David R. Henderson  

Paid Parental Leave: No Free Lunch

I now consider myself to be a ... "Fund" and games with fiscal p...

When the government mandates one component of compensation, other components adjust.

Various friends on Facebook have expressed the view that government-mandated paid parental leave, which seems to have been advocated by Donald Trump supporter Ivanka Trump and is advocated by Hillary Clinton, would help women with young children. These friends' implicit assumption seems to be that other components of the pay package would not adjust. But that is unrealistic. Wages will adjust downward for the relevant group. free lunch.jpg

Consider Hillary Clinton's proposal, which, not surprisingly, is more spelled out than Ivanka Trump's. Here's what her web site says:

As president, Hillary will:

Guarantee up to 12 weeks of paid family and medical leave to care for a new child or a seriously ill family member, and up to 12 weeks of medical leave to recover from a serious illness or injury of their own.

Ensure hardworking Americans get at least two-thirds of their current wages, up to a ceiling, while on leave.

Impose no additional costs on businesses, including small businesses.

Fund paid leave by making the wealthy pay their fair share--not by increasing taxes on working families. Hillary will pay for her paid leave plan with tax reforms that will ensure the wealthiest Americans pay their fair share.

I'll start by assuming that she means it when she says that the pay will be financed solely with higher taxes on wealthier Americans. Then I will relax that assumption and assume that she doesn't mean it and that employers will have to pay.

Assuming that Employers Don't Pay

A business that hires people who are likely to be in age zone when they will want leave to take care of their children (disproportionately people between age 23 and 38) will reduce their demand for these people. Why? The business doesn't have to pay out of pocket for the leave but the business loses something valuable nevertheless. The reason they are not offering unpaid leave now is that their business is disrupted when people take unpaid leave. Ask yourself this: If you went to your employer tomorrow and said that 3 months hence, you want to start taking 12 weeks off without pay, would your employer always say yes? If the answer is no, that suggests that the employer bears a cost from losing your services.

So the demand for such people falls. What about supply? To the extent that people value this option, the supply curve of these people shifts to the right. Net effect: wages fall. In other words, part of the cost of this mandated benefit is borne by the workers who are most likely to take advantage of the benefit.

Assuming that Employers Do Pay

If employers pay, then the result above obtains a fortiori. The demand curve for the relevant population shifts down by more and the supply curve shifts to the right by more. So wages fall even more.

The bottom line: Paid parental leave is not a free lunch.

Of course, not everyone in the 23 to 38 age group plans to have kids and people who don't have kids would not be able to take advantage of this to the extent that parents would. Ideally, from the business's viewpoint at least, the employer would be able to identify (1) those who plan to have kids and (2) within (1), those who would take time off, and pay them (2) less. In practice, that would be difficult, and so one result of this policy would be that workers who don't gain from it would implicitly pay, in lower wages, for those who do.

The reasoning above makes sense as straight microeconomic reasoning. Is there any evidence for it? Actually, yes. And it comes from MIT health economist Jonathan Gruber. Yes, that Jonathan Gruber. In fact, I leaned heavily on his research when he and I both testified before Senators Kennedy and Kassebaum on the Clinton health plan in July 1994. In "The Incidence of Mandated Maternity Benefits," American Economic Review, Vol. 84, No. 3 (June 1994), pp. 622-641, Gruber studies the effect of "state and federal laws that mandated comprehensive coverage for childbirth in health insurance policies." His thinking was the same as what I laid out above. To the extent that employers could pay lower wages to women of child-bearing age in response to the law, they would. To study the effect on wages, he took advantage of the fact that some states had such mandates and some did not. He found that wages of women of child-bearing age in states that had no mandate rose more, adjusted for other factors, than wages of child-bearing women in states that imposed a mandate. As a further check, he used the 1978 federal Pregnancy Discrimination Act, which extended the mandate to all states. This gave him a natural experiment. Gruber found that after 1978, wages of women of child-bearing age in states that previously had no mandate rose more slowly than wages of women of child-bearing age in states that already had a mandate. In other words, employers, once they became subject to the mandate, were able to single out women of child-bearing age and pay them less than otherwise. He concluded:

The findings consistently suggest shifting of the costs of the mandates on the order of 100 percent, with little effect on net labor input.

Comments and Sharing

CATEGORIES: Labor Market

COMMENTS (18 to date)
Daniel Kuehn writes:

As far as I know it's pretty rare for social insurance to be paid for anything other than a payroll tax (at least all the big ones seem to be). That makes sense of course. So I think it's wise to relax the assumption of her claim that it'll be paid for by taxes on the rich. This is, of course, the election of soaking the rich and the elite. But that doesn't mean it's what will ultimately come out of Congress.

Emily writes:

If the issue you're trying to fix is one of liquidity, this isn't necessarily a problem. Instead of having a small group experiencing a large drop in income, you're spreading it out across a longer span of their working lives. You're also lowering the cost of having kids, since the wage hit is now associated more generally with being female/of child-bearing age, rather than actually having kids.

I understand that these aren't generally the arguments made for paid parental leave. And if those were the problems I diagnosed, this wouldn't necessarily be the best policy. It might be the best politically-feasible one, though.

Of course, it's detrimental for women who don't want to have kids or have lower-than-average preferences when it comes to taking leave.

Don Boudreaux writes:

David: Excellent post, as usual.

The prevailing belief among many non-economists (and, sadly, increasingly among economists) is the deeply unscientific one that certain wages and prices - or, more generally, certain contract terms - are arbitrary. This belief is reflected, for example, in the fact that many people assert (or imply) that if government forces wages up, at least just a bit, from their prevailing market levels that the only consequences are distributional: sellers (workers) get more of the gains from trade while buyers (employers) get fewer. But (it is assumed) nothing happens to resource use. Minimum wages cause no reduction in the quantities of labor hours purchased (or a worsening of job conditions); ditto for mandated overtime pay; ditto for mandate leave, be it paid or unpaid.

A similar unscientific belief that prices are arbitrary lies at the foundation of the argument of those who insist that prohibitions on so-called "price gouging" merely prevent sellers from making unjust profits and prevent buyers from paying unjustly high prices.

An interesting and important question is: where does this unscientific belief in the arbitrariness of wages and prices come from? Such a belief is understandable and forgivable in non-economists; they're not trained in price theory. But our profession today has a number of men and women in it - several of whom, such as Krugman, are prominent - who write and speak about wages and (some) prices as if these are arbitrary - as if these wages and prices reflect no particular underlying reality. It's a wholly untheoretical belief. Or, better, this "theory" of wages and prices is that many wages and price are set more or less randomly.

Note that it won't do to say that the economists whom I refer to above theorize that wages and prices are determined mostly by the greater bargaining power that employers and firms are (usually without much thought) assumed to have over workers and consumers. Even unusual bargaining power (for example, that which comes with a monopsony position in the labor market) will be fully exploited in the market by those who possess it. This full exploitation of bargaining power means that prevailing wages and prices are not arbitrary - forcibly raise such a wage and employers with monopsony power will adjust employment contracts on other margins in ways that will make most workers worse off. Forcibly lower the prices charged by a genuine monopolist and other of the terms of sale (e.g., product quality) will worsen.

Yesterday you had a lovely short post on a Harry Markowitz comment about Armen Alchian. One lesson that cannot be missed by anyone who reads Alchian's works with any care is that competitive forces are robust. These forces are so robust, and the number of real-world margins on which they play out are so numerous and (often) subtle, that there is no excuse for any economist ever to treat wages and prices prevailing on markets as arbitrary. Such a treatment is as unscientific as would be, say, an astronomer's insistence that there's nothing special about the particular orbit of the earth around the sun. Everyone would reject the counsel of an astronomer who proposed to use a giant rocket to force the earth into an orbit at twice the distance from the sun as its current orbit. A similar rejection should greet the counsel of any economist who proposes to force wages, prices, and other contract terms from their current positions.

Jon Murphy writes:

Interesting post. I wonder what the effect would be in someplace like Massachusetts which just passed a bill requiring equal pay for comparable work (see here).

I don't know the details of the bill (so there may be something to prevent this), but I would strongly suspect we would see considerably less hiring of young married women if this bill were combined with Hillary's proposal.

Hazel Meade writes:

All though there is no officially mandated maternity leave at the federal level, in practice most companies do offer paid leave for women who have babies - in the form of short term disability, which is often privately insured by the company.

I wonder if anyone knowledgeable would care to comment on what effective difference federally mandated leave would have versus STD as it's currently done. Maybe it would be simpler, but would the net effect on salaries and leave time really be any different? Maybe tyhis proposal is really just window dressing considering most women actually do get paid leave in practice - they just don't call it maternity leave.

Dan writes:

Lower wages do not necessarily mean that the policy is bad. There are externalities, co-ordination failures across workers, etc.

Making workplace sexual harassment illegal also reduces wages, but it's not necessarily a bad policy.

Kevin Dick writes:

One might think from the Democratic platform that they believe people aged 40-50 are the optimal workers.

The minimum wage makes younger workers more costly. Parental leave makes workers of prime childbearing age more costly. Mandated health insurance make older workers more costly.

These are _already_ the peak earning years. So, ironically, the Democratic policies magnify employment inequality. (Though there's probably a policy I've forgotten whose cost incidence falls mainly on 40-50 year olds.)

Tom West writes:

Of course there is never a *free* lunch, but you can often get a steeply discounted one.

Don Boudreaux seems somewhat vexed that many economists are aware enough of the world to understand that non-economic factors (such as custom, prestige, ethics, personal preferences, etc.) often dwarf economic factors in ostensibly economic decisions.

And while I can sympathize with the "economics is right, it's people who are broken", given that we're stuck with human beings, government policy might as well capitalize on their brokenness to put forward policies it thinks are worthwhile.

Sure, on the margins a few people will "get it", think economically and claw back our free lunch, but a *huge* number won't, giving us a very nice discount.

zeke5123 writes:

It's odd; the justification for government intervention is often externalities. Yet here, where the government is creating externalities, defenders of the policy seemingly trumpet its virtues precisely because of the externalities (because it spreads the costs amongst many workers).

Just an observation.

Tim Worstall writes:

"Ideally, from the business's viewpoint at least, the employer would be able to identify (1) those who plan to have kids"

As far as I'm aware (meaning of course I could be wrong) it's illegal to ask a female employee if she is intending to have kids.

Toby writes:


while it is definitely not a free lunch, do you think that some individuals would be willing to pay the price for paid maternity leave if they could bargain for it? If they would, then it seems like a sensible policy don't you think?

I am happy that there are some mandatory defaults that I don't need to bargain over with my employer. That saves me and my employer a lot of time haggling.

Sure, I don't get exactly what I want, but it seems worth the cost because otherwise I'd have to figure it out all by myself what I might need or might not need instead of being able to rely on what other individuals thought useful.

Of course, in a free market there are forces operating that employers will want to offer me these options because employers are competing for workers. I am not sure though how strong these forces are as the feedback is much slower than say the feedback an entrepreneur receives when she positions her ice cream truck on the beach. The optimal distance is much easier found there.

Now the government and/or labor unions don't have this information either. However, voters are more willing to express their preferences directly to them than to their employers. Personally, I prefer that labor unions do the negotiation but a government mandate isn't too bad either. It all depends on how the available knowledge about preferences can be used at the lowest possible cost.

S D writes:

In most European countries there is paid paternal leave for mothers AND a considerable level of compensation for the employer as well.

This smooths income for mothers and makes employers less worried about taking on a woman of childbearing age.

There is probably some productivity (economy- and individual-level) to this kind of system vis-à-vis the US.

However it is all paid for out of social insurance contributions (employer and employee) that are generally much higher than what prevails in the US.

There is a clear trade-off of course here. Higher taxes on labour weigh on employment.

The Clinton proposal is of course a free lunch which does not exist. A more honest proposal would be to be up-front and say that workers can have paid paternal leave but they and their employers will have to pay for it via increased payroll taxes, and let voters see if they like this.

Zeke5123 writes:


I think you are correct about seeing government and free markets as allocating resources (in this case allocating resources away from wages to benefits). But you seem to think in this case government can do it better. It seems odd to me from a knowledge problem / revealed preference perspective.

What I mean by this is that republican democracy is a really bad way to utilize local knowledge or to identify revealed preference. You are given a binary choice once every two years that is indeterminate as to the outcome. Further, you are voting for a candidate; not a specific policy explanation. This method doesn't encourage individuals to utilize their local knowledge because there is little incentive to do so. Further, it doesn't encourage revealed preference (as an individual's preference doesn't matter much in an election for the reasons discussed). Instead, it encourages social desirability bias, as there is no individualized harm to voting with the crowd or for high minded ideas like maternity leave.

Contrast this with market behavior where choices aren't binary yes/no but occur on a spectrum, your vote is determinative for yourself, you can modulate your social vote (i.e. It isn't one person one vote -- you can buy a lot of X, some of X, or none of X), and the harm of social desirability bias falls on the voter in this case.

Labor unions as currently constructed might be better than government but face similar issues.

Jay writes:

@ Hazel

Yes, most mid to large sized companies already offer leave but those that push for these mandates assume that if it isn't required than it isn't happening and that's how its argued.

To those saying that just because there's a trade off doesn't make it bad policy, that isn't the point, it is rarely if ever argued honestly as a trade off and that's David and Don's point here. You can't even begin to debate a policy when one side says it'll be paid for from the rich or corporate profits. They act like the generous European leave and vacation requirements just rain down from heaven and we're dumb or vicious not to offer them here.

BC writes:

"one result of this policy would be that workers who don't gain from it would implicitly pay, in lower wages, for those who do."

This reminds me of a recent marginalrevolution post about how prohibiting employers from asking about job candidates' criminal records actually hurts minorities without criminal records [http://marginalrevolution.com/marginalrevolution/2016/06/ban-the-box-or-require-the-box.html]. If women are expected to be more likely to take parental leave than men, then women in general will be paid lower wages than equivalent men and those women that never end up taking parental leave, for example single and childless married women, will be hurt most by this policy. This effect might be especially pronounced because it is illegal for employers to ask about job candidates' marital and family status so in negotiating job offers there is no way for childless women to let potential employers know that they should earn higher wages to offset their expected non-use of parental leave. It seems like mandatory leave policies would most benefit men that unexpectedly take parental leave more often than their male peers since their wages will be set based on a lower expectation of leave taking.

Ironically, many of the people that claim to want to close the "wage gap" between women and men also support mandatory maternity leave.

Cyril Morong writes:
Toby writes:


It entirely depends on whether individuals feel freer to express their preferences in the voting booth, as a member of a labor union, or as an employer. I think that it makes more sense to express these preferences in a voting booth or as a member of a labor union, rather than in salary negotiations. The same goes for working hours and other such aspects. Think about what you might be signaling to your employer if you ask for a maternity / paternity leave or for fewer hours. It's not really going to help your career.

Tom West writes:

it is rarely if ever argued honestly as a trade off and that's David and Don's point here. You can't even begin to debate a policy when one side says it'll be paid for from the rich or corporate profits.

Well, let's be real. Once you are in the political realm, you can't acknowledge your chosen position has any cost under any circumstance. Of course it's absurd, but that's just reality.

It's only when you're small enough scale not to matter that somewhat nuanced debate can take place, but even here, how often do you have the posters point out the cost of freedom? That's not their job.

In other words, it's the job of partisans to point out the costs in the opposition's arguments, but you can't really expect anyone holding a position to point out the weaknesses or trade-offs in their own. That's simply sabotaging your own side.

Comments for this entry have been closed
Return to top