Paid maternity leave again

It’s not long since Paul Frijters raised the subject of paid maternity leave here, inspiring a long and stimulating discussion in the comments. The topic is in the air again, largely because the Productivity Commission has been looking into the issue. Unfortunately I stll haven’t formed an opinion, so I’m hoping that a little more discussion will help me do so. Joshua Gans has a very useful piece on it in The Age today, which I may well adopt as my own opinion.

Before coming back to Joshua’s suggestions, I’d like to summarise the main arguments and the issues they raise.

Employer-provided or employer-subsidised maternity leave is properly regarded as a wage premium using second-order price discrimination. That is, just as cinemas offer discount tickets to pensioners because on average their ‘reservation price’ is lower, employers offer higher wages to a particular group of workers, whose reservation wage is higher, on the basic of an objective characteristic (they are having babies). For private sector employers, this allows them to retain productive workers who might otherwise quit, but is cheaper than paying higher wages to everybody. It would make a lot of sense for the public sector to use this approach too, especially with nurses, and maybe also with teachers. In both cases the decision should be based on profit or cost-benefit considerations for the enterprise in question.

But universal state-subsidised maternity leave, which is what the PC is considering, is a different story, and would need to be based on clear and widely shared social objectives. Unfortunately, the case for it is usually emotive and muddled, like this recent polemic from Pru Goward, whose recommendations were ignored in 2002, and who is still seething:

It is still unbelievable to me that both sides of government are apparently content to see Australian women with babies only a few weeks old struggle back to work because their partner has lost his job, the overtime has dried up or because the average mortgage can no longer be met from one salary […] Innumerable studies showing that women who left their infants too early to return to work were more likely to suffer depression and related illnesses were ignored, despite the cost to employers of paid sick leave. Productivity arguments, labour force retention arguments, fairness arguments (remember, you can be given paid leave for a broken foot but not for a Caesarean delivery) and rights arguments fell on deaf ears. Fertility arguments were received differently apparently it is OK for women to be barefoot and pregnant thus the lump sum maternity payment.

This could just as easily be an argument for unemployment insurance, or mandatory 50% deposits on home loans. The case for paid maternity leave needs to be made much more rigorously than Ms Goward’s, with clear answers to a number of questions. First, is it meant to be form of income distribution in the name of fairness, equity or social insurance; or is meant to modify people’s behaviour? If the former, is it well targeted? If the latter, is the behavioural change in the interest of the mother, the child, or broader society? If it’s the mother or the child whose interest is being served, is it really the state’s business to interfere at all? If it’s broader society, what exactly is the positive externality associated with working women staying at home for a few months? And how can we quantify this externality?

As far as I can make out, a government funded paid maternity leave scheme could be one or more of seven things:

1. A device to redistribute income to households with young children. For those who believe that the main purpose of income redistribution is to promote equal opportunity for children, maternity leave might seem a practical way to support households with children. It would presumably be means tested. In this aspect, maternity leave is one of several kinds of subsidies for families with children, including the baby bonus, family allowances and, indeed, free school education (given that school is compulsory, making it free is effectively an income subsidy for families with school-age children, should they choose to accept it). However, in the greater scheme of things, a few months of paid maternity leave isn’t a big enough sum either to count as significant income distribution, or to motivate a couple to have an extra baby if they otherwise wouldn’t. In any case, if assisting children is the only goal, a family allowance or baby bounus is a better way to meet this objective, since it helps non-working mothers too.

2. A device to redistribute income to specifically to working mothers with children, because they ‘deserve a break’. As far as income redistribution is concerned, it’s not obvious that two-income families are more needy than one-income families. As Joshua Gans points out in his submission to the PC, ‘on a social contracting basis, there is no rationale for a cross-subsidy.’ It’s true that losing the income for a few months is difficult if you’re used to a higher level of consumption, but the obvious solution is for the household to budget for that.

3. An incentive to have children, or more children. The premise here is that Australia needs to increase its population faster — to make us more productive, to redress a generational imbalance, or in the interests of defense against invasion. But the same objections apply as to (1): it’s a small amount in the general scheme of things; and it discriminates for no apparent reason against non-working women. However, it might be limited to working women on pragmatic grounds, on the reasoning that having an extra baby is a bigger sacrifice for a working woman than for a non-working woman, so they need a bigger financial inducement.

4. An incentive for working mothers to have children who otherwise wouldn’t, because it’s good for them to have children. This falls into the category of paternalistic rationales: the assumption, obviously, is that women don’t necessarily know what’s good for them, or at least don’t realise what they are missing out on until they have that baby — that they need to be coaxed off the career treadmill to experience the joy of motherhood. This is not a logically incoherent argument, but it will be hard to get past our libertarian brethren, who don’t even approve of mandatory baby capsules.

5. An incentive for women who were going to have babies anyway, to pursue careers, because it’s good for them to have careers. It’s better that they should be in the workforce so they don’t lose their skills and become unemployable; later when their children are of school age, they will appreciate the income. Implicit in this line of reasoning is the assumption that low-income households, if subsidised for breeding per se, would not be sufficiently forward-looking to choose the career option voluntarily. That’s OK, but, once again, it’s a paternalistic policy, because it supposes the beneficiary to be myopic.

6. An incentive for women who are planning families anyway to keep their jobs, because the nation needs them to work. This is similar to the previous argument, except beneficiary is the nation rather than the woman or her family. It’s the same tactic that the firms use when they retain employees by offering maternity leave (instead of higher pay, which they have to give to everyone), except that the aim is to boost GDP. The purpose such a policy would have is not obvious; it would create more taxable income, but that could hardly be a legitimate objective in itself.

None of aims 4-6 imply a need for paid maternity leave. They could all be be served just by giving working mothers money, and letting them decide for themselves whether to take unpaid maternity leave, or spend the money on childcare or nannies.

7. An incentive for working women to stay at home and bond with their babies rather than put them in a creche. On this rationale, paid maternity leave is essentially a subsidy in kind like food stamps: help is available, but only if you make the sensible choice. This is the only argument for which there is no straightforward objection. It’s a policy measure that’s aimed at the right target, and it’s not necessarily paternalistic, especially if we consider the child as the beneficiary — and society at large, which benefits from having well-adjusted, breastfed children.

We would, however, see low-income households with no children, and low-income households with stay-at-home mothers, cross subsidising high-income households with working mothers. In correcting an externality, an inequity would be created .

Applying a means test to the maternity leave allowance would be one way to overcome that objection. But means tests are anathema to some. Eva Cox, for example, sees paid maternity leave as ‘the basis for a wider set of policies on better conditions for work/family integration’, whatever that means. She thinks means testing is ‘grossly insulting to women in paid work as leave is an industrial condition and Army Reserve leave is not means tested, though government subsidised.’ What’s more, ‘it disrespects women in paid jobs by denying them, and men, paid leave for time spent in parenting a new baby’. (A quick Google tour of Cox’s interventions on this topic reveals that most of her arguments are variations on the ‘insulting’ theme.)

The real problem with means testing is that it limits the efficacy of the policy as a corrective to market failure. One of Joshua Gans’ ideas is to offer income contingent (i..e HECS-style) loans as a way of recouping some of the money after it has achieved its purpose.

But another of Joshua’s ideas — and this is the one he wrote about today — might bypass the cross-subsidy issue altogether, while still fixing the externality.

Joshua wants to subsidise the firm rather than the worker. He advocates a “return to work” tax credit, which the government would give firms when they re-hire women who have taken maternity leave. Women eschew maternity leave not so much because it’s unpaid, as because taking time off disadvantages them in the competition for higher responsibility and promotion. He cites a study by Isabel Metz confirming that taking leave, even just a few months of it, derails women’s careers, and deepens labour market segmentation.

… not only would employers have an incentive to attract workers back, they would have an incentive to encourage them to take leave in the first place. Indeed, they might even choose to provide paid leave themselves, as many organisations already do. Moreover, the more highly paid the employee, the greater the incentive to do this. That means that workplaces will want to encourage leave by fathers as well as by mothers.

The trump card for advocates like Cox, Goward, and Natasha Stott-Despoja is that Australia and the USA are the only OECD countries without paid maternity leave. That makes us sound pretty uncivilised. On the other hand, it would be OK to be the only country without farm subsidies. So let’s get the arguments right before joining the throng. Perhaps we could be leading it instead, in a slightly different direction.

This entry was posted in Economics and public policy, Politics - national. Bookmark the permalink.

17 Responses to Paid maternity leave again

  1. Great post, James – one for my BBP08 collection.

  2. Pingback: CoreEcon » Blog Archive » Talking points

  3. Patrick says:

    Great post, James. Personally I like the baby bonus as-substitute-for-maternity-leave because(I’m a beneficiary, and) I am keen to encourage non-working mothers as well as working mothers. I do however recognise the benefits of mothers returning to/joining the workforce after a few years.

    That said, my employer already offers about the most generous fully paid primary carer leave in Australia (and more generous than that in most French companies, besides banks). Based partly on that and my experience with other very large employers, I would rather any subsidy be targeted to smaller companies. On that basis, I think Gan’s return-to-work credit is probably on the money.

    This is because I think those ‘career-derailing’ effects would necessarily get stronger the smaller the company. In (most) large companies,
    – there is a fairly high level of turnover anyway,
    – they tend to be more geared to ‘alternative’ career development paths, and
    – they are highly likely to have already adopted maternity leave because they practice ‘second-order price discrimination’ as (implicitly) a part of their standard remuneration policy.

    All of which suggests that maternity leave is more easily accomodated by large companies.

  4. Peter Whiteford says:

    James (and skepticlawyer)

    I think you are missing the point – or at least you have never read any of the non-Australian literature on the rationale for paid maternity leave.

    James, you say ” Employer-provided or employer-subsidised maternity leave is properly regarded as a wage premium using second-order price discrimination.”

    A wage premium for mothers (or women)???

    The premium is negative. Having children has large direct costs and even larger indirect costs – and most of the indirect costs are borne by women. For example, using British data from the 1980 Women and Employment Survey Joshi (1990) found that a woman who had two children would give up nearly half the lifetime earnings that she otherwise might have had, with these lost earnings due, in roughly equal parts, to reduced participation in the labour market, shorter hours of work, and lower wages. Using data from the Australian Negotiating the Life Course Survey (NLCS), Gray and Chapman (2001) find that women with one child are estimated to earn 63% of what they would have earned had they remained childless.

    Around 30-40% of the female workforce in Australia currently are entitled to paid materniy leave – but guess which 30%? Strangely, its the people who work in the public sector and the well-paid people in the private sector – people who work in banks, insurance companies and law firms and large companies.

    So who would benefit from paid maternity leave – given that low income mothers are covered by parenting payment or the current baby bonus then the people who would benefit arwe the lower half of of the wage distribution of women who already have jobs.

  5. I’m not agreeing one way or the other, Peter – I just found James’ post well written, thoughtful and well-explained. I’m one of those nasty libertarians who thinks it’s all best left to private contract (while recognizing that Joshua’s proposal is damn clever).

  6. James Farrell says:

    Thanks, Helen and Patrick. I’m always pleased when I can cater to your corner of the market.


    Having done my best to work through the arguments systematically, it’s naturally disappointing to learn that I’ve missed the point. But let’s deal with the wage premium thing first, since this is not a matter of opinion or interpretation. If two workers get the same salary for the same work, and one gets paid maternity leave on top of that, that’s a premium. From the firm’s point of view, the cost of raising children is neither here nor there. It’s not like recreation or sick leave, because the probability of becoming a parent is much higher for some people than others. So paying them a premium is analogous to offering concession tickets to pensioners — but I’m not sure whether you were disagreeing with this for some reason, or whether you simply didn’t understand it. In any case, the 30-40% percent of employers you refer to apparently see the advantage of this kind of scheme.

    The rest of your comment purports to show that low-wage working mother are missing out on something that everyone else gets. But they do get the baby bonus, and they presumably would be entitled to the means-tested parenting payment if they went on unpaid leave, although I admit I’m not an expert on the rules governing that entitlement. What you and Eva Cox and Pru Goward seem to have noticed is that the income of parents who take unpaid maternal leave is more volatile than other people’s. Volatility motivates much of our social security system. We tend to have social insurance for things that effect everybody at some stage (getting old), or might effect everybody (getting sick). But getting pregnant when you’re in paid work isn’t in either of these categories. So, unless you can set up a wage-income tax that applies only to people whose probability of having more babies is above some specified threshold, then maternity leave will always be a redistribution of income to parenting workers from everyone else, and therefore needs to be properly justified.

    I said at the outset that I haven’t made up my mind, but so far I find only #7 to be persuasive, and also #3 if it can be demonstrated that our low fertility rate is a national problem.

  7. Jennifer says:

    Very interesting post. I think you’ve missed one of the potential points of government funded maternity leave – and that is to enhance women’s attachment to the workforce. If women lose their jobs completely, and drop out of the workforce for a few years, then it is much harder for them to get back in, and the country has lost a productive worker for much longer than the year of maternity leave.

    It is a long term investment. But most analysis I see about it focuses on the actual year of the maternity leave, and misses that long term investment.

    Higher income (and skilled) women, who are most likely to receive maternity leave, and employers of those women both have an incentive to keep their attachment to the workforce – the employee has a big investment in her skills, and the employer will spend a lot of money hiring a replacement.

    The lower skilled/income woman and her employer will have less incentive. So the negative externality is that the economy as a whole has an interest in an investment in keeping that woman attached to the workforce, but the investment required is not worthwhile either for the woman or her employer.

    Which is why it is worth calling it maternity leave rather than providing a baby bonus regardless of employment status. And why it is most important for the lower paid, so it is probably reasonable for the amount to be capped at some reasonably low level.

  8. James Farrell says:

    Jennifer, if the problem is that women can’t take time without getting sacked, then that’s an excellent argument for unpaid maternity leave (which I fully support as an entitlement). If the problem is that firms will find other ways of penalising them even if they have to take them back, then Joshua’s scheme is a good remedy. Your externality theory sounds a bit like my Item 6, but it needs a bit of fleshing out. If the woman needs a bit of extra inducement to return to work, and her high productivity justifies this, why wouldn’t this potential profit be captured by the employer in question, and why doesn’t that employer offer the relevant inducement? If it’s just a matter of ignorance and stupidity, well, that is indeed a market failure, but if you read Joshua’s column you’ll have seen that he foresees that, under the regime he proposes, firms will offer inducements of their own — possibly including PML.

  9. Jennifer says:


    I have no evidence to offer, only anecdote, but most paid maternity leave is only paid on condition of a mother returning to work. So paid maternity leave will increase the chances of the mother remaining attached to the workforce. For high skilled workers, the benefit to the employer of that specific employee is high, so they will offer it (I myself would be entitled to 12 weeks paid maternity leave, should I plan to have another baby, which I don’t). And my employer reckons that the direct costs of replacing me would be around 30% of my salary (recruiting costs, which includes the time of my boss in trying to find a replacement, rather than doing his/her own job), plus another 70% or so in indirect costs (lower productivity of my replacement and my peers and my team while my replacement is trained up).

    But for a lower skilled woman, the employer can find a replacement reasonably easily, so both the direct and indirect costs of replacing her are a much lower proportion of salary. So the employer doesn’t care about retaining that specific woman, and is less inclined to offer paid maternity leave to increase the chances of her coming back from maternity leave.

    But the economy as a whole loses if she doesn’t come back from maternity leave. She is less likely to go back to work at all after her baby is born – perhaps for a few years. And once she does, she is less employable, and will find it harder to find another job. So she might not bother at all – perhaps until the children are in high school. So the taxpayer has an interest in maintaining her commitment to the workforce, by effectively providing her a bribe to go back to work somewhere.

    I have comments on some of your other reasons, but perhaps I should write my own post (as I have been intended for a while!) rather than clogging your comments.

  10. Jennifer says:

    Oops – just realised you are right that your reason 6 is the same as mine. But I still think the long term effects are real, and of value to society as a whole.

  11. Backroom Girl says:

    I have a slightly different perspective on this. Looking at it from the perspective of a government (who should be representing the concerns of the society as a whole), I think that governments want two things – for women to keep on having babies so that the ‘home-grown’ population doesn’t go into decline and for women to stay attached to the workforce for as much of their lives as possible.

    Now, I guess it is possible at one extreme to have one part of the female population bearing the babies and the other part in the workforce, but I don’t think that would be desirable. So if you want the same women to have the babies and to supply the paid labour, how might you go about ensuring that happens?

    One thing that is undeniable is that the indirect costs of child bearing and rearing are mostly borne by women in terms of income foregone. Many women, and it would seem many fewer men, probably see this as unfair. And while there is clearly no simple formula for calculating what these costs are, better educated and more highly paid women stand to lose more financially if they take a period out of work and/or work part-time for any significant length of time. So, for me, paid maternity leave, from the point of view of the mother, is simply a small compensation for her personal opportunity cost of helping to produce the next generation. Two categories of women currently do not suffer an income loss in the immediate period following birth – better-educated, better-paid women in good professional and white-collar jobs and income support recipients, some of whom even get to move onto better income support (single parent pensions) as a result of giving birth.

    So the current paid maternity leave debate is really about whether it is appropriate to offer some level of financial compensation to those working women who by choice or chance don’t happen to work in jobs that offer paid maternity leave. These women are probably those most likely to drop out of paid work for too long, as Jennifer fears, but they may also be the ones most likely to have to go back to work too early (from the point of view of the baby) if they really can’t afford to take much time off.

    I actually think that apart from the issue of financial compensation for income lost that James’ seventh objective is an important one and the most likely measurable outcome from offering some form of paid maternity leave safety net. I think many mothers who currently return to work within the first two years would use additional financial assistance (whether by way of baby bonus or paid maternity leave or both) to enable them to stay home longer with their newborns.

  12. Alison says:

    Sorry I’m late to this discussion but I’ve just been catching up on the week’s news and felt impelled to add a quick comment here.

    James, you lay out seven possible arguments to justify government funded paid maternity leave in order to clarify what the clear and widely shared objectives are that would justify a universal government-funded paid leave – mostly from an econonc or sociological perpsective – but it seems to me there is one argument that you’ve left out, and that is work as a basic right. Women have the right to work and the Australian government endorsed it as an ‘inalienable right’ by ratifying the Convention on the Elimination of All Forms of Discrimination against Women. The same convention also outlines women’s rights to protection of health and the function of reproduction in relation to employment. Since pregnancy and childbirth are physical conditions that bring about a temporary incapacity to work, they should be considered as categories of social insurance and subsidised by the government as part of its obligation to safeguard both the right to work and health.

    This basic right should be guaranteed for all women in the workforce and called maternity leave since only women can give birth. But once a woman has recovered from childbirth and is physically able to work again, the length of leave and question of who pays for it, is a different issue. For clarity’s sake it should be called parental leave and debated separately from maternity leave. Either one parent stays home – and to assume it is the mother is discriminatory – or the child goes to day-care. In any case, someone has to pay and the different ways this is handled will reflect the values, priorities, labour structure etc. in different countries.

    So I can see how your seven arguments are a useful point of reference for debate – once we have acknowledged maternity leave as a basic right for all women.

  13. Pingback: Paid Maternity Leave « Penguin unearthed

  14. James Farrell says:

    Sorry for the delay in responding to the very thoughtful comments that have appeared since I last looked. It’s been a busy weekend.

    Jennifer: Let’s suppose for argument’s sake that taking maternity leave put your unskilled worker into a temptation zone where she might shortsightedly postpone her return to work, and eventually find her options closing down. But as long as she is entitled to at least unpaid maternity leave, the question is why she would be more likely to come back if the leave was paid. Is the problem that she has adjusted to the lower income, so that the thought of losing her income for a further five or ten years is not as unthinkable as it once was? Or is that having paid leave would cement her relationship with the firm give her a moral incentive to come back? On he other hand, if the problem is the firm not wanting her back, then why wouldn’t the solution be a longer unpaid maternity entitlement, or indeed a financial incentive as proposed by Joshua?

    BRG: I was hoping you might make comment on this post, since you were in the thick of the discussion on the previous occasion. We seem to agree that, other things being equal, working women probably need a bigger financial inducement to have a baby, as I mentioned in points 3 and (implicitly) 6. One can see why non-working women might find this a bit unfair, though. In any case, as you note, this would be a good argument for a bigger cash payout to working mothers, which they could use as they see fit. So I think that proponents of PML really need to hammer Point 7 rather than go on about how insulting the status quo is.

    Alison: It’s never too late to comment. I agree with the thrust of your argument about rights, and its clear that non-discrimination is the starting point of Joshua’s piece as well. Also. I know I should have talked about parental leave more broadly rather than focusing on mothers, but it was complicated enough already. And thanks for raising the health issue — that does bolster the social insurance argument for government funded PML, although the whole thing is tricky, as I argued in my response to Peter. Finally, your point about different cultures having different conventions and arrangements for achieving a given goal is also a useful antidote for those of us who tend in terms of narrow economic incentives.

  15. Backroom Girl says:

    James et al

    Of course, it is relatively seldom that any particular policy has only one objective (that’s one of the things that makes policy formation difficult). So I’ll just try to have a go at setting out what I think are the reasons for paid maternity leave.

    I agree with Alison that, in the first instance, paid maternity leave recognises that a woman is incapacitated for work for a period (variable) around the birth of a baby. We would not want mothers either working too close before the birth or returning to work too early afterwards simply because of financial necessity, so a period of paid leave would obviate that. This is presumably what the ILO standard of around 3 months maternity leave is aimed at.

    Next, there is the issue of the public benefits of enabling mothers or fathers to continue to provide parental care for a period of time (James’ objective 7). If a small baby goes into formal child care, the government will subsidise that care fairly heavily – so it may be possible to devise a parental leave scheme that doesn’t cost the taxpayer all that much more in net terms. But in any case, the majority of people believe, and there seems to be some research evidence, that babies are best off with parental care for the first year or so.

    I think those are the two main reasons for having paid maternity/parental leave. Others might be to:

    * enable mothers to start their families earlier than they would if they had to meet all, or a larger share, of the costs themselves – there may well be public health benefits in this and, to the extent that some women end up having more children in total it might have a marginal positive effect on the birth rate; and

    * enable those partnered mothers for whom child bearing is currently financially the most difficult (ie those in lower paid and/or less secure jobs) to have their ‘fair share’ of the babies being born. As I pointed out in my earlier comment, this is the group that currently gets little or nothing in the way of income replacement.

    The last of those proposed objectives is, I’ll admit, a values-based objective.
    I think it is still a commonly held value in the community as a whole that people should be able to have (at least one or two) children if they want to. A problem I’ve always had with the ‘parents should only have children if they can afford them’ argument is that it seems to me that the people for whom cost is most likely to be a barrier are those in the lower income segment of the working population.

    Well-to-do parents can have as many children as they like without suffering too large a drop in their standard of living. At the other end of the spectrum, income support recipients can also have as many children as they like because the money they get increases with each additional child more or less in proportion to their costs.

    If you are a couple in low-paid work, you are likely to get family payments to help meet the direct costs of your children, but financially negotiating that first year or so of the child’s life is likely to be very difficult.

  16. Alison says:

    Yes the whole thing is tricky but I believe it is important to separate out the issues so that those which should be a constant, ie. protection of rights and obligations to safeguard women’s health, are fixed even as other economic, demographic or value-laden objectives change.

  17. Pingback: Paid Maternity Leave | Actuarial Eye

Leave a Reply

Your email address will not be published. Required fields are marked *

Notify me of followup comments via e-mail. You can also subscribe without commenting.