Welfare Magnet Theory: Do Large Welfare States Attract Low-Skilled Immigrants?

To leave one's home to live in another country is a huge undertaking. It involves leaving behind that which you once knew, and once you arrive, you have to get used to a new culture, a new way of doing things, and in some instances, a new language. This acclimation and shock are all the more pronounced if you come from a lower socioeconomic status. Why do these individuals decide to leave their comfort zones and migrate to another country? There are a number of reasons one makes the choice to live in another country. One possible answer is that of the welfare magnet theory.

The welfare magnet theory hypothesizes about the effect that the existence of a welfare state has on migration flows. The larger the welfare benefits, the more incentivized migrants, particularly low-skilled migrants, are to move to a given country. It seems intuitive enough of an idea. The question is whether it plays out in reality.

Harvard economist George Borjas was one of the first ones to argue on theoretical grounds that the phenomenon exists (Borjas, 1998). There has been some empirical work since then to rebut the theory. Here are four particular studies:

• The Personal Responsibility and Work Opportunity Act of 1996 included a provision to allow states to provide cash welfare to immigrants. Since some states opted not to, it would follow that immigrant families would have migrated to states that had the benefits. However, the law did not have that magnitude of migration (Kaushal, 2005). Perhaps the cash incentive was not large enough to induce migration. 
• One study showed that poor, single mothers were not particularly inclined to move across state lines for the welfare benefits (Levine and Zimmerman, 1999). 
• Another study showed that the empirical evidence for the welfare magnet theory "does not uniformly support this theory" (Bitler and Hoynes, 2011).
• One study from the Journal of the American Medical Association: Pediatrics came out earlier this week showing that the expansion of public health insurance for non-U.S.-born children does not create a "welfare magnet effect" (Yasenov et al., 2019).

To be fair, the aforementioned studies are measuring interstate migration patterns. Migrating across international borders involves a whole different set of incentives and challenges. On the other hand, if one is not willing to migrate interstate, one would a fortiori would not be incentivized to migrate internationally. What does international research have to say? One paper uses an econometric model for the European Union to suggest that it does exist (Razin and Wahba, 2011). One study countered by postulating that not only does the magnet not exist, but in some countries, immigrants experience less welfare dependence than their native counterparts (Giulietti, 2014).

If it didn't feel coincidental to have one study on the topic released this week, how about two studies? We already covered the first one (see above), but there was one released in the National Bureau of Economic Research (NBER). A Princeton economist, Henrik Kleven, worked with two of his subordinates to look at a Danish case study. According to this research, it turns out that the welfare magnet theory is much more than a theory (Kleven et al., 2019). The authors go as far as stating that this is the first real piece of causal evidence on the welfare magnet theory.

After sifting through this evidence, I still maintain some skepticism because there are multiple factors that induce migration. Let's use the United States as an example. The Congressional Research Service, a nonpartisan public policy research institute of Congress, found that major factors of Central American migration include poverty, natural disasters, political persecution, and gang violence. At best, I would say the recent NBER study would hold for Denmark only, and that it is a single case study.

I personally don't hold any stake into the veracity of the theory. I recognize that it has political implications for immigration policy throughout the Western world regardless. It could be used by anti-immigration movements to restrict migration flows. As I have brought up in the past, immigration is a net benefit for the host country, even when it comes to low-skilled immigrants. If the welfare magnet theory ends up being true, I would find it more of an indictment of the welfare state and provide an argument that we should scale back the welfare sate. After all, the famous economist Milton Friedman found the welfare state to be incompatible with open borders because of limited resources. At the same time, Friedman was against having a welfare state in the first place.

Plus, it is worth noting that the welfare magnet theory is not the same as the fiscal burden hypothesis, which states that migrants pose a net fiscal burden on the native population. Evidence shows that immigrants, especially low-skilled immigrants, do not cause a fiscal burden (see here, here, and here). If you're worried about welfare usage, look no further than Cato Institute research from 2018, which found that natives are more likely to use welfare benefits than immigrants.

Getting back on topic, the welfare magnet theory should not be weaponized against immigrants. Even if the theory does turn out to be true, it is also true that low-skilled immigrants are a net benefit to this country. If people are worried about draining the welfare system, perhaps they should ask how it got so big in the first place and what could be done to shrink it so insolvency doesn't swallow us whole.

A Soft Drink Tax Is Hardly A Sweet Policy Idea

Sugary drinks, and sodas in particular, are tasty. popular among people in the United States. About half of Americans drink a soda every day. The catch is that they are some of the unhealthiest drinks out there. In response to the unhealthy nature of sodas, municipalities have enacted soda taxes. The purpose of such a "sin tax" is to levy an excise tax in order to produce healthier outcomes for Americans.

Mathematica Policy Research, a nonprofit research firm, released a study a few weeks ago about the effects of such taxes in four cities: Oakland, Philadelphia, San Francisco, and Seattle (Cawley et al., 2019). The study's main finding is that a beverage tax rate of 1¢ per ounce, it translates into a reduction of 53 ounces consumed per month (or a 12.2 percent decrease).

I want to take the Mathematica finding at face value and find the impact of the soft drink tax. Let's walk through a back-of-the-envelope calculation of mine. The average American consumes 38.87 gallons annually, which is about the equivalent of eight 12-ounce cans in a week. With the average can of soda having 150 calories, that amounts to 1,200 calories a week (or 171 calories a day). The average American consumes 3,600 calories, which means that means sugary drinks account for 4.8 percent of daily caloric consumption. The soft drink tax reduces consumption by 21 calories per diem, which is 0.6 percent of caloric consumption.

Asking what optimal caloric consumption is tricky to determine since it depends on sex, height, current weight, level of activity, and metabolic rate. Plus, the quantity is different if you are trying to maintain current weight versus if you are losing weight. Even with the assumption that the average woman needs 2,000 calories to maintain and the average man needs 2,500 calories to maintain (and that's for maintaining weight, not losing weight), it illustrates how little the soft drink tax has done to cut caloric intake adequately to reduce obesity.

None of this factors in what economists refer to as the substitution effect. The substitution effect is when a change of prices of one good results in the increased consumption of another good, typically because the alternative good is cheaper. Much like I asked with trans fats, my question here is "When the tax is implemented, what gets replaced with soda?" As the Left-leaning Urban Institute points out:

While sugar is consistently identified as contributing to obesity, it is not the only factor. And the health effects and medical costs of obesity are not uniform. Some consumers with no risk of harm or medical cost will pay the tax. Meanwhile, others may substitute equally or more unhealthy options (such as  alcohol) to avoid the tax.

I bring up these questions and points because the truth of the matter is that the primary goal of the soft drink tax is not to reduce soft drink consumption per se, but rather to reduce obesity. This line of thought corresponds with the past research that has shown that taxing sugary products does not a) cause significant weight loss (Fletcher et al., 2014; Fletcher et al., 2010; Sturm et al., 2010), b) greatly shift overall caloric consumption (Epstein et al., 2015; Fletcher, 2011), or c) does not automatically cause consumers to purchase healthier options because they might be driven to purchase unhealthier options (Seiler et al., 2018; Wansink et al., 2012). Another fun finding with the Philadelphia soft drink tax is that enough people went to neighboring areas to purchase soft drinks, which offset the decrease of sugary drinks attributed to Philadelphia's soft drink tax by 56 percent (Roberto et al., 2019). And if that weren't enough, here is a literature review conducted for New Zealand's Ministry of Health. After examining 47 studies on sugar taxes, the conclusion of the authors was that "we have yet to see any clear evidence that imposing a sugar tax would meet a comprehensive cost-benefit test."

I am not here to cast doubt on Mathematica's methodology or its findings because Mathematica is a reputable research organization. What I would like to remind us is that we need to have a fuller understanding of the effects on a policy such as a soft drink tax, not simply one aspect that sounds nice but in reality has little to no bearing on the primary goal of obesity reduction.

2-24-2022 Addendum: In 2018, Seattle adopted a soda tax. A study found that people were substituting soda for beer (Powell and Leider, 2022). The people of Seattle were swapping out one calorie-heavy drink for another. In other words, we have another study that shows that the soda tax does next to nothing to help with obesity. 

Warren's "Medicare for All" Math: A Lot of Wishful Thinking Plus Avoiding Her Tax on the Middle Class

There has been a major push for single-payer healthcare vis-à-vis Medicare for All (M4A), one that was most prominently started by Senator Bernie Sanders (I-VT). Other Democrats have followed suit, especially Senator Elizabeth Warren (D-MA). Much like so many proponents of single-payer healthcare in the United States, Warren has not able to answer the ever-elusive question of "How to pay for it?" Warren has been so evasive of the question that Saturday Night Live made fun of it this past weekend.

This past week, Warren finally released a response as to how she would pay for M4A. As the memorandum points out, Warren anticipates six forms of funding for M4A:

1. Employer Medicare Contribution. Instead of making employer-sponsored healthcare contributions, Warren would have employers make the contributions to the government for M4A. Total amount: $8.9 trillion
2. Additional Take-Home Pay. Employees would no longer pay their part of salary to their health care premiums or health savings accounts (HSAs), hence the additional take-home pay. Total amount: $1.4 trillion
3. Taxing the 1%. Combine taxes on financial firms, on corporations, and 1% of individuals. Total amount: $6.8 trillion
4. Improved Tax Enforcement. Attempts to fight tax evasion and improve tax compliance. Total amount: $2.3 trillion
5. Immigration reform. Under a Warren presidency, she would allow for greater immigration, which would mean more taxpayers. Total amount: $400 billion
6. Eliminate the Overseas Contingency Operation Fund. Total amount: $800 billion

In total, Warren expects these six policy prescriptions to generate $20.5 trillion over the next decade to fund M4A. Much like I have criticized Sanders' math when he made his proposal in 2016, I am now going to do the same for Warren.

• Warren underestimates the costs of M4A. Warren estimates that M4A is going to cost an additional $20.5 trillion over the next decade. According to the bipartisan Committee for a Responsible Fiscal Budget (CRFB), ten-year estimates of M4A range from $13.5 trillion to $36 trillion. The tricky part of the range is that the rosy $13.5 trillion estimate was conducted for the Sanders campaign. Independent estimates put the amount ranging from $25 trillion to $36 trillion, with the most recent estimate (coming from the Left-leaning Urban Institute) at $34 trillion. What this means is that even if her calculations are correct (which they are not...see below), Warren's estimation is still about $10 trillion short of being able to pay for M4A.
• Warren's plan includes a middle-class tax. What Warren calls an "employer Medicare contribution" is in fact a tax. What else would you call a payment to the government to finance government programs? That's what a tax definitionally is. More to the point, this "contribution" would be a payroll tax. Much to Warren's dismay, general economic consensus is that incidence of the current payroll tax for Medicare is largely borne by the employee, which means that the "employer Medicare contribution" is de facto a multi-trillion-dollar tax on the middle class. The Biden campaign is taking a shot at Warren for this exact reason.
• Warren has rosy projections for improved tax enforcement. Through increased tax enforcement, Warren calculates she could acquire an additional $2.3 trillion over the next decade. This is in considerable distinction to previous estimates. The Congressional Budget Office (CBO) calculated what would happen if the IRS increased appropriations for enforcement initiatives by 35 percent. The CBO calculated that this policy alternative would generate $55 billion over the next decade. Warren's calculations are whimsical enough where she thinks she could get 40 times CBO estimates.  
• Savings from "comprehensive payment reform." Warren credits itself for $2.9 trillion in savings from comprehensive payment reform relative to the Urban Institute study. As former public trustee for Social Security and Medicare Charles Blahous points out in his analysis on Warren's Medicare for All calculations, has already taken credit for these credits, which makes it tantamount to double-counting the savings. 
• Warren's wealth tax is impractical. Warren would like to impose a 6 percent wealth tax on billionaires, and estimates that it would generate $3 trillion in revenue over 10 years. Forgetting the constitutional barriers, a wealth tax would be impractical (see my past analysis on the wealth tax here). No country has ever imposed a wealth tax that high. Countries that had lower wealth taxes repealed them because, as the OECD states, "[the repeals were] justified by efficiency and administrative concerns and by the observation that next wealth taxes have frequently failed to meet their redistributive goals (OECD, 2018, Executive Summary)."
• Warren has already stated that she plans on using wealth tax revenue for other initiatives.
• Warren's previous wealth tax estimate had overstated what it would acquire in revenue. Per OECD data, the average European wealth tax raised 0.27 percent of GDP, with Switzerland raising 0.98 percent. Warren thinks she can pull off 1.4 percent of GDP. How is Warren going to succeed where so many other countries have failed? 
• Skepticism on Warren's payment cuts to providers. One of the ways that Warren attempts to keep costs low is to pay physicians at Medicare rates and hospitals at 110 percent of Medicare rates. Per Warren's calculations, it would save $4.2 trillion over ten years. The closest system that the U.S. has is Maryland's all-payer system, which means that every price is the same, whether private or public. While private systems are 13 percent lower, the public-sector hospitals are 40 percent higher for inpatient services and 60 percent higher for outpatient services. The rates under this system are much higher than average Medicare costs (Pope, 2019). Additionally, the state of Washington could not handle smaller provider payment cuts with its public option. The federal government also could not overrode Medicare physician payment cuts, and eventually did away with the Sustainable Growth Formula (SGF) [see Congressional Research Service report here]. Combined with other countries' inability to keep costs down once single-payer was implemented, it provides enough reason to pause and wonder if Warren could actually implement such payment cuts.
• Warren is going to need to impose additional taxes, including taxes on the middle class. Shortly before Warren released her calculations, CRFB released their findings on whether funding M4A would require a tax on the middle class. Even with high taxes on the 1 percent, financial institutions, corporations, and other policies not included in Warren's calculations (e.g., closing corporate tax loopholes), CRFB calculated that the government could raise $11 trillion over a decade. The CRFB makes caveats with this $11 trillion estimate, the foremost being these are aggressive policy prescriptions that might not be technically or politically feasible. A wealth tax would have constitutional challenges. Additionally, these aggressive policies would likely reduce incentive to work and save, thereby stifling economic growth, which would also have indirect effects on tax revenue. CRFB's conclusion was that we would need to increase taxes on the middle class. 

Postscript: Not only is Warren making unreasonable assumptions about her cost savings, but she is also imposing a tax on the middle class without being honest about the nature of her "employer Medicare contribution." Additionally, she is using an unrealistically low cost estimate to avoid the criticism that single-payer critics are all too right about: single-payer health care is an insolvent payment mechanism and something the people of the United States can ill afford.


11-7-2019 Addendum: If you want another good read on Warren's budgetary manipulation, read this piece from the Federalist. 

California Provides an Argument Against Mandated Paid Family Leave

At least in a U.S.-based context, California is known as a state that is at the forefront of trying policies that are heralded by the Left. One such policy is that of mandated paid family leave. Under the California Paid Family Leave (PFLA), employees are provided partial pay to take off of work for up to six weeks to either tend to the serious illness of a close family member or to bond with a new child. Essentially, the premise behind paid family leave is work-life-balance vis-à-vis providing employees to take on a variety of family caregiving obligations without work getting in the way or needing to quit one's job to meet said obligations. If you want more information on paid family leave, please see my analysis on paid maternal leave from five years ago (see here), my analysis on Family and Medical Leave Act (FMLA), this policy report from the Cato Institute, or you can read this primer from the Congressional Research Service.

Having recently come back from a trip to France and see how they better manage work-life-balance than in the United States in the sense that they work to live (instead of the increasingly common practice in the United States to live to work), it got me thinking about whether it's an important value. Nevertheless, the tricky thing about public policy, especially when it has good intentions, is that it all too often comes with unintended consequences. Looking at the latest study on the PFLA, it seems that paid family leave is no exception. Last week, researchers from the University of Michigan, University of Utah, Middlebury College, and the U.S. Department of Treasury released a study showing that there is little evidence towards the benefits of paid family leave (Bailey et al., 2019). To quote the report:

We find little evidence that PFLA increased women's employment, wage earnings, or attachment to employers. For new mothers, taking PFLA reduced employment by 7 percent and lowered annual wages by 8 percent six to ten years after giving birth. Overall, PFLA tended to reduce the number of children born, and by decreasing mothers' time at work, increase time spent with children.

This finding is significant because one of the arguments used for legally mandated paid family leave is that at least for new mothers, it helps with labor force attachment. Based on these findings, reducing annual wages by 8 percent sure doesn't help with the gender wage gap that liberals are vehemently against (see my analysis on the gender wage gap here, here, and here). And I imagine that reduced employment doesn't do any favors when it comes to trying to get greater female representation in the workforce, nor does it help with make new mothers more likely to stay attached to employers, as proponents predict. While increased time with children is important, there is also the tradeoff of a lower fertility rate, which is problematic for a country that already struggles with a fertility rate below replacement rate.

Yes, this study draws upon robust tax data, has a large sample size, and does so over a relatively long period of time, all of which helps make it methodologically superior to previous paid family leave studies. While case studies have a role in discovering the efficacy of new ideas with little previous empirical data, there are limits to trying to draw general conclusions from this study. For one, PFLA lasts for six weeks. One could argue that six weeks is not long enough (or that it could be too long). Another issue is that PFLA provides 60-70 percent of a worker's wages. Perhaps providing a different amount would create different incentives. Perhaps an automatic enrollment would change the interactions. There could also be other elements within either the culture or economy of California that could make paid family leave less effective than it could be otherwise.

By itself, using this study to rally against mandated paid family leave is inadequate. Nevertheless, it does add to the empirical research showing the unintended consequences of mandated paid family leave. With that being said, here are a few points to consider when thinking of the tradeoffs of mandated paid family leave:

• Paid family leave lowers women's wages. The latest study is not the only one to confirm this point. One study analyzing 21 countries showed that paid parental leave is more effective when the time period is moderate, as opposed to being long (Misra et al., 2011). On the other hand, the same study showed that the same policies contribute to lower wage levels for women relative to men (ibid.). There are also older studies showing the same effect, including those from economists well-known on the Left (e.g., Ruhm, 1996; Gruber, 1994; Summers, 1988).
• Paid family leave affects women labor participation rate. A study from the National Bureau of Economic Research came to the conclusion that paid parental leave was responsible for about 28 percent of the drop of women labor participation between 1990 and 2010 (Blau and Kahn, 2013).
• Paid family leave makes it more difficult for women to receive promotions. A study of paid leave expansions in the United Kingdom not only resulted in fewer female managers, but also exacerbated gender inequality (Stearns, 2017).
• Support for paid family leave is in the details. Much like with so many policies, they sound nice in concept or in theory. That is why support for many Left-leaning proposals has higher support in the abstract. When you ask survey respondents about the details of the Left's latest and greatest policy ideas, support declines (see my analysis on that survey data here). Mandated paid family leave is no different. People assume that paid family leave is a wonderful thing, assuming they don't have to pay for it. When confronted with costs they would have to shoulder (e.g., lower salary, fewer benefits, less promotional potential for women), the support for federal paid family leave diminishes to the point where a majority are opposed (2018 Cato Institute survey).

I will leave you with this thought: whether we are discussing minimum wage, menstrual leave, or other rigid employee protections, they unquestionably come with a tradeoff. That is the economic nature of labor laws, and more specifically, employee benefits. If mothers want to prioritize more time bonding with their newborn children, that's fine. That is a decision they have to make for themselves. But let's not ignore the fact that that choice all too often comes with the tradeoff of less career development potential, a shift in career choices, and lower wages for women. While paid parental leave is becoming more popular, it comes with a price, a price that employers are too happy to ultimately pass either to the customer or their employees. The question is whether the price of a policy such as mandated paid family leave is worth the cost.