Evidence
Why parental leave matters: what the research shows
There is a large body of research on what happens when parents take longer leave after a birth. It covers parent health, infant outcomes, fathers, and what happens at work. We have collected the findings we can trace to a primary source, along with the places where the evidence is weaker than people assume.
We are not arguing for any policy here. We build a planning tool, and our interest is narrow: parents keep asking whether the extra weeks are worth the trouble of claiming them. This page is what the studies actually found.
Last updated August 12, 2026. For claim rates and program data rather than outcomes, see our paid leave statistics page.
Duration is the variable that keeps showing up
If there is one finding that organizes the rest of this page, it is that length matters more than access alone. Benefits tend to appear at two to three months, not at two weeks.
2 to 3 months
Where the mental health benefit becomes consistent
A systematic review of 45 studies across high-income countries found improved mental health among mothers under more generous leave policies, and reported that paid leave of at least two to three months was the most protective. Shorter paid leaves produced weaker and less consistent results. The review examined leave policies by three measures: whether any leave existed, the benefit amount, and the duration.
Source: Heshmati, Honkaniemi and Juárez, The Lancet Public Health · 2023
3 to 6 or 7 weeks
What California's program did to how long mothers actually stayed home
California introduced paid family leave in 2004. Using March Current Population Survey data from 1999 to 2010 and a difference-in-differences design, researchers found the program more than doubled the overall use of maternity leave, from around three weeks to six or seven for the typical new mother, with particularly large growth for less advantaged groups. That is a real increase, and it still lands short of the two to three month range where the health literature finds the clearest effects.
Source: Rossin-Slater, Ruhm and Waldfogel, NBER Working Paper 17715 · 2011
Why this connects to coordination
In most states, no single benefit gets you to two or three months on its own. Reaching that range usually means claiming a medical recovery benefit and then a separate bonding benefit, and often stacking employer policy on top. That second claim is the one parents most often miss. In Massachusetts, nearly a third of parents who take paid medical leave never file the family leave claim that follows it.
So the practical question is rarely whether longer leave is better. It is whether anyone told you there was more to claim. The claim-rate data is here.
The birthing parent
Mental health and breastfeeding are the two outcomes with the most consistent US evidence.
11.7% vs 13.3%
Postpartum depression symptoms, most generous states against least
Comparing US states by the generosity of their paid family and medical leave, researchers found postpartum depression symptoms in 11.7% of respondents in the most generous states and 13.3% in the least. Having only unpaid leave was associated with higher odds of depressive symptoms than having at least some paid leave.
Source: Association Between State Paid Family and Medical Leave and Breastfeeding, Depression, and Postpartum Visits · 2023
18 days
Additional breastfeeding duration after California's policy
Using a sample of more than 270,000 children born between 2000 and 2012, drawn from restricted-use National Immunization Survey data, researchers estimated that California's paid family leave policy increased overall breastfeeding duration by nearly 18 days and raised the likelihood of breastfeeding for at least six months by 5 percentage points. Effects were substantially larger for some disadvantaged mothers.
Source: Pac, Bartel, Ruhm and Waldfogel, NBER Working Paper 25784 · 2019
59.5% vs 51.0%
Breastfeeding at six months, most generous states against least
The same state-comparison study found breastfeeding rates at six months, or at time of survey, of 59.5% in states with the most generous paid family and medical leave against 51.0% in states with the least.
Source: Association Between State Paid Family and Medical Leave and Breastfeeding, Depression, and Postpartum Visits · 2023
Infants
This is the section where being straight about the evidence matters most, because the headline finding is real and narrower than it is usually reported.
12%
Reduction in postneonatal mortality after California's policy
Researchers compared birth and death certificates from all in-hospital deliveries in California against Missouri and Pennsylvania from 1999 to 2008, a sample of 6,164,203 births, using a difference-in-differences design around California's 2004 policy. They found roughly a 12% reduction in postneonatal mortality, meaning deaths occurring after the first 28 days of life, after adjusting for maternal and neonatal factors.
Source: Montoya-Williams et al., Health Services Research · 2020
What the same study did not find
No significant effect on preterm birth, on low birthweight, or on overall infant mortality. You will sometimes see this research summarized as though paid leave broadly improves birth outcomes. One outcome moved. Several did not. We would rather you get that from us than find it yourself in the paper.
Fathers and non-birthing parents
The research here is smaller than the maternal literature but points consistently in one direction, and it is the most underused evidence in this whole area.
2 weeks
The threshold where father involvement changes
Drawing on evidence from four OECD countries, this study found that fathers who take leave, especially those taking two weeks or more, are more likely to carry out childcare related activities when children are young. The association persists after the leave period itself ends.
Source: OECD, Fathers' Leave, Fathers' Involvement and Child Development: Are They Related? · 2013
Some evidence
Cognitive scores yes, behavior not really
The same study found some evidence that children with highly involved fathers perform better on cognitive test scores. It was explicit that evidence on the link between fathers' involvement and behavioral outcomes was weak. The authors also concluded that the kind of involvement matters more than the amount, and that quality rather than quantity of father-child interaction is what shows up in the results.
Source: OECD, Fathers' Leave, Fathers' Involvement and Child Development: Are They Related? · 2013
Wage replacement
Fathers' mental health improved where the money was adequate
The Lancet systematic review found some evidence that fathers' mental health improved under policies providing adequate wage replacement or other incentives to take leave. It did not find strong evidence for indirect effects of one parent's leave on the other parent's mental health. Replacement rate is the operative variable, which is consistent with low pay being a common reason fathers skip leave they are entitled to.
Source: Heshmati, Honkaniemi and Juárez, The Lancet Public Health · 2023
Work, retention and employers
Useful if you are weighing whether taking the full leave will cost you at work, and useful if you are the one writing the policy.
Easier, not harder
What New York employers reported after their policy took effect
Researchers surveyed New York and Pennsylvania firms, matched each New York firm to a comparable Pennsylvania one, and used difference-in-differences models to compare them. New York's 2018 paid family leave policy led to an improvement in employers' rating of how easily they handled long employee absences. The effect was concentrated in the first policy year and among firms with 50 to 99 employees. Employee leave-taking also rose in the second year, driven by smaller firms.
Source: Bartel, Rossin-Slater, Ruhm, Slopen and Waldfogel, NBER Working Paper 28672 · 2021
6% to 9%
Suggestive evidence of higher work hours later on
The same California analysis found suggestive evidence that paid family leave increased the usual weekly work hours of employed mothers with one to three year old children by 6 to 9%, and that their wage incomes may have risen by a similar amount. The authors label this as suggestive rather than robust, and we are repeating that label deliberately. The leave-taking result was the strong finding. This one is weaker.
Source: Rossin-Slater, Ruhm and Waldfogel, NBER Working Paper 17715 · 2011
What this research does not establish
Every page like this should have a section like this. Here is where we think the evidence is thinner than the way it usually gets quoted.
Most US evidence is California
California passed the first state program in 2004, so it is the only place with enough elapsed time for strong causal work. Much of what gets described as evidence about American paid leave is evidence about one state with a particular labor market, a particular replacement rate, and a particular population. Whether it transfers to Maine or Colorado is an open question.
Policy effects are not individual effects
Difference-in-differences studies measure what changed in a state after a law took effect. That is not the same as measuring what happens to one family that takes six more weeks. Both are useful. They answer different questions, and the second is much harder to study because parents who take longer leave differ from parents who do not in ways that are hard to control for.
Several outcomes show no effect
Preterm birth, low birthweight, and overall infant mortality did not move significantly in the largest US infant-health study. Null results get quoted far less often than positive ones, which distorts the picture if you only read summaries.
Nearly all of it is high-income countries
The Lancet review restricted itself to high-income countries, and the same is true of most of the literature. Findings should not be read as universal.
We are not neutral about our own product, and we are not going to pretend otherwise
We sell a leave planning tool. Research showing longer leave is valuable is research that makes our product more appealing, and you should read this page knowing that. It is also why we link every figure and why we included the null results. Check the sources.
Frequently asked questions
How much parental leave do you need for the health benefits to show up?
A 2023 systematic review in The Lancet Public Health found improved mental health among mothers with more generous leave policies, and reported the effect was clearest with paid leave of at least two to three months. Shorter paid leaves showed weaker and less consistent effects. This is one of the reasons coordination matters: in most states, reaching two to three months means claiming both a medical recovery benefit and a separate bonding benefit.
Does paid family leave improve maternal mental health?
The evidence points that way. A study of US states found postpartum depression symptoms ranged from 11.7% in states with the most generous paid family and medical leave to 13.3% in states with the least. The Lancet systematic review reached a similar conclusion across 45 studies. These are population-level associations rather than guarantees for any individual.
Does paid leave affect infant health?
In the strongest US study on this question, California's 2004 paid family leave policy was associated with roughly a 12% reduction in postneonatal mortality, meaning deaths after the first 28 days. The same study found no significant effect on preterm birth, low birthweight, or overall infant mortality. The honest summary is that one specific outcome improved and several others did not move.
Does paternity leave actually change anything?
An OECD study across four countries found that fathers who take leave, particularly two weeks or more, are more likely to carry out childcare activities when children are young, and found some evidence that children with highly involved fathers score better on cognitive tests. The same study was explicit that evidence on behavioral outcomes was weak. The Lancet review separately found fathers' mental health improved where policies offered adequate wage replacement, which matters because low replacement rates are a common reason fathers skip leave.
What does the research not show?
Several things. Most US evidence comes from a small number of states with long-running programs, especially California, so it may not generalize everywhere. Many studies measure the effect of a policy existing rather than the effect of one family's choice. Several outcomes, including low birthweight and preterm birth, show no significant effect. And nearly all of this research is from high-income countries.
Sources, by type
Grouped so you can weigh each figure. Peer-reviewed and intergovernmental carry the most weight here. We have deliberately not cited advocacy organizations on this page, because outcome claims are where their framing shows up most.
Peer-reviewed journals
NBER working papers are circulated for discussion and have not been peer reviewed. We label them as working papers throughout for that reason. Several of the ones cited here were later published in peer-reviewed journals.
Citing this page
Journalists, researchers, clinicians and HR teams are welcome to use this page. Every finding links to its original source, and we would ask that you cite that source directly rather than us.
Suggested citation
"Why Parental Leave Matters: What the Research Shows." Parental Leave Pro, August 12, 2026. https://www.parentalleavepro.com/why-parental-leave-matters
If you are writing about paid leave and need a figure for a specific state, wage level, or household situation, we maintain a rules engine covering 18 state programs plus federal FMLA and can usually turn a scenario around quickly. Get in touch.
How we chose what to include
We include findings we can trace to a paper a reader can open. We prefer peer-reviewed journals and large administrative samples, we name the sample size and the design where it affects how much weight a finding carries, and we report null results alongside positive ones. We do not cite advocacy summaries of research on this page, and we do not present our own estimates as though they were research. Our rules-engine methodology is documented separately on the methodology page.
This page describes population research. It is not medical advice, and it cannot tell you what is right for your family or your recovery. Talk to your clinician about that.
Find out how many weeks you can actually get
The research points to two or three months. Whether you can reach that depends on your state, your employer, and whether you claim every program you qualify for. The free check takes about 60 seconds. No email required.
Check my leave. Free →