professor Blandy acknowledged this misrepresentation when queried by Peter Martin of The Canberra Times.

"If it does have an effect, I would expect the unwind to happen much more slowly," the professor said. "Partly because employers will have found out what these people who they have hired are like. They're not necessarily going to give them the bullet."

He could have added that Labor wouldn't actually "abolish" Work Choices anyway. It is planning to retain a fair bit of it, particularly its anti-union measures. It would, however, restore a streamlined version of the unfair dismissal law for firms employing fewer than 100 workers.

But now a couple of economists have questioned the method Professor Blandy used to achieve his guestimate of the number of jobs created by Work Choices.

It's obvious there has been very strong growth in employment - particularly full-time employment - since Work Choices was introduced, but some of that would have occurred anyway (because the economy was growing) and you'd expect a lot of growth to have been caused by the direct and indirect effects of the resources boom.

In Professor Blandy's regression analysis, he attempted to account for the influence of the resources boom by using the "proxy" (substitute) of the change in the trade-weighted exchange rate.

Since the exchange rate didn't change much over the period he examined, he concluded the resources boom had made little contribution to jobs growth - only about 38,000 since March last year.

But Mark Wooden, a professor of labour economics at the Melbourne Institute, has been reported as commenting that this was "a brave assumption". A more reasonable assumption would have been that the exchange rate obviously didn't capture the effect of the resources boom.

And now one of Australia's top econometricians, Adrian Pagan, a professor at the University of NSW and the Queensland University of Technology and former member of the Reserve Bank board, has argued that Professor Blandy's conclusion is "a type of trick". (His comments can be found at johnquiggin.com)

Professor Blandy sought to capture the effect of Work Choices by adding a "dummy variable" to his regression from March 2006 onwards.

This means the dummy variable will pick up the effect of all the factors that contribute to employment growth apart from the trend growth in the economy (said to account for about 12,000 extra jobs a month) and the small number of jobs attributed to the resources boom.

Professor Pagan comments that "no matter what the effect of Work Choices is, one will find that the amount attributed to it by Blandy's regression will be very close to the average of the employment change after its introduction.

"This is simply an artefact of the way in which the analysis is done and has no causal significance whatever," Professor Pagan says.

It's clear that some other factor or factors - call it factor X - caused the big rise in employment occurring before Work Choices was introduced. If factor X was still operating after March last year then the way Professor Blandy did his regression means its effect would be attributed to Work Choices.

"Unless an attempt is made to control for X, there can be no confidence that Work Choices has had any effect," Professor Pagan says.

To emphasise the foolishness of this, Professor Pagan replaced the Work Choices dummy with a Rudd dummy from the time Kevin Rudd became Opposition leader in December last year.

Using Professor Blandy's slipshod methodology, he found that Mr Rudd's appointment created 10 per cent more jobs than Work Choices was supposed to have.

Economists who misuse econometrics - and whose conclusions are then quoted to score ideological points - bring their profession into disrepute. Their more scrupulous colleagues need to hold them to account.