Big taxee, no workee

Now here’s a paper whose conclusions should make John Howard and Ross Gittins both nod their heads.

Labour Taxes and Work Hours in Australia
Anton Hallam & Ernst Weber
In the 1970s, work hours in Europe were similar to work hours in America, but today Europeans work less than Americans. Prescott (2004) attributes the decline in European work hours to an increase in the effective marginal tax rate on labour income. The Australian labour market experience confirms that the taxation of labour income is an important determinant of the decision to work. In Australia taxes and work hours did not change much in the long-run, but Australian work hours rebounded after a temporary increase in taxes in the 1980s. The resilience of Australian work hours suggests that a return to the tax rates of the 1970s would restore the European labour supply.

The analysis is pure time series, with the problems that normally entails (what if something else changed to affect both taxes and work hours?). But the authors’ methodology is refreshingly upfront – right down to showing us their Excel formulas in the appendix tables.

About these ads
This entry was posted in Tax. Bookmark the permalink.

18 Responses to Big taxee, no workee

  1. Kevin Cox says:

    What I find fascinating is the size of the effect. If tax increases by 1% then the work hours appears to increase by about 1.7%. For those people in “poverty traps” where their effective rate of taxation is way above 50% because of loss of benefits then if their effective tax rates were reduced to say an effective 20% then this would increase the available labour supply.

    It would be really interesting to see what would happen if we introduced some form of minimal social wage for everyone with a 25% taxation scale on all monies earned outside the social wage.

    It would seem that people who lived solely on benefits are likely to work a considerable number of hours as they would not lose their benefits.

    If the government thinks it is a good idea to increase the paid labour supply then that may be a policy well worth considering.

    However, we do not have to implement such a radical policy change to find out. From July 1st this year we will have such an experiment. All superannuates over the age of 60 will be able to work for an effective 15% tax rate if they have a 100% salary sacrifice. It is going to be fascinating to see the effect on the hours worked of people 60 years and over who are superannuates and who were previously unable to arrange their tax affairs to reduce their taxes on work done.

  2. derrida derider says:

    By coincidence I’m reading up on this stuff for work at the moment. Its precisely the imputed size of the effect that makes Prescott’s study (and hence this somewhat less sophisticated one) controversial, because they appear to imply labour supply elasticities way out of line with those that forty years of microeconomic research into these elasticities finds.

    This link is to a paper supporting Prescott, this link is to a paper arguing againt the thesis, this link is to a paper providing an alternative explanation of the effects Prescott found, while this link is to a paper that argues that the discrepancy is only apparent rather than real.

  3. Matt Cowgill says:

    formulas or formulae?

  4. backroom girl says:

    “It would be really interesting to see what would happen if we introduced some form of minimal social wage for everyone with a 25% taxation scale on all monies earned outside the social wage.”

    Well, I’m sure it would be interesting Kevin, with the most interesting thing being how on earth the Government would fund it.

    The only study into the feasibility of an Australian basic income that I know of, by Peter Dawkins when he was at the Melbourne Institute, found that you would have to tax every dollar of private earnings at around 55 per cent (from memory) in order to pay for a basic income for everyone. While that would undoubtedly be simpler than the tax-transfer system we have now, in EMTR terms it’s not all that huge an improvement. (Some people’s EMTRs would go down, but an awful lot of people’s would go up.)

    I suspect that the labour supply implications of people knowing they faced such a high tax rate could actually be significant (in the opposite direction to those you propose). It would certainly be a big incentive to work in the black economy, I would have thought.

  5. Sacha says:

    formulae of course

  6. Fred Argy says:

    Perhaps it’s my age but I can’t follow the debate. I thought the big debate among economists was about whether higher taxes encourage people to work harder or less. Where does the new literature take us on all this? Prescott obviously thinks high taxes deter work.

    Derrider you have studied the wider literature (which I can’t access). What do you make of it in a nutshell? Please.

  7. conrad says:

    I’d be interested in the answer to Fred Argy’s question too.

    I’d also be interested in a second cross-cultural factor. Based on pure speculation, I wonder to what extent high taxes are a bigger loss in countries like Australia where people (at least the ones you most need in your workforce) can move easily to other countries that are culturally almost identical (apart from tax rates) than countries where this is not the case (say, like France). It seems to me than in places like France, people have no real alternative but to work more if they want the extra money but that isn’t true of Australia.

  8. derrida derider says:

    Prescott is a macroeconomist, Fred, of the “real business cycle” persuasion. He thinks recessions are an optimal reaction to real supply-side (rather than aggregate demand) shocks and that most recessionary unemployment is because people in rationally withdraw their labour as part of the mechanism of adjustment to these shocks (don’t laugh – they gave him a Nobel prize for this stuff!). Think of it as the (relatively) respectable face of supply-side economics.

    Now the problem all the supply-siders have is that they have to assume large uncompensated elasticities, especially elasticities of labour supply and elasticities of saving. And as you know, sixty years of empiric microeconomics has generally failed to find such large elasticities.

    Some supply siders (eg Martin Feldstein) simply do their own micro studies that invariably find the large elasticities. Somehow these studies tend to defy replication by others.

    Some assert that longrun elasticities are much higher than short run ones because of the effect on social norms, and that we can only measure short run ones. This might even be true, but unfortunately its (a) almost untestable, and (b) irrelevant to short and medium term macroeconomics.

    Prescott takes a third tack – that the micro people are measuring the wrong thing. He concendes that the elasticity of labour supply at what labour economists call the intensive margin (ie how many hours an individual chooses to work) is small, but (translating from his macro jargon) claims that it is larger at the extensive margin – ie whether people choose to work over their lifetime or not. Apart from the fact that labour economists took that point on board 30 years ago and the measured extensive margin elasticities are still way too low to make Prescott’s case, this is very demolishable even within his own chosen framework – see
    paper by two rightwing but non-RBC macroeconomists (one of them was Thatcher’s macroeconomic adviser!)

    If I sound hostile to this garbage, I am.

  9. Peter Whiteford says:


    The third link in your original post doesn’t seem to work, or maybe I’m not authorised to go to places where you can – can you give a citation, so I can google it?

    There has been some discussion of the Prescott analysis in other blogs. Part of my problem with it – or maybe just the US-based discussion of his paper – is the idea is that there is this country called Europe where people are all highly taxed and none of them work very long hours.

    Now Europeans do pay higher taxes than americans, but the ratio ranges from about twice the US level (in Denmark and Sweden) to about 50% higher in Greece. See,2340,en_2649_201185_37504715_1_1_1_1,00.html

    Aggregate hours of work are higher in the US than in a lot of european countries, but Greeks, Spaniards, Poles, and Czech (and also Australians and Kiwis) all work longer hours than Americans, although paying higher taxes – see, for example - and

    Now taxes may be part of the story of differences in working, but my not very original conclusion is that other influences are also important – particularly insitutional differences. for example, in Nordic countries, significant parts of taxes are going to finance child care, and employment-related support, while in southern Europe significant parts of taxes finance pensions, and so on.

  10. Christine says:

    Don’t have terribly much to contribute of substance here (have read much less widely on this recently than DD, but previously came to a fairly similar conclusion – but I’m a micro type, so that’s not surprising perhaps).

    But I do want to say that it’s fantastic that there are discussions like this one that I get to listen in to. Thanks very much to all, esp to DD, for sharing.

  11. Fred Argy says:

    Thanks DD. You are a sport.

  12. Fred Argy says:

    Peter Whiteford, your comment is spot on and raises a much wider issue. Whether one is discussing work effort or workforce participation or productivity growth or whatever in economics, it is not the level of taxes per se which determines outcomes but how and where the government spends the money and the efficiency and neutrality of the revenue-raising system.

    That is, it is the METHOD of redistribution as much as the SCALE of redistribution that is crucial to the reconciliation of efficiency and equity. If there is one message I have been trying to convey for the last ten years it is this one.

  13. Andrew Leigh says:

    DD/Peter, thanks for sharing your knowledge on this. Prescott must of course be right in terms of sign (extensive>intensive, lifetime>single year), but his magnitudes are enormous.

  14. derrida derider says:

    Thanks all – it’s been good stuff. Peter, that reference is:

    Christopher A Pissarides. International Economic Review. Philadelphia: Feb 2007. Vol.48, Iss. 1; pg. 1

    My distaste for supply-siders apart, I can’t help feeling that Prescott keeps making the error that macroeconomists of all schools are traditionally prone to – not looking closely enough at the nitty-gritty of the data and hence ignoring both useful information and serious problems. Treating “Europe” as one case is a good example.

    Of course we micro dabblers have our own traditional errors – eg failing to think about general equilibrium effects is a common one.

  15. derrida derider says:

    Oh, as a PS, Pissarides is coming out to Oz in September/October. Its a good chance to talk to him about this stuff for those interested.

  16. Fred, while I agree with you up to a point, you can push the “efficiency of the redistribution system” point too far. Suppose that a Ramsey tax method was used to raise revenue. This will minimise the deadweight loss associated with raising a given amount of revenue. Nonetheless, the larger the required revenue, the greater the deadweight loss.

  17. On the macro versus micro estimates/calibrations of labour supply elasticities, I tend to believe the micro results more than the macro ones. This is not surprising, since I am a microeconomist. However, it is perhaps worth noting a couple of things about this issue. First, while the micro studies might suggest that much of the adjustment takes place on the extrinsic margin (participation as opposed to hours worked), in a typical macro model this will still appear as though it is taking place on the intrionsic margin. The reason for this is that typical macro models simplify things by assuming a representative agent. What in reality amounts to many participation decisions will appear in a rep agent model as a single hours worked decision. Second, the micro studies typically relate to a specific labour marketr. This is especially true of the really good micro studies (see for example Oettinger 1999). When elasticities from a variety of disparate labour markets are aggregated up, it is possible that weird things happen. It is particularly important to note that the labour markets that are used for the micro studies are typically chosen because they allow the researcher to control for various things and avoid endogeneity problems and selection bias problems and the like. These markets can often be rather unrepresentative. For example, Oettinger (1999) studies the daily labour supply of food and beverage vendors in a single baseball stadium over an entire baseball season. This does not mean that the labour supply elasticities needed for the macro models are sensible. But the mere fact that they do not match those obtained from microeconometric studies does not automatically rule them out.

    Reference: Oettinger, GS (1999), “An empirical analysis of the daily labour supply of stadium vendors”, Journal of Political Economy 107(2), April, pp. 360-392.

    A particularly useful survey of the RBC literature is:

    King, RG and S Rebelo (2000), “Resuscitating real business cycles”, NBER Working Paper Number 7534, National Bureau of Economic Research, USA, February.

    This is available online at the following web site: .

    Disclosure: Gerald Oettinger taught me a course on graduate labour economics.

  18. derrida derider says:

    Damien, you’ve pretty precisely made the point that Prescott makes – much better than I did. His critics’ point, though, though, is that if you take apart these aggregation effects in detail you’re still left with a model that can’t fit the known facts.

Comments are closed.