How Much Does Immigration Increase Poverty? Less than Robert Samuelson Thinks it Does

November 17th, 2014 at 3:22 pm

It is a common mistake to overestimate the contribution of immigration to the increase in poverty. Today’s purveyor of this erroneous association is the WaPo’s Robert Samuelson, who writes in the context of a discussion about immigration reform:

The influx of unskilled Hispanics has sharply boosted U.S. poverty. From 1990 to 2013, Hispanics accounted for 57 percent of the 11.7 million increase in the number of people below the poverty line.

But of course, not all Hispanics are immigrants, and since he’s allegedly talking about the poverty implications of immigration reform, he should really be looking at, ya know…immigrants!

In fact, there are crosscurrents to the impact of immigration on poverty. On the one hand, as Samuelson implies, immigrants tend to be poorer than average, so increasing their share, all else equal, leads to higher poverty. But there’s another factor in play: the trend in poverty rates among immigrants.

In fact, their population share has been rising which increases poverty, but their poverty rates have been falling, which pushes the other way.

The table shows these results. (The Census data on immigration start in 1993.)


Source: Census poverty data, tbl 23; my analysis.

The immigrant share of the population increases by 4.4 percentage points over these years, but immigrant poverty declines by almost five points, a larger decline than native born poverty which is only slightly down.

So how do we parse out the net effects of these contrary forces? The bottom panel of the table does two little simulations that have a go at this question.

The first line holds population shares of natives and immigrants at their 1993 level but multiplies those shares by 2013 poverty rates. The result is a simulated total poverty rate based on 2013 native and immigrant poverty rates but their 1993 population shares. The difference between this simulated rate and the actual rate can then be assigned to the growth in the immigrant share of the population (and the commensurate decline in the native-born share). As you see, that growth added only two-tenths of a percentage point to total poverty over this period.

The second line uses native shares and rates for 2013, but for immigrants, multiplies the 2013 population share by the 1993 poverty rate. Thus, the only difference between the actual total and this simulated total is the decline in immigrants’ poverty over the past 20 years. This shows that the decline in immigrant poverty reduced overall poverty by six-tenths of a point. In other words, had immigrant population shares shifted as they did since 1993 yet their poverty rate hadn’t declined, overall poverty would have been 15.1% last year, instead of its actual value of 14.5%.

These are very simple simulations which do not account for interactions—one could argue that native poverty rates would have fallen more but for immigrant competition. Also, endpoints matter—you get somewhat different results if you do this over different dates. But the story is pretty much the same (here are the data if you want to fool around with them yourself).

But the larger point is that there are numerous moving parts here. You have to account for not just increasing immigrants’ share of the population, but also “within-group” declines in their poverty rates. Once you do so, at least in my little exercise, you find that the latter lowers overall poverty more than the former raises it.

Print Friendly, PDF & Email

26 comments in reply to "How Much Does Immigration Increase Poverty? Less than Robert Samuelson Thinks it Does"

  1. Larry Signor says:

    When I read Mr. Samuelsons’ post I was skeptical of his 57% figure. Glad you did the heavy lifting for me, Jared.

  2. Gimlet says:

    Honest question: How well do the Census data account for illegal immigrants? My guess would be an undercount, given an understandable reluctance to interact with census takers.

  3. Tiree says:

    Robert seems to imply that we should expect unskilled labor to live in poverty, and that being unskilled is the cause of the poverty. Yet most of these people work at jobs offered by US companies. All he’s really saying is that he looks down on unskilled labor, something without which an economy cannot function. Why should a person picking his fruit live in poverty so that he may eat well while disparaging their existence?

  4. Kevin Rica says:


    You could have simplified your headline: “Immigration – It’s Not Doing As Much Damage as You Thought.”

    That’s a stirring endorsement!

    There are lots of moving parts that you haven’t mentioned either. (Saying that “there are numerous moving parts” doesn’t say much. It doesn’t tell you in which direction they or moving or that they support your argument.)

    For example, what percentage of the “native born” poor are the minor children of immigrants? My guess is that if you eliminate them, the native-born poverty rate probably declined sharply too, because if you are a poor immigrant, legal or illegal, 100% of your U.S.-native-born children are poor too.

    Remember, for the 7-year period from the beginning of the recession until earlier this year, the U.S. economy created no net new jobs. Yet during that period, we took in almost one million LEGAL immigrants per year. Therefore, if they had jobs, they had to displace a native worker, one-to-one onto the unemployment line or out of the workforce. More people, fewer jobs, means a lower employment/population ratio and probably lower wages, particularly at the bottom.

    If people fleeing poverty and unemployment in their homelands created jobs everywhere that they went, would they be fleeing poverty and unemployment in their homelands?

    • smith says:

      One can get an idea of the percentage of native-born poor who are minor children of immigrants looking at the census figures in the report provided below and the graph from the Wall Street Journal. One could subtract the number of immigrants in poverty from the total number of hispanic in poverty (left as an exercise for the reader). The catch is there may be more than one generation separating the immigrant from the latest generation of poor native born descendents, though this doesn’t necessarily mean immigration of the last half century isn’t a source of poverty. Also just using hispanic poverty rates doesn’t excludes other immigrant groups experiencing the a similar burden (new to country, low skills, language barrier, no labor rights, discrimination), thus underestimating the effect.

      Finally, it’s not one for one job replacement, because even poor immigrants create demand, add to the economy. It’s possible even in a recession that their presence could add jobs by increasing demand. However, common sense would seem to point to a bigger negative impact in a contracting economy, vs an expanding economy.

      Triggers that would constrain immigration according to economic conditions were watered down in the proposed bill so as to make them useless to the average worker, low or high skills. Nothing like full employment and rising wages are contemplated.

      • Kevin Rica says:


        “Finally, it’s not one for one job replacement, because even poor immigrants create demand”

        No, actually, if an immigrant takes a job held by a native or previous immigrant, the it’s not new income, it’s just transferred from the previous job holder.

        If wages decline as a result of the additional competition for a fixed number of jobs, then income declines for the employers other workers and the workers of competing employers. Furthermore, immigrants often send money home, diverting a portion of their income to other countries.

        So, on closer examination, on a macroeconomic basis, increased immigration may cause a net loss of aggregate demand and jobs.

        • smith says:

          It’s not one for one or worse because even after taking away money sent to the home country, and even accounting for the lower wages, which can easily be 1/4 to 1/2 less than the going rate, the remaining money earned gets spent somewhere here in the U.S. At the same time, those initially displaced and unemployed do not spend zero amount of dollars either. Demand rises. Before the recession we had nearly the same level of immigration, and fairly healthy unemployment figures. Wages are depressed for everyone because immigrants are not free, don’t have any labor rights, can’t strike, bargain, quit, argue, complain about worker conditions, or safety concerns, leave for another job, very difficult to find a better or higher paying job, or start a business legally (until maybe after five or ten years they get a green card).

          If it was one for one unemployment would have been much higher before the recession.

    • Smith says:

      It turns out Krugman applauds the president but concedes the benefits of restrictions, referred to later here by other’s comments (1924 legislation). He writes:
      “The New Deal made America a vastly better place, yet it probably wouldn’t have been possible without the immigration restrictions that went into effect after World War I.”
      “Once immigration restrictions were in place, and immigrants already here gained citizenship, this disenfranchised class at the bottom shrank rapidly, helping to create the political conditions for a stronger social safety net.”
      “I like to say that if you don’t feel conflicted about these issues, there’s something wrong with you.”

      This is yet another economist that has trouble coming to terms with the evidence that the numbers admitted during recession matter, the labor rights denied employee sponsored immigrants matter, and the bill supported by Democrats and the president would make conditions much worse. There is nothing wrong with admitting this, or not feeling conflicted about supporting commons sense measures to end immigration without full labor rights, and controlling flow according to economic conditions (neither 0 nor present 1/million a year).

      Noteworthy and almost buried is the president’s expansion of high skill labor immigration (spouses can work), and increased mobility (but still way short of necessary labor rights).

  5. Smith says:

    The graph below seems to indicate Hispanic = immigrant, give or take one generation

    Poverty census figures can be found here:
    Figure 4, page 12 shows the end of falling poverty rates just as Hispanic immigration takes off (thanks to supposed immigration reform 1960s style), whether correlation or causation, certainly worth investigating further.
    Table 3, page 13 shows the disproportionate burden of poverty that minorities (black and hispanic as listed in the census data) carry.

    Any true reform measure would end employer sponsorship instead of expanding it. The idea that immigrants are admitted as temporary workers for jobs that can’t be filled with people already in the U.S (both low skill and high skill) is a farce. Instead not only do immigrants compete for jobs with those already here, they are forced to work for less, and with no basic labor rights, they can’t strike, therefore they can’t bargain, and they are not free to quit, or find another job.

    The studies on immigration showing the least impact say that immigrants drive down the wages of other immigrants. The studies showing the most impact say immigrants drive down the wages of those already earning the least.

    All the studies grossly underestimate the impact on wages by not considering downwardly nominal wage rigidities, which masks the unreliability of imperfect substitution. In English, some professor came up with a formula which everyone uses that says the immigrant is not competing with native worker if he’s not earning the same amount. Yet when the boss hires Jose who is doing the same job as Joe, he pays him less, but doesn’t cut Joe’s higher and unmysteriously stagnant salary. It’s a form of discrimination, exploitation, and lack of free labor.

    Good for Obama halting deportation, and may the awful bill that tramples labor rights never see the light of day again.

  6. Larry Signor says:

    Smith, NAFTA had a much larger impact on the wages of low skilled American workers than immigration. We were forced to compete with domestic wage levels of low foreign wage earners in their home countries by NAFTA . In the US we are at least competing with them on US domestic wage levels. It is far from proven that immigration has a substantial impact on the wages of Americans. New immigrants may tend to drive the wage level of present immigrants down, but are generally considered to slightly raise the wage level of domestic workers. You are right, employer sponsored jobs are akin to economic slavery. Our anti-immigration policies need to be seriously rethought.

    • smith says:

      The current trade deficit with Canada is $26 billion, and Mexico $40 billion = $66 billion.
      If 6 million of 11 million immigrants are in labor force (single or head of household) earning sub minimum wages of $10,000/year = $60 billion.
      Very rough back of the envelope calculation doesn’t support impact of NAFTA being much greater than immigration. Also, substantial North American U.S. trade deficits predate NAFTA

      There published papers claiming to prove that “immigration has a substantial impact on the wages of Americans” and even they underestimate the effect due to ignoring downwardly nominal wage rigidities, not to mention common sense.

      I support true reform to allow free immigrant labor, not the corporate sponsored legislation promoted by Democrats and Republicans. There should be buttons printed “Kill Bill”

      • Kevin Rica says:


        The better way of thinking about it is to ask yourself, “What would happen of all the illegal immigrants in this country got fed up waiting for all the neat benefits we promised them for evading the border patrol and decided to leave?” Let’s say one million per month left until 12 or 18 months later they were all gone and 8 million FTE unskilled and semi-skilled jobs were opened to citizens and legal immigrants. What would happen?

        Well, for one thing, if anyone thinks I’ll going to go vegetarian because the meat-packing plants need to replace 40% of their workers, that person needs to be autopsied for the presence of brains. The meat-packing plants will have to pay whatever it takes to attract a sufficient number of replacement workers and they, and the other 60%, will all get higher pay.

        So in the end, wages will go up and unemployment will go down. GDP might be affect a little bit (my guess: 0.25 percent), but GDP per capita will increase (3-4%) and that increase in income will be concentrated among the lowest paid workers.

        In short, we just don’t need mass immigration anymore. It just depresses wages. It’s archaic and harmful, like whaling and denying women the vote.

        I’m a Truman Democrat and I approved this outcome!

        • smith says:

          You don’t want to be a Truman Democrat, they weren’t able to sustain the veto of Taft-Hartley.

          But without the immigrant vote, there would be no Truman Democrat. Immigrants and second generation voters made the election of F.D.R. and enactment of the New Deal possible. They played a central role in the formation of a viable labor movement.

          We don’t want to become Germany, Japan, and China, a presently or future shrinking country of diminishing influence. We could use more pro-family policies like France to promote some population growth without immigration. Eventually population will level out for everyone anyway, but Germany, Japan, and China are poised to shrink.

          Ask yourself why immigration worked before, but promotes poverty and stagnant wages now. The truth is immigration accompanied tremendous poverty in the past, just as it does now. Eventual success relied on key elements missing now, like a growing economy, a rising labor movement, expanded free education, and free labor itself.

          If you shackle immigrants to their employer and deny them the most basic labor rights, they will add to poverty, and suppress everyone’s wage. The new bill seeks to vastly expand the same arrangement for high skills. To mimic the immigrant success story of the past, we would need a growing economy with actual labor shortages, free college (the new high school) and no restrictions on labor rights, and the end of employee sponsorship (work quietly or get deported). Neither can the U.S. survive when states compete with states for corporate give-aways, and foreign governments deny basic labor rights. Americans could stop globalization in it’s tracks with selective boycotts. Unions would do this but they are banned by Taft-Hartley! Secondary boycotts banned thanks to the defeat of Truman Democrats in the midterms.

          • Kevin Rica says:


            “You don’t want to be a Truman Democrat, they weren’t able to sustain the veto of Taft-Hartley.”

            Get real Smith!

            “Ask yourself why immigration worked before..”

            To the extent that it worked before (the record is mixed), we were trying to take the continent away from the Brits, Mexicans, and Indians and we needed the European population to stake our claim to Manifest Destiny. We weren’t necessarily worried about providing immigrants with a minimum standard of living.

            We started cracking down on immigration after the West was won. In the past immigrants came here and often lived in abject poverty.

            And frankly, I’m a big believer in ZPG and if you can get there without a Chinese-style one-child policy, we shouldn’t ruin a good thing with immigration.

            “If you shackle immigrants to their employer and deny them the most basic labor rights, they will add to poverty, and suppress everyone’s wage.”

            If you are worried that they will fall into poverty, then we shouldn’t let them in the the first place.

            We should have a much smaller immigration intake with a Canadian-style, skills-based system. If the employer does not guarantee that the immigrant will receive a middle class salary and benefits (including employer-paid health) and at least a 10% premium to the normal salary scale, they shouldn’t come. If the immigrant leaves and accepts less, the the immigrant should be deported for violating the terms of their admission.

            But in the end Smith, we have a philosophical difference about the purpose of the US Government. I want it to put the well-being of Americans before that of foreigners. You believe it should do the opposite. (Also, I understand the Law of Supply and Demand as applied to the US labor market and you either don’t understand it or don’t care. I haven’t figured out which.)

          • smith says:

            The law of supply and demand is operable, and I’ve stipulated already that numbers alone, whether from population growth or immigration, will tend to suppress wages. However, a vibrant economy usually functions better with a moderately growing population, which encourages innovation, investment, and productivity gains to meet larger demand. Japan, Germany, China, and Italy especially, have not ZPG, but shrinking or soon to be shrinking populations. Immigration works when the economy grows, education is free, and labor rights are enforced. Opening the West was not the story for the great wave of immigrants from Eastern and Southern Europe at the turn of the century (1900).
            The new bill would vastly expand and prioritize high skills immigration, to further exploit and hollow out the middle class. Worker protections are a total sham in the new bill, just as they are in the current system. Immigrant labor is not free, and workers already in the U.S. are left unprotected.

          • Kevin Rica says:


            After the turn of the 20th century (1924, to be exact) we shut the door part way and the number and percent of immigrants fell, including during the expansions of the 50s and 60s. Immigration was at a low point. It was after the big increase in immigration of the 1970s that wages started stagnating and falling. Does not match your story.

            I agree that the new bill stinks, but you actually make the case against all immigration. If we don’t need skilled people, then we certainly don’t need more unskilled people competing for a limited number of jobs. Your assertions are bold (immigration causes growth), but you wind up arguing against them.

            I really can’t figure out what you want – although I know that you are not a big fan of intellectual consistency.

          • Smith says:

            Wages tend to follow business cycle and not immigration patterns.
            Wages climbed during booming immigration years 1900 – 1924
            Wages declined in Great Depression, after the 1924 curtailment of newcomers.
            Similarly wages continued to climb into the mid seventies when wages struggled to keep up with historically high inflation, ignited by surging oil prices.

  7. Larry Signor says:

    Smith, You are not accounting for stagnant wages and US labor market churn due to NAFTA. The majority of immigrants (at least the ones I have worked with) probably do not earn sub minimum wages. NAFTA has a substantial impact on the wages of Americans, immigrants don’t have this same level of impact since they contribute to overall demand. How does NAFTA contribute to domestic demand? How many Canadian immigrants are there? Canadians seem to be satisfied with their countrys’ economic performance. Our trade deficit with Canada has more to do with oil than off-shoring manufacturing, as with Mexico.

    • Smith says:

      This EPI report says NAFTA cost roughly 680,000 jobs, which would add a little over .5% to the unemployment rate. This is contrasted with the roughly 700,000 jobs per year that are taken by immigrants. While in the long run immigration will expand the economy, labor participation of immigrants is 66%, so times 1 million/year, that’s 660,000 per year. This seems to indicate that immigration most likely equals or exceeds the impact of NAFTA. Admittedly immigrants create demand. Unfortunately if they are paid 1/4 to 1/2 less, then in a few years we are already at the 600,000 700,000 level. Obviously since the poverty rate for immigrants is 18%, a substantial portion receive sub-minimum wages. What a majority of immigrants do receive, including the ones I’ve worked with, are sub-standard wages. Numerous studies support the idea that immigrant labor drives down average wages, and even these studies underestimate the effect. And it’s not a probably, anecdotal information comes from actually knowing the wages, and studies tend to show that numbers alone are sufficient to effect wages. That means even if immigrants were not denied basic labor rights, just their added presence affects wages.
      Strike for labor rights in Mexico, strike for the labor rights of immigrants in the U.S.
      Kill Bill.

      • Larry Signor says:

        @ Smith, “That means even if immigrants were not denied basic labor rights, just their added presence affects wages.”…no argument from me here. “Strike for labor rights in Mexico, strike for the labor rights of immigrants in the U.S.”, or here.

        “Obviously since the poverty rate for immigrants is 18%, a substantial portion receive sub-minimum wages. “…this is somewhat spurious, since $7.25/hr. minimum wage= $14,500.00/yr., which raises no one out of poverty (even though this is above the HHS stated poverty level for one person, who can get by on $11,670.00/yr.? Sub minimum or minimum wage, this is embarrassing and damaging to America. NAFTA is responsible for 25% of the lost jobs due to free trade agreements through 2011 (from the EPI), as you no doubt know. (These agreements were touted to “…support hundreds of thousands of new jobs in the United States.”)
        We are being sold a bill of goods on FTAs.

        • smith says:

          Repeal the portion of TaftHartley that allows secondary boycotts.

          • smith says:

            Error in that last statement. allows instead of disallows.

            TaftHartley disallows secondary boycotts, repealing that portion would allow unions the ability to organize selectively against companies that take unfair advantage of overseas workers, and/or threaten to move to countries or states with lower wages.

      • Kevin Rica says:


        Why are you so upset when US jobs are sent overseas, but so supportive when foreigners are brought here to do the jobs that can’t be sent overseas, such as construction, farm work, landscaping, food-service, meat-packing etc?

        Do you realize the inconsistency of your positions?

        • smith says:

          Make a simple model to explain.
          U.S. has two workers. One makes food, the other makes clothes.
          Food worker grows enough food for two people per year.
          Clothes maker makes enough clothes for two people per year.
          Food worker is paid enough to buy a clothes for a year.
          Clothes maker is paid enough to buy food for a year.

          Immigration allows a clothes maker and food grower into country and pays them the same amount as previous workers, no problem.

          Pays them less and everyone’s wages stagnate, plus the immigrants go hungry and unclothed part of the year.

          But if there is no immigration, and a years clothes and food are imported at lower prices, because foreign workers are paid less, the original clothes maker and food grower starve.

          That is how it works. Just add a rich person to buy all the excess and assume the each worker makes more than enough for one person to produce enough for the rich person for a more realistic model.

          • Kevin Rica says:


            I get it. The Phlogiston Theory of Economics.

            It’s really original. You should publish it. There is nothing like that in any econ book that I have ever seen and I’ve seen my share.

            Just a follow-up question. You said, “That is how it works. Just add a rich person to buy all the excess..”

            What if you don’t have a rich person to buy all the excess food and clothes? Could you add eye-of-newt or tongue-of-frog?

  8. Dave Rimshnick says:

    I think there is also an important point here about growth in median wages. If most of the population growth has been below the median, then factoring this growth out may lead to the conclusion that most families who have been here the entire time actually are doing better than they were 10 years ago or whatever.