Guest post from Adrian Pagan

Last week the Prime Minister is reported to have said that economist Richard Blandy had estimated Labor’s plan to abolish Work Choices would destroy between 200,000 and 400,000 jobs. This claim has provoked a response from Australia’s (and one of the world’s) leading econometrician, Adrian Pagan. My one line summary: Blandy has derived this estimate by comparing levels, but has ignored the pre-existing upward trend in employment.

The full piece is over the fold – if there are any opinion editors among my readers, this would make a great contribution for the final week of the campaign.

Economic tricks: A comment by Adrian Pagan

Many years ago a visiting econometrician from the U.S. gave a public lecture in which he mentioned that a secretary who was typing his material thought that “econometricsâ€? should be typed as “economic tricksâ€?. Last week the Prime Minister is reported to have said that economist Richard Blandy had estimated Labor’s plan to abolish Work Choices would destroy between 200,000 and 400,000 jobs. This was apparently based on an econometric model that was reported on in the AFR of 29/8/07. It needs to be appreciated that the conclusion is indeed a type of “trickâ€?.

Blandy claimed that Work Choices has added something like 290,000 jobs. When teaching econometrics to students one always argues that they should plot the data to get some idea of what they are trying to explain. Later, after they have finished the modeling, it is a good idea to see what the model did explain.

Following this advice one would get the figure below.

200711160942

The quite volatile dashed line is the variable Blandy was explaining – deviations from employment after trend – and the solid line called “fitâ€? are the predictions of his model. Work Choices was introduced in May 2006 where the jump in the solid line takes place.

Now the first observation is that there was a high growth in employment before Work Choices. Moreover, the model looks singularly unimpressive as a model of employment over that period. Things seem to change after the advent of Work Choices. Such marked differences inevitably makes one suspicious.

There is a simple explanation of the “improved fitâ€?. It is a consequence of the way in which Blandy allows for Work Choices in his regression – through a so-called “dummy variableâ€?. This variable goes from zero before the introduction of Work Choices to the value of one when Work Choices comes in during March 2006. 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 artifact of the way in which the analysis is done and has no causal significance whatever.

Obviously there was some other factor (or factors) – X – that caused the big rise in employment before Work Choices. If X is still operating after March 2006 then its effects will be attributed in Blandy’s regression to Work Choices. Unless an attempt is made to control for X, there can be no confidence that Work Choices has had any effect.

To emphasize the foolishness of all this I replaced the Work Choices dummy with a Rudd dummy. It is constructed in the same way as the Work Choices dummy – zero until Rudd took over as leader of the ALP in December 2006 and then unity. Interpreted in the same way as Blandy interprets the Work Choices results it would imply that Rudd has added many jobs- in fact 10% more jobs per month than Work Choices did.

This type of work is akin to the famous quote from Wallace Irwin that “Statistics show that of those who contract the habit of eating, very few survive�

Adrian Pagan

Professor of Economics, UNSW
Processor of Economics, QUT
Chair, National Centre for Econometric Research

41 thoughts on “Guest post from Adrian Pagan

  1. Can I reiterate what John Quiggin says in response to Mugwump. Adrian was my Phd supervisor and we have coauthored several papers over about 11 years. I don’t know what his personal politics are. Nor do I care. But what I do know is that I have never seen a shred of evidence that personal politics have affected his analysis. On the contrary, one of the most enduring lessons that I have learnt from Adrian is to focus on the idea rather than any other feature such as the person. The other lesson is to have the personal courage (of which he has heaps) to be willing to debate ideas irrespective of who you are debating.

    Australia would be a much better place if more academics were willing to debate things in public. One reason they are reluctant to do that is the noise created by people who have no ideas themselves but are all too willing to shout “political bias” drowns out the content of the debate.

  2. Harry Clarke, raises two questions.

    The answer to the first is that if you read what I said it was that I didn’t see Labor’s IR laws as being particularly threatening to business and it could even be favourable. So I don’t see why there would necessarily be noticeable stock market reaction.

    The answer to the second is that econometricians conduct structural break tests for single breaks all of the time so I don’t see the problem. Moreover, in finance event studies are widely used to assess the effectof say a takeover on the value of a company. One can do the same thing here except that it is employment that is of interest. The only tricky thing is that one needs a forecast of employment that is independent of Workchoices. Back in June 2006 in the Fin Review I suggested using the Treasury budget employment forecasts because Treasury was forebidden to do any modelling of WorkChoices and so they could not affect its forecasts. Its a bit rough but it remains the best approach I have seen to date.

  3. Mugwump, your claim that ‘The use of statistics to sell a message rather than tell the truth knows no political boundaries, but the efforts of academics dispelling such messages are curiously one-way.’

    The first part of this claim has already been completely demolished by other folks, so I won’t bother commenting any further on this. But I will take issue with your claim regarding academics.

    This silly claim about academic bias, so often trotted out by the likes of you, is as offensive as it is pathetic. Do you and your ideological amigos really believe this tripe? Throughout this election there has been bucket loads of other academic opinions being trotted out that are very supportive of Coalition policies (e.g. a number of folks from the Melbourne Institute in the Oz). And while Richard Blandy might have to go back to econometrics 101, he is still an academic (adjunct professor of economics at the University of South Australia). Can anyone else smell the irony?

    What you’ve jumped into is a debate BETWEEN academics: there is nothing ‘one-way’ about that. In the future, how about trying to play the ball and not the man.

  4. Hi Don (et al)

    Is there another way to characterise aggregate demand so we could compare the response of employment in this boom to the response of employment in another earlier boom. The microeconometric equivalent of this is using propensity scores to match your treated and controls – this seems to be one way to sort out the effect of Work Choices.

    cheers

  5. P.S.
    Alternatively if we had firm or plant level data and an argument could be made for differential impacts of work choices on different types of firms (like we did for the Workplace Relations Act) this might be another way to get the same thing. I don’t know enough about Work Choices to do this (and don’t have the data!)

  6. Don you said:

    ‘Since businesses are forward looking it makes sense for them to increase hiring in advance of the detail of the laws and their introduction.’

    I am saying that if this was true in labour markets it is even more likely to be true in equity markets and so far it has not been.

    I cannot see how Treasury forecasts of employment (ignoring policy effects) have much to do with the employment decisions of firms anticipating policy changes. If the shocks were repeated you might get away with it but a once off shock I don’t see how. That it is done all the time does not neceesarily make it sensible.

  7. Gee and here was I thinking that our strong employment growth was due to the terms of trade boom and easy monetary policy.

    Let’s face it in the current climate Workchoices effects only the most vulnerable members of our society.

  8. Mugwump, your claim that ‘The use of statistics to sell a message rather than tell the truth knows no political boundaries, but the efforts of academics dispelling such messages are curiously one-way.’
    The first part of this claim has already been completely demolished by other folks.

    You claim Labor or the green left are pure in their use of statistics yet the Coalition are corrupt? Good luck with that – I look forward to your thesis.

    But I will take issue with your claim regarding academics. This silly claim about academic bias, so often trotted out by the likes of you, is as offensive as it is pathetic…

    as it is true.

    I was once an academic. Worked in two Australian Universities, including the IAS at ANU. I’d guess less than 10% of my fellow academics at either institution were of the right.

    If I am so off the mark, where are the references to Pagan’s myriad attacks on poor applications of econometrics by the left?

  9. “You claim Labor or the green left are pure in their use of statistics yet the Coalition are corrupt? Good luck with that – I look forward to your thesis.”

    No, I believe the claim is that all statistics should be viewed with a degree of skepticism.

    But when one individual is shown to be misusing statistics, the proper response isn’t “everybody’s doing it”.

    As I’ve noted before, most of us here are interested in establishing the facts, you on the other hand seem primarily driven by a desire to defend your pre-existing ideological prejudices.

    Personally, I take the view that if my position need to be defended by distorting or ignoring the facts, that means I need to re-examine my position.

  10. Great work. I was thrilled to read this article, and Gittins’ discussion of it in SMH. The underlying data say it all – Blandy’s model is just a level shift plus noise. I will be interested to see Adrian’s estimate of the effect on jobs of Howard’s exit in about a week.

    I am glad to see Australia’s academics playing a key role in this election. It is so easy for politicians to make glib claims about policy effects, yet those of us who spend our lives estimating these things know how perilous it can be. I can only wonder how the Liberals would have got the econometric result they wanted had the legislation been introduced in a downswing. Howard is a clever politician indeed.

  11. @14 Don Harding says:

    “I don’t agree with your statement that “Obviously there was some other factor (or factors) – X – that caused the big rise in employment before Work Choices.â€? The reason is that after the Coalition won the Senate majority in October 2004 it was evident that they would get rid of unfair dismissal laws and change labour laws in a way that reduced the cost to business. Since businesses are forward looking it makes sense for them to increase hiring in advance of the detail of the laws and their introduction. If one looks at the data the increased employment growth started about November 2004.”

    The extent of the changes to unfair dismissal and other laws was not apparent in November 2004. Howard’s announcement that the unfair dismissal laws would be removed for firms with less than 100 employees (as opposed to the 20 employees they had promised) was not made until the end of May 2005 and came as a considerable surprise. Businesses would not have factored such a change into their decision making in November 2004.

    @2 SATP says:

    “Abolition of workchoices may cause some problems at my place. If we have to return to the previously used awards that is.”

    Of course SATP you won’t have to go back to awards after a change of government. You can make overaward payments, like the majority of businesses did. If you want to reduce penalty rates and the like you can just use an employee collective agreement, provided that you can get the majority of employees to agree to it. If you can’t, then it’s probably not a good idea anyway.

    @12 Melanie says:
    “According to the Workplace Authority’s data, 9.3% of workers are currently on AWAs.”

    The Workplace Authority estimates of AWA coverage are grossly exaggerated, as they assume that every worker who signs an AWA remains in their job for a minimum of three years.

  12. Ian Gould, Sam et al
    While Pagan may not be biased, there are plenty of others in the IR field who are. A classic is David Peetz, who (for example) argued that IR reform in New Zealand reduced productivity whereas various researchers have shown that productivity in NZ improved after they did IR reform.
    See: Diewert & Lawrence (1999) “Measuring New Zealand’s Productivity� Treasury Working Paper Series 99/5-1 reach the conclusion that “After 1993 there was a productivity surge. This is likely to have been aided by the effects of the labour market reforms of the early 1990s, among other things� (p xx)
    Also see ). Conway & Hunt (1998) “Productivity Growth in New Zealand: Economic reform and the Convergence Hypothesisâ€? Reserve Bank of New Zealand Discussion Paper G98/2 and Perry (2006) “Do Workplace Contracts Harm Labour Productivity Growth – a reconsideration of the macroeconomic evidence from New Zealandâ€? Australian Economic Review 39(4)

  13. Harry, read what I said carefully. I never said that Workchoices *will* reduce jobs in the long run – I said it was “another question”.

    You’re balancing two effects that work in differing directions – more people will be in low-skill jobs because they’re easier to get, fewer people will be in such jobs because they pay less. The result on total employment is theoretically ambiguous, and given where we’re starting from (ie essentially full employment, so unemployment can’t be reduced much) you’d be brave to predict the empiric outcome.

    Of course, to the extent that Workchoices delivers increased employment of marginal workers it will tend to *reduce* average productivity by a simple compositional effect.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

w

Connecting to %s