Why Does Purim Occur on the Jewish Month of Adar?: An Insight on Purim and Happiness

A little over a week ago, the Jewish calendar reached the month of Adar. The Talmud (Ta'anit 29a) says that when the Jewish month of Adar begins, we increase our happiness. The Talmud continues to say that Jews should proceed with litigation in the month of Adar because of the good luck that is inherent within the month of Adar. During the month of Adar is the holiday of Purim. Purim commemorates the biblical account of the Book of Esther. In this Book, we see the story of how the Jewish people, who were at the time in exile in Persia, were saved by the cruelty of Haman, the king's vizier. What does the miracle of being saved from extinction have to do with happiness?

1. The Hebrew month Adar (אדר) has the same root as the Hebrew word for "strength" (אדיר). This correlates with the fact that the zodiacal sign of Pisces corresponds with the month of Adar. Pisces is the sign of the fishes. What do fishes have to do with Purim?
• Fish are seen as a fertile species, and thus are a symbol of blessing and fruitfulness. 
• I wrote a homily about four years ago on why fish have fins and scales. Fish are a metaphor for how we should behave. Fins help us propel forward, and scales protect us when life is bumpy. 
2. In the Exodus narrative, Adar was the last month that the Jews spent in Egypt. The month of Adar commemorates as the prelude to the Exodus, which leads to the receiving of Torah and commemorates freedom. Happiness is rooted in our freedom because it gives us the ability to reach our fullest potential. 
3. Jewish tradition teaches us that when the Jews accepted the Torah at Mount Sinai, it was done so under duress. The Talmud teaches (Shabbat 88a) that G-d held Mount Sinai over the heads of the Jewish people to accept Torah. However, it was during Purim that the Jewish people accepted the Torah voluntarily. The voluntary nature of accepting Torah and mitzvahs teaches us that we accept G-d out of love and not fear. Pirke Avot teaches that while fear is sometimes necessary, it is out of the love of G-d that we truly become close to G-d. Purim is such a happy moment because we come to G-d out of love, and not fear. 
4. Haman selected the month of Adar in attempts to annihilate the Jewish people not just because the Jewish people lost their Temple and were already down. Haman chose Adar because it was the month that Moses passed away. Haman took that as a sign to destroy the Jewish people. While that 7th of Adar was when Moses passed away, it was also the day when Moses was born. This reminds us that a fork in the road could be a crisis or an opportunity. Looking at research on happiness, about 40 percent of happiness is our own choice. We have the ability to take that 40 percent and choose what to do with it, much like the Jewish people decided what to do in terms of making the best of their situation.  
5. Some of the Purim practices liberate us from our usual routine and has the potential to encourage happiness. Mishloach Manot, the practice of sending gift baskets to at least two friends on Purim, encourages us to give. Giving is significant because research shows that giving leads to greater health and happiness. The practice of drinking on Purim tries to find a balance between not being sober and being drunk in order to connect to G-d in a different way. Wearing costumes is a way to reveal a different part of ourselves that we might not reveal other days of the year. Exploring different venues can help us find a different way of being happy that we might not have previously thought possible. 
6. Purim is a time where we turn things upside-down. Things that are normally not encouraged (e.g., drinking a fair amount of alcohol) or even forbidden (e.g., cross-dressing) are permitted on Purim. It turns life upside-down for a day. R. Geoff Mitelman brings up the idea that is the fact that life gets turned upside-down is the reason why we should be happy on Purim. The word Purim (פורים) comes from the Hebrew word "lot" (פור). During the entire Purim story, the fate of the Jewish people was uncertain. Purim is a reminder that we don't have full control over our lives, and that chasing after a security blanket or a bubble in which nothing hurts us ever is not only impossible, but precisely the wrong way to go about happiness. Sukkot teaches us a similar message: The sooner we get comfortable in the insecurity of life, the happier we will become because we are not chasing after something so elusive. 

To tie all of these ideas together, Purim is about how we choose to react to our circumstances. Life is not always fun. There are times where it is downright cruel and uncertain. The story of Purim is taking a moment of darkness, a moment where it looked like the Jewish people would be no more, and to transform it into light. And that's the point: the ability to make life a happier one comes from us. It is an internal process, and chasing after external validation or success leaves one empty. Am I saying that it's easy? Of course not! I'm working on it myself, and it comes with difficulty. However, the month of Adar reminds us that we take control over what we can and adapt to the rest. I'll end with the Serenity Prayer:

G-d, grant me the serenity to accept the things I cannot change, the courage to change the things I can, and the wisdom to know the difference. 

Prevalence of School Shootings and Why We Don't Need to Arm Our Teachers

It is amazing how mass shootings cause a media frenzy and a huge amount of debate in the United States. I know that for me at least, it has resulted in a few blog entries. With the Las Vegas shooting last year, it was about how there isn't a real link between mental health and mass shootings. In response to the Orlando shooting, it was how a high-capacity ban wouldn't make a real difference. The Sandy Hook shooting had me wondering about whether "common sense gun reform" was possible, and the Aurora gun shooting resulted in a brief attempt of me putting mass shootings in the context of overall deaths in the United States.

With the Parkland shooting last week that occurred at a public school and involved 17 innocent children taken from this world much too soon, the topic of gun control has reemerged. Some have proposed "fighting fire with fire" by arming teachers with guns. This was part of the NRA's National School Shield Plan from 2013, so it is not as if this were a brand new idea. The state of Kentucky recently introduced a bill to arm teachers and administrations. The premise behind this idea is to provide training, practice, and advice directly related to a school shooting scenario in order to minimize the carnage unleashed during a school shooting. It is seen as an alternative to schools that cannot not afford to have police officers or other security guards on the school premises. I want to see whether or not this would be a good use of taxpayer dollars, but first, I want to see how prevalent school shootings are.

I bring up prevalence because in order to assess a problem and what are justifiable costs and benefits, we need to know how frequent it is happening. After the Parkland shooting, school shootings in the United States are being branded as "a new normal." The anti-gun organization Everytown for Gun Safety put out a statistic that the United States has already experienced 18 school shootings this year. Considering that we're not even two months into 2018, that sounds like a lot. The problem is that the statistic was inaccurate as it was sensationalist, and the Washington Post and Politifact called them out on it. So how prevalent are they?

Prevalence of Gun Homicides and Mass Shootings
It makes sense to ask how prevalent mass shootings in general are since some mass shootings are also school shootings. Plus, schools are the second most common site for mass shootings (FBI). Since the Parkland shooting was both a mass shooting and a school shooting, I will be looking at the prevalence of both.

It is a point I first brought up in 2012 and again in 2016: mass shootings are an uncommon form of gun-related death. The CDC's National Vital Statistics System found in their most recent report on fatalities (2015) that gun homicides do not make the Top Ten list for leading causes of death. In 2016, gun-related homicides increased from 12,979 in 2015 to 14,415 in 2016 (see CDC WISQARS database). Homicide rates did increase in 2016, but a preliminary review of 2017 data shows a decrease in homicide rates. Since most homicides in the United States are committed with guns, it is reasonable to assume that gun-related deaths also dropped in 2017. This brings us to mass shootings.

In its piece on mass shootings (updated after the Parkland shooting), Washington Post breaks down the historical demographics of mass shootings. Washington Post admits that mass shootings are a small portion of overall gun deaths. In 2018 so far, there have been 21 deaths from mass shootings and 1,827 gun deaths. For 2018, that would mean 1 percent of gun deaths are from mass shootings. However, that is just for two months, which is hardly an adequate timeframe to define a "new norm." The Congressional Research Service (CRS) released a report on mass shootings from 1999 to 2013. Using the FBI's definition of mass shooting as "multiple-homicide incident with four people killed within a single event," CRS found that there have been an average of 21 mass shootings annually, and that there has not been a particular increase in this time period (see below). The Left-leaning Mother Jones uses this criterion (along with some others) in its open source database.

Comparing across nations is also interesting. As CNN brings up in a recent article, from 1966 to 2012, the United States accounts for 5 percent of the population and 31 percent of the mass shootings.   The Crime Prevention Research Center does something interesting with the international data. They compare the United States to Europe by using death rates per million and frequency per million, and found that the United States is not at the top of the list.

Mass shootings are infrequent, but are terrifying not just because of the number killed, but also because they take place without advanced warning and in places we would not expect. One of those unexpected places is in a school, which brings me to my next point......

School Shooting Prevalence
Now that we have looked at gun homicide statistics and mass shooting statistics, let's take a brief look at school shooting data. The data I found did not segment by number of deaths in the school shootings the same the data are segmented with mass shootings, so I won't make that distinction. As for prevalence, here is what I could find: I'm normally not a fan of using Wikipedia as a source. In this case, they provide a good listing of school shootings in the United States. We see an increase of school shooting deaths in the 1980s, and especially in the 1990s and beyond. 150,000 students have experienced a school shooting since Columbine.

This sheds some light on the raw data, but how about the rate at which school shootings happen? At the very least, we need to adjust for population growth over time. The Department of Education's National Center for Education Statistics [NCES] provides an annual report called Indicators of School Crime and Safety. In its most recent report, we see a breakdown of deaths in schools (see below). What NCES concludes is that from 1992 to 2014, school shootings have accounted for less than 3 percent of overall youth homicides (NCES, 2016, p. 37). Just as interesting, the Cato Institute looked through NCES data to calculate what percentage of schools undergo a school shooting. As of 2014, only 0.009 percent of schools have historically undergone a school shooting. In terms of percent of children killed, only 0.000044 percent of children are murdered in a homicide at school. In its 2016 fact sheet on Understanding School Violence, the Centers for Disease Control and Prevention (CDC) said that "school associated violent deaths are rare." How rare? According to 2012 CDC research, the CDC found that the probability of a child being a victim of a school-related homicide is 1 in 2.5 million. Even if you were to argue that school shootings have been twice as prevalent since then (which is generous to assume), that would still mean only a probability of 1 in 1.25 million, which would mean that a child would be about 10 times more likely to be struck by lightning and 17 times more likely to die in a car accident.

Globally, the best data we have is from the Academy for Critical Incident Analysis. This Academy aggregated data from each incident where there were two or more victims. From 2000 to 2010, 28 out of 57 incidents happened in the United States alone. When it comes to mass shootings that are school shootings, the United States has a bigger problem than other countries. Nevertheless, school shootings remain a small percentage of youth homicides in the United States, as well as a statistically rare occurrence. 

Should We Arm Teachers in Response to School Shootings?
I don't want to get into whether there are too many school shootings. Given the horrific nature of the crime and the fact that the victims are children and school staff, I'm sure we can all agree that one school shooting is too many. For argument's sake, let us assume that the prevalence of school shootings is high enough where we need to take some drastic action, such as arming teachers. While the idea has some intuition behind it, I have a few issues (also see Center for Homicide Research paper here):

• Cost of arming teachers. Per the NRA's 2013 report, one-time training costs $800 to $1,000. To be generous, let's assume $800. NCES estimates that there are 3.592 million teachers in 2018. 30 percent of Americans own a gun (Pew Research). Only 61 percent of firearm owners have had training (Reuters). Assuming that the demographics of U.S. firearms owners applies to teachers, 657,336 teachers have had training, which means that we would need to train 2.93 million teachers. That would mean that the cost of training these 2.93 million teachers would cost $2.3 billion nationwide. This does not consider that training would need to be ongoing, the cost of purchasing the handguns, the permits for these handguns, and the storage and insurance for the handguns. Many schools already have strained budgets. Can schools afford such a measure? 
• Storage of firearms. It is not just the cost of the storage that concerns me. Where are the teachers going to store their firearms? If the storage area is too heavily guarded, it would be too difficult to access in the event of a school shooting. If it is too easy to access, there is legitimate concern that someone could steal the firearms.   
• Success rate of hitting target and casualties. When I was covering the topic of the inadequacy behind a high-capacity magazine ban, I brought up how police officers hit their target 30 percent of the time. Even if teachers are trained, it is a reasonable assumption that their success rate would be lower than that of a police officer. Why? Because teachers are not accustomed to such scenarios like police officers and security guards are. Combine that with a bunch of students potentially in the line of fire, I cannot imagine how this would end well. It is difficult enough for a police officer to hit their target. How can we expect a teacher with minimal to no training to do the job? 
• Teacher pressure and turnover rates. On average, the United States experiences an attrition rate of about 2 percent (BLS). The Department of Education's NCES found that about 17 percent of teachers quit teaching within 5 years, which is below the commonly believed 50 percent. In spite of the lower attrition rate, teachers still face a number of priorities, including grading, academic commitments, curriculum planning, mentoring students, and meetings. With everything else teachers have to manage, you want to throw gun training and keeping track of a gun on top of it?

Postscript
It is important to make sure our children are safe while in school. At the same time, it is not the job of the teacher to defend students. It is their job to teach. That is what they are trained to do. Having a jerk-knee response to a phenomenon that has not particularly been on the rise is a poor use of resources and a way to let fear override better judgement. Arming teachers is a way to cut corners, especially given the potential lethalness of the government funding teachers to carry firearms while in school.

And let's not forget about prevalence. Government statistics show that school shootings are rare, even if they are more common than in other countries. School shootings account for a small percentage of youth homicides, and mass shootings account for a small percentage of firearm homicides. We should keep the prevalence in mind as we make policy priorities. This is not a call to discontinue the discussion on gun reform, but rather a call for more effective and targeted way to reduce gun deaths than arming teachers.

"Chain Migration": How Trump Needlessly Demonizes Family-Based Immigration

For those of you have been reading my blog, you have probably noticed that I have been covering the topic of immigration more extensively since Donald Trump became President. Since the beginning of 2017, I have covered such immigration-related topics as the border wall, refugee ban, high-skilled immigration visas, the idiocy behind cutting immigration in half, DACA, why we don't need more border agents on the wall, and temporary protected status. Today, I feel the need to cover another topic of immigration that Trump doesn't have a clue: chain migration.

Chain migration is a term that was popularized in the 1960s to describe how U.S. citizens and green card holders use current law to bring extended family into the country, much like links follow one another in a chain. It refers to a specific type of family reunification, i.e., when a family member is following another immigrant. It is not a derogatory term, but rather a term for a commonsense idea that people are more likely to move whether their relatives are. Chain migration is a mechanism that many families have historically used to bring their families over to the United States, including President Trump.

I bring the topic of chain migration up in the first place because for President Trump, an end to chain migration is an essential to sign off on any immigration deal for the 690,000 affected by DACA. Trump views chain migration both as a way for terrorists to enter the countries and as a way to open the floodgates of immigration by allowing an unlimited number of extended relatives. If we are to get into chain migration, we should figure out the demographics of these individuals and figures out what is going on. With that, let's cover the myths that are commonly used by the President and other like-minded individuals.

1. "Chain migration" immigrants are a national security threat. The Cato Institute covered this topic last year, and found that the odds of being killed by such in immigrant in a terrorist attack is 1 in 723,000,000. To put this in perspective, the odds of being struck by lightning is 1 in 161,856. You are nearly 4,500 times more likely to get struck by lightning than killed by an immigrant brought in through chain migration, and over 50,000 times more likely to get killed by a native citizen.
2. Immigrants using "chain migration" to enter the country do not have the skills to succeed in the U.S. Last month, Attorney General Jeff Sessions made the argument that these immigrants are incapable of succeeding because they lack literacy and work skills. Nearly half of adults who come to the United States through family-sponsored and diversity visa categories have a college degree. This is higher than the third of U.S. natives that have a college education. 
3. The low-skilled "chain migration" immigrants are ruining this country. 11 percent of those who come through family-based immigration lack a significant formal education. Nevertheless, these individuals, by and large, succeed. This does not surprise me in the least since low-skilled immigrants on the whole positively contribute to the economy.
4. "Chain migration" is a loophole resulting in an influx of millions of immigrants. Trump claims that "a single immigrant can bring in unlimited numbers of distant relatives." That simply is not true. If you take a look at current law, it only allows for sponsoring immediate relatives (i.e., spouses, children, siblings, parents) and their spouses and minor children. Current law does not allow for aunts, uncles, and cousins, not to mention that immigrants cannot sponsor relatives until they themselves have green cards or become naturalized citizens. 
• Although there is no cap for immediate family members, there is a maximum of 480,000 family preference visas. In 2015, 69 percent of these visas were for U.S. citizens' spouses and children, which further diminishes the argument. 
• Per a report released from the Congressional Research Service [CRS] on the topic of family-based visas earlier this month, we have a backlog of 3.95 million people, with average wait time being 23 months and waits as long as 23 years (CRS, p. 13, 19).
•  The CRS report points out that Trump's hypothetical of immigrants flooding our cities is theory without empirical backing. The report then lists reasons for family-based immigration being more modest, including needing U.S. citizenship, the fact that not all eligible candidates want to immigrate to the United States, and long wait times for visas (CRS, p. 24-25). As the Right-leaning Manhattan Institute states, "Chain migration is an information system that enables networks of people to find each other to come and work. If there is no work, individuals will not invite family members to come."
5. "Chain migration" is overwhelming the United States. You would think that from President Trump's rhetoric, we are being overrun by immigrants vis-à-vis "chain migration." In absolute numbers, we admit more immigrants than other nations in large part because we are a larger nation than most. In 2016, the United States allowed for 1.2 million legal immigrants. 68 percent of those immigrants were brought here due to "chain migration." To account for nations' sizes, we need to look at immigrants as a percent of the nation's overall population. The Cato Institute did so in mid-2016, and found that as a percent of the population, the United States admits less immigrants than other developed nations.

There are not masses of foreigners coming on our shores (CRS, p. 27-28) and eroding our economy or the "American way of life." "Chain migration" has become  the nativist boogeyman term used as cover to want to cut down on legal immigration. There is no justification on cutting back or banning "chain migration." Ultimately, it does not matter if immigration is "chain migration" or not. It does not matter if the immigrants are high-skilled or low-skilled. What we find is that more immigration is a net benefit for immigrants, native workers, and the economy as a whole.

Biennial Budgeting: An Example of Bogus Budget Reform

Last week, Congress passed a federal budget for the fiscal year 2019. You can read an analysis from the Committee for a Responsible Federal Budget here, but essentially, the claimed deficit reduction will not be adequate to deal with the upcoming increase in debt. Creating and passing a budget is a messy process, to be sure. It makes me think of possible ways to make federal budgeting more efficient.

I found one such policy reform when reading a piece from libertarian think tank R Street: biennial budgeting. Biennial budgeting is preparing and adopting budgets for two-year periods. This is in contrast to an annual budget, which prepares and adopts a budget for a one-year period. With all the threats of shutdowns, the budgetary process is held hostage by political desires instead of the desire to serve the people it is meant to serve. Proponents argue that a biennial budget process would solve this problem. Additionally, it could potentially make the budgeting process more efficient because Congress would lessen the number of repetitive votes required by an annual budgeting process, thereby allowing for more time on oversight of federal spending and less politicking. And of course, there is the decrease in man-hours and resources spent on the additional budget that is part of annual budgeting. These all sound like wonderful benefits, but does biennial budgeting outshine annual budgeting? I'm not so sure for a few reasons:

1. Inaccurate budget forecasting. Economic forecasting is a tricky endeavor. It is difficult to forecast one year out, never mind two years. The Congressional Budget Office doesn't exactly have a spotless record on budget projections beyond one year (see testimony from Urban Institute scholar here). There has even been congressional testimony by former CBO analyst Philip Joyce attesting to that fact. In a biennial budget, we would see more budgetary uncertainty, not less.
2. Frequently revised budgets. There is a good chance that because of the inaccurate forecasting, there would need to be off-year appropriations to fill in for unforeseen gaps. This does not account for additional political pressures for additional appropriations that were not political fodder in the previous year.
3. High likelihood of increased spending. Supplemental appropriations increase government spending (De Rugy, 2011). Given the probability of supplemental appropriations in a biennial budgeting process, it is unlikely that biennial budgeting would be cheaper. The libertarian Mercatus Center highlights a couple of studies showing how biennial budgeting translates into higher spending (Fichtner et al., 2016). Plus, a biennial process would give Congress less opportunities to reform non-Social Security entitlement programs, which are major drivers of government spending. 
4. Outdated budget decisions. The Left-Leaning Center on Budget and Policy Priorities makes a good point. Not only is biennial budgeting done too far in advance, but policy considerations and priorities can shift a lot in a year. An annual budgeting process is more adaptable than a biennial one.  
5. Questionable increase in oversight. Congress has hundreds of committee and subcommittee hearings every year (see Brookings Institution stats here). The issue here is not oversight, but lack of follow-through on oversight findings. Biennial budgeting is not going to make Congress more likely to act on these findings. Plus, let's consider that Congress would have to audit twice the spending in the same amount of time. I would not consider it plausible under those circumstances that Congress would have more time for oversight. If anything, there would be less. 
6. Little to prevent budget shutdowns. One would think that biennial budgeting would put an end to the political squabbling and the threat of government shutdown. However, let's think of this intuitively. In a biennial budget, there is more at stake because there are two years of government spending at stake, not just one. Raising the stakes could make the politicking of budgeting worse, not better. 

Conclusion: One of the advantages of a federalist system is the states act as "laboratories of democracy" to see which policies work and which do not. In the past fifty years, fifteen states switched from biennial to annual budgeting, whereas only three states switched from annual to biennial budgeting. As state budgeting processes get more complex over time, we see a switch over to annual budgeting, and for good reason: it works better. And if we see larger states switching over to annual budgeting, wouldn't we think that, a fortiori, that annual budgeting would be better for the federal government? I hope that Congress does not convert the federal budgeting process into a biennial one simply as feel-good policy. Otherwise, it is the American people that figuratively and literally pay the price.

What Uber and the Ridesharing Industry Can Teach About the Gender Wage Gap Myth

Our economy is evolving in such a way where there is an increase of positions in which companies contract with independent workers and freelancers for short-term engagements. This phenomenon is known as the "gig economy." One of the main features of the gig economy is that these contracted positions give workers flexibility in their work-life balance. Economists were hoping that the gig economy could help eliminate the wage gap. However, that might not be the case. Last month, five economists released a study entitled The Gender Earning Gaps in the Gig Economy: Evidence from Over a Million Rideshare Drivers (Cook et al., 2018). 

Instead of the wage gap being non-existent, the study found that there was a wage gap of 7 percent. The other interesting part of the study was that the wage gap in the ridesharing industry had nothing to do with gender discrimination, which is all the more significant considering that the algorithm and the dispatch are gender-blind. The study found that the wage gap is due to three factors: experience on the platform (learning-by-doing), preferences of where and when to work, and preferences for driving speed. Male Uber drivers work more, they are more likely to drive in areas and times in which pay is higher, and they drive 2.5 percent faster than female Uber drivers.

The study is significant given that the sample size is 1.9 million Uber drivers and nearly 2 billion ride-sharing trips. This study provides statistical evidence that there is a wage gap without gender discrimination. However, these findings are confined to one market. As fascinating as these findings are, we cannot say that there is a lack of a wage gap based on gender discrimination....at least not with this study alone.

I wrote on this topic five years ago. Although it was five years ago, my general conclusion about the wage gap remains the same: there are other factors that account for the wage gap. Men work longer hours, men are more attached to the labor market, there's a discrepancy in occupational hazards, and there is a discrepancy in the choice that men and women in job selection. As the American Enterprise Institute pointed out in August 2017, when factoring these considerations in account, the gender wage gap is all but nonexistent (also see Blau and Kahn, 2016; Furchgott-Roth, 2016). These findings are not confined to the United States. A 2016 study from the Hay Group compared across 33 nations, and found that the when looking at an apple-to-apple comparison with the same level, company, and function pay gap, the wage gap on a global level is only 1.6 percent.

The Uber study is additional evidence that the gender wage gap is more myth than anything else. Let's just remind ourselves of that fact when Equal Pay Day comes around on April 10th.

Thoughts and Reflections of the Family and Medical Leave Act at 25

Life happens sometimes, and that was the impetus behind the creation of the Family and Medical Leave Act (FMLA). 25 years ago yesterday, Congress signed the FMLA into law. The purpose behind this labor law is to cover job-protected leave for medical and family reasons, including pregnancy, adoption, placement of a foster child, medical leave, and family military leave. The FMLA allows for eligible employees to take up to 12 weeks in a calendar year to tend to these matters. Unlike other countries' leave policies, the FMLA's leave policy provides unpaid leave only. The unpaid provision has caused much controversy over the years. Nevertheless, I do wonder if FMLA is successful policy. Here are some thoughts I have on the FMLA and paid leave more generally.

• Limited eligibility. Since the FMLA only applies for employers with over 50 employees, 44 percent do not even qualify for FMLA leave. Furthermore, FMLA only covers employees that have been at their employer for 12 months. This exemption does not cover a considerable minority, which means coverage is not as extensive.
• Many eligible people do not take FMLA. According to a 2014 survey commissioned for the Department of Labor (DOL), 45 percent of eligible employees did not take FMLA leave because it would cause financial hardship. This finding should not be surprising since FMLA only provides for unpaid leave. The DOL survey also showed that a third of those who ended up taking FMLA cut their leave time short to cover lost wages.
• Wage gaps. A 2016 study from the National Bureau of Economic Research found that FMLA had a small, positive effect on the women's employment level, but had no effect on the wage gap (Blau and Kahn, 2016). 
• Employment gaps. Even the New York Times points out that although FMLA makes it 5 percent more likely that a woman remains employed, FMLA makes it 8 percent less likely that a woman receives a promotion. 
• There is more than a philosophical consideration to the notion of paid maternal leave. Other countries can provide us an idea of how family leave can affect female employment. Looking at Chile as an example, paid maternal leave helps ease the transition back into work. However, the law also had the result of their wages being 9 to 20 percent lower than without paid maternal leave laws (Prada et al., 2015). Spain's maternity leave law was not any better. In Spain, companies were 6 percent less likely to hire childbearing women, 37 percent less likely to promote them, and 45 percent more likely to dismiss (Fernández-Kranz and Rodríguez-Planas, 2013). In a study of 22 countries, family-friendly policies in Europe makes it more possible to work, but also more likely to be stuck in a dead-end job (Blau and Kahn, 2013).

Postscript: Although proponents of paid leave can and do point out the flaws of FMLA, there is also the market failure of a lack of paid leave to consider. Only 13 percent of employers offer paid family leave. This is more pronounced by socioeconomic class. Only 4 percent of those in the lowest 10 percent have access to paid leave. What I can say is that 25 years of FMLA, my guess is that FMLA is not going to get repealed. It will probably not even get expanded. If anything, it is likely to get replaced with paid family leave for at least two reasons. One is that the United States is the only developed country without paid leave. This fact helps give the "paid leave" movement some traction. The second is that mandatory paid leave is gaining in popularity. Pew Research found that over 80 percent are in support of mandatory paid leave, although there is more of a division as to whether it should be mandated by the government or employers should choose themselves. The Right-leaning American Enterprise Institute is on board with paid paternal leave. 25 years after FMLA, the conversation about family and medical leave is not going away. We should continue to have that conversation.

I analyzed mandated paid leave four years ago, and am tempted to take another look at paid family leave in light of the increased momentum. Looking at the FMLA reminds me a golden rule of policy: every policy has its tradeoffs. This is especially true with employee protection laws, whether that is menstrual leave or France's employee protection laws. Yes, paid leave makes it easier for women to return to work. There is evidence that it provides some job security. However, as we see above and in the other hyperlinks in this paragraph, they come with a price. Women are less likely to get promoted or hired in the first place. Labor economists generally find that the cost of fringe benefits primarily fall on the employee, which is something to consider. As we move forward with the "paid leave" debate, we need to keep these costs and benefits in mind.