The Atlantic Dares to Question the Economists

Interesting, provocative piece by David Frum of The Atlantic, titled “Does Immigration Harm Working Americans? Many economists say no—but they may be too glib.”  That subtitle is quite a contrast to the blog today by Jennifer Rubin, the conservative columnist at the Washington Post, in which she says,

Virtually every credible economist has found that immigration increases growth, tax revenue and wages — with the exception of a modest decrease in wages for those without a high school education. In particular, immigrants are over-represented in high-tech start-ups and new patent filers. So for now, why not reform the immigration system for the brainiacs, the highly credentialed and the entrepreneurs only? No serious person can argue that we don’t have a shortage in some high-tech fields or argue against allowing foreign students with advanced degrees to remain in the country.

(Of course, I have made the very arguments Rubin thinks “no serious person” would make.  I support facilitating the immigration of “the best and the brightest,” but otherwise do not feel we should make a surplus any worse than it already is.)

I’ll discuss the substance of Frum’s piece below, but first I want to emphasize how remarkable it is, “daring” in the phrasing of my blog title here.

I use the word dares in my title reluctantly, as I know some will take the word in the sense of, say, “putting his job, career and social life at risk,” which is not what I mean at all.  Instead, I mean “‘dares’ as in, willing to think out of the box, questioning the conventional wisdom.”

In being interviewed by the press often in the last 20+ years, I have seen very few journalists, editors and so on who are truly biased, out to write polemics.  (Rubin may be an exception.)  However, many have unconscious assumptions, often formed in society at large by PR experts and so on.  This forms a climate that profoundly effects our thinking.

When former NY Governor Mario Cuomo passed away last week, I thought of a statement Cuomo — the quintessential liberal — made in 1994 (quoted in To Be an American, by my former UC Davis colleague Bill Hing):

[Immigrants] are part of our strength.  They will be a nourishment for our future…They are also expensive [in terms of use of government services],

No Democratic politician would even consider making such a statement today, nor would one see such a statement in the liberal press.  Nor would the Democratic Leadership Council invite me to write an article on immigrant abuse of the SSI program, as they did in the 1990s for their magazine The New Democrat.  I’m not saying this pejoratively, being a liberal myself, just pointing out how mind sets have changed.

This change became clear to me in a rather unsettling way a few weeks ago in a conversation with a journalist on a major prominent big-city newspaper.  He said he didn’t think that the age discrimination I write about as the scourge of programmers and engineers applies to the foreign-born.  He seemed to buy into the notion that immigrants are hardier than U.S. natives, and when they reach age 35 or 40, they’ll still find programming and engineering jobs.  An outrageous statement — yes indeed, the immigrants suffer from age discrimination too — from an otherwise very thoughtful, reasonable guy, starkly illustrating today’s liberal (and in the case of Rubin et al, business conservative) climate.

Now, what about Frum’s remarks?  He begins by citing CIS research, that shows that not only have most new jobs in recent years gone to immigrants, but also the natives have actually lost ground, i.e. fewer are employed now in absolute numbers. He then points out that this is in stark contrast to what the “experts” say.  In my view, he falters a bit by citing as experts people who don’t actually do real research on immigration, e.g. Nowrasteh, but merely cite (very selectively cite, at that) research of others.  But then he does cite a prominent researcher, another UC Davis colleague, Giovanni Peri.

Giovanni has become THE go-to guy for pro-immigration views these days.  Notably he played that role for the Obama White House, as I recently reported, and here Frum has gone to him as well.  Frum then presents Giovanni’s complementarity argument, and tries to deconstruct it.

Though Frum gets bonus points with me for noting the fragility of statistical regression models, I think he misses a couple of important points about Giovanni’s analysis.

First, Giovanni’s “engineer and construction worker” parable is reasonable if we have a shortage in either occupation.  But we don’t.  We have a surplus, which is eroding wages and job opportunities in both professions.

Second, in contrast to the rosy picture Giovanni painted for Frum, he himself has conceded that highly-educated Americans are indeed displaced by the immigrants, and that there is at least some concomitant wage loss (Giovanni Peri and Chad Sparber, Highly-Educated Immigrants and Native Occupational Choice, working paper, 2008):

…we assess whether native-born workers with graduate degrees respond to an increased presence of highly-educated foreign-born workers by choosing new occupations with different skill content.

…we add to evidence from past studies by showing that [U.S.] native occupational adjustment in response to immigration occurs among highly-educated workers and occurs for those already employed.

As the foreign-born share of highly-educated employment rises, native-born employees respond by moving to jobs with less quantitative and more interactive content.

The wage consequences of immigration were not estimated in this paper…If the evidence from the labor market for less-educated workers is an indication, the occupational skill response among highly-educated natives is likely to mitigate their potential wage loss from highly-educated immigration.

(This last comment is similar to a Peri quote in Frum’s article, but that quote focuses on low-skilled workers, whereas the above quote is for the high end.)

And of course, this flies in the face of the claims that (a) only the low-skilled natives are harmed, not the high-skilled ones, and (b) upward mobility is the solution.

The article touches on the “immigrants as consumers” argument for a positive effect of immigration.  I don’t necessarily disagree, but I’ve been saying for many years that it isn’t so simple.  Just as weather forecasters need to “look out the window” once in a while instead of relying only on numbers, immigration economists need to take an up-close look at immigrant-enclave labor markets.

Take for instance the Ranch 99 chain of Chinese supermarkets, popular in the Bay Area and southern California.  They don’t employ many non-Chinese, due to linguistic requirements, and their clientele are largely shopping at Ranch 99 rather than the mainstream stores.  Just this one example alone makes it understandable that natives might be losing jobs while immigrants gain them.  And the Ranch 99 stores typically anchor entire immigrant-oriented malls, where even the Starbucks is staffed by immigrants, serving immigrant customers.  My wife and I shop in these malls frequently, and especially enjoy the one in Concord, CA, which is largely Latino but with a Ranch 99 too; we patronize businesses of both ethnicities.  But the point is that there are a lot of factors that have the effect of shifting jobs from natives to immigrants (though of course this is a highly complex matter, with effects in both directions).

All in all, though, my hat is off to David Frum and The Atlantic.


26 thoughts on “The Atlantic Dares to Question the Economists

  1. Looking at the Atlantic, and other pieces by Frum, he has taken many positions which question the current received wisdom that open borders are wonderful for Americans. Frum has questioned the logic behind Obama’s amnesty, written several pieces about low-skilled immigration, and other pieces. I’m going to subscribe to the Atlantic to support him.


  2. My translation of this passage from the 2008 Peri and Sparber working paper:

    ….. As the foreign-born share of highly-educated employment rises, native-born employees respond by moving to jobs with less quantitative and more interactive content.

    The wage consequences of immigration were not estimated in this paper…If the evidence from the labor market for less-educated workers is an indication, the occupational skill response among highly-educated natives is likely to mitigate their potential wage loss from highly-educated immigration…..

    The experienced U.S. citizens such as myself are permanently displaced by young imported workers from the positions that made use of our Ph.D.-level STEM training and experience to become poorly-paid call center workers or adjunct faculty members. (My total 2014 adjunct professor compensation was less than $9,000.00)


  3. The Mario Cuomo citation can be found here:

    Even Gov. Mario M. Cuomo of New York, a son of immigrants who is known for praising immigrants, was more sober in a recent radio interview, saying: “They are part of our great strength. They will be a nourishment for our future,” adding, “They are also expensive.”

    Also, Frum offers a pithy rejoinder for those on top who advocate immigration as a means to lower labor costs …
    “The ratio of CEO pay to other workers has skyrocketed. Obviously we are suffering from a glut of workers and massive CEO scarcity. We should issue work permits automatically to any executive with a job offer that pays more than $500,000 a year. Americans with organizational skills will be pressed to shift to the public sector, improving the quality and lowering the cost to taxpayers of government services.”


  4. Yes, well, since when is David Frum an authority on economics? His article isn’t bad, but it’s extremely shallow. As I find I must always point out the reality is much, much worse than his deductive skimming of the subject suggests. Look at the construction industry over the last twenty years. Look at the STEM fields over the last twenty years. These have both been overrun by “immigrants” legal and illegal, cutting wages by 50% or more. If you can’t see that, you haven’t a clue as to what you’re doing. And these are the STRONGEST fields in our economy, which is exactly why the immigrants go there. The weaker job categories have already been shipped to China, there are NO opportunities there, which is the only reason any Americans at all are still in construction or STEM – desperation.

    Mr. Frum, get out in the real world and see with your own eyes. Or keep reading some of the “alternative” economic sites and literature, look under the headlines and you will see a labor market which has been destroyed, compared to what we grew to expect post-WWII, and the few remaining categories have been swamped with immigrants, which is no better for the economy and no more sustainable than is our shipping 100% of manufacturing out to China. The data *is* there in the government statistics, but a few government officials lying about the matter seems to completely derail the mainstream press from running accurate headlines.


  5. > Giovanni has become THE go-to guy for pro-immigration views these days. Notably he played that role for the Obama White House, as I recently reported, and here Frum has gone to him as well. Frum then presents Giovanni’s complementarity argument, and tries to deconstruct it.

    As it happens, I’ve just finished looking at a Wall Street Journal article that was based on Giovanni’s study. I’ve posted the preliminary results at but I plan to add a description and links to the R code that generated the results. As you can see, the WSJ posted a chart that appears to have erroneous and missing data based on that study. As a source, they give “University of California, Davis”. It would be very interesting to know who created this chart, where the data came from, and why San Jose appears to have been omitted. This further convinces me that, as Frum suggests, studies such as this one need to be carefully verified. A key tool for achieving this would be to require all studies that affect public policies to make available the data and code required to reproduce their results. In addition, national publications like the Wall Street Journal should demand this before publishing unverified results from studies.


    • You’ve come up with another excellent set of analyses. Note that your last analysis, showing a negative relation between number of foreign and native STEM workers, is consistent with a paper by Prof. Peri that found the same thing (but which is never cited in press accounts).


      • Thanks. I have noticed that the authors of these studies will often include qualifiers which get dropped by the media. That’s one reason why it was very good for you to state your positions in your post of December 17th. However, I do see that the conclusion of Peri’s paper states “We find non-significant effects on the employment of those two groups [college and non-college educated natives]”. Hence, the paper likewise doesn’t mention the prior findings.

        Regarding the article and chart in the WSJ, I wrote the author of the article and he said that Dr. Peri provided the data and chart. He gave me his contact information and suggested that I direct any questions to him.


      • I just added a detailed description of the R code that generated the plots and tables for the analyses at . Doing so reminded me of the large number of decisions that must be made in such an analysis. In his article, Frum states:

        > First: The story about immigration benefiting all (or almost all) native workers could be true. But that doesn’t mean it is true. Economists prove their claims about immigrant law by drawing regression curves that compare ratios between data sets based upon the number and the pay of immigrant and native workers. Have they drawn their data sets correctly? Did they choose the correct basis for comparison?

        > These technical decisions at the beginning of the calculation have huge impacts on the final conclusion at the end. Between 1990 and 2006, the wages of non-college-educated Americans declined. The less education the worker had, the steeper the decline. How much was immigration responsible? The data the economist chooses to look at will determine the answer.

        To pretend that there are no such decisions and that the data and methods for obtaining the results need not be released for public scrutiny is simply wrong, if not disingenuous. One very big decision that was made in the Wall Street Journal chart was to omit San Jose, the center of Silicon Valley, from the plot. This was done without any mention to the reader. In fact, the article saw fit to state that the “areas with the biggest influx of foreign STEM workers were Austin, Texas; Raleigh-Durham, N.C.; Huntsville, Ala.; and Seattle”. No need for the reader to worry their pretty little heads about San Jose!


      • I have a quick update on the omission of San Jose from the chart that appeared in the Wall Street Journal article. Peri supplied me with the spreadsheet and it appears that Stamford, CT and San Luis Obispo, CA were also omitted. I put all of this and other information from the spreadsheet in a red-typed update in the middle of my page at .


        • You’d make a good academic researcher, Admin. 🙂
          The data in all published papers (including mine, of course) should be available to the publish and subject to scrutiny.


    • Are the compensation figures inflation-adjusted by location? If so, how? I mainly ask because SF is known to have restrictions on residential construction which has driven up that aspect of costs-of-living (a jazz musician’s small house on a postage-stamp recently sold for over $1M, while a comparable house in a comparable neighborhood in Dayton OH might go for $70K). I’ve also experienced some of southern California’s (and Virginia’s and Minneapolis’s…) over-regulation, over-taxation, etc., which are equivalent to an increased cost-of-living or decreased quality-of-living. And besides, they don’t know how, or refuse?, to make decent pizza or barbecue in California. (OTOH, the weather is very comfortable.)

      From talking with people there has been displacement from residential construction, certainly roofing, some kinds of road-work.

      OTOH, to bolster the Rubin/Wadhwa argument, they’ve also driven some grads into “innovative” “entrepreneurial” activities, like running a smoked meat vending truck and other forms of under-employment.


  6. Thanks for pointing out this excellent piece by Frum. (Though it is depressing that, once again, the best elite-media analysis in this space seems to be coming from the right wing–though Frum has annoyed fundamentalists by retreating slightly from his most extreme forays in neoconservativism and Zionism, he’s no “apostate.”) IMHO Frum’s best bit is

    As the technical economists understand, wage cuts and job displacement are the exact and only ways that immigration confers any benefits on native workers at all. It is wage decline and job displacement that drives natives to shift to higher-paid sectors. No wage cuts, no job displacement. No jobs displaced, no benefit to natives. Here’s Peri saying just that: “Large inflows of less educated immigrants may reduce wages paid to comparably-educated, native-born workers. However, if less educated foreign- and native-born workers specialize in different production tasks, because of different abilities, immigration will cause natives to reallocate their task supply, thereby reducing downward wage pressure.”

    When economists minimize the impact of immigration on wages, they aren’t denying that immigration pushes wages down in the jobs that immigrants take. They concede that immigration does do that. They celebrate that immigration does that. Instead, they join their celebration of immigration’s wage-cutting effects with a prediction about the way that the natives will respond.

    But what if the prediction is wrong?

    It’s an empirical question, which the Church of Academic Economics (and the corporate-funded media who love them) conveniently ignores.


    • Why wouldn’nt all corporate cheap labor interests consider higher skill higher wage jobs a more lucrative target for global labor arbitrage through both immigration and offshoring trade policy than lower skill lower wage jobs? After all the nominal bottom line potential is greater, but so is the damage to upward mobility for all workers. It is a top down squeeze play.


  7. When I read some of these “immigration improves…” articles written by economist I get the impression that the authors do not understand the difference between correlation and causation. While they often have weasel words about the difference the only “proof” they present for their conclusion is that events happen at the same time. The fact that some high tech companies have one or more foreign born founders does not prove that an increase in the H-1B cap would foster more high tech companies.


    • “Virtually every credible economist” is a not-so subltle, not-so-clever propaganda technique. You see every “credible economist” agrees with me, while those other economists are not credible at all, so everything they write or say must be ignored.

      The fact is that economists, just like politicians, or leftists/right-wingers, and STEM professionals, do not neatly sort out to be on one side or other of this cluster of issues.

      David Frum does not consistently remain cubby-holed in any one of those boxes… though he seems fairly consistent on this issue since 2007.

      (Ugh! This comment text-box has shrunk down to 2 rows by about 40 characters again. Reminds me of trying to do a write-in on one of those old mechanical voting machines with the paper about an inch back behind the little opening.)


      • @mib8 January 8, 2015 at 11:18 pm: Ugh! This comment text-box has shrunk down to 2 rows by about 40 characters again.

        1. Open Emacs buffer. (Actually I have one I keep open all the time on a scratch file, called ‘notReallyJunk’, precisely to avoid composing text with unfriendly tools.)

        2. Compose response at leisure, utilizing Emacs’ excellent text-editing UI (including character counting).

        3. When done, paste response text into WordPress’ response UI.


          • @matloff January 12, 2015 at 11:31 pm: WordPress keeps changing their text editors, and none of them is very good

            Personally I find the [switch to editor, compose, cut, paste] approach useful not only here, but in many other usecases. Even better, sometimes one’s tools can consume files, obviating the [cut, paste].

            E.g., I continue (foolishly) to be shocked by how often commit messages fail to communicate anything useful. So, instead of `git commit -m ‘I am just too tired to tell you what this is about`, just open a file *while* you’re coding (and presumably testing) describing what you are doing (presumably including how you tested your changes :-), and then `git commit -F /path/to/scratchfile`


  8. Semi-off topic, but 99 Ranch is a great place to shop if you recognize the products on the shelves. I am more puzzled at Trader Joe’s than 99 Ranch. And Mexican places . . . one needs local knowledge to order a meal.


  9. You make a good point about Ranch 99, a store I have only gone to once. Speaking of Concord, the same situation happens in supermercados (Latino-oriented supermarkets) along Monument Blvd and other Latino areas throughout the nation.

    But take it down to the smaller, more common small businesses which probably number in the tens of thousands. I can’t recall a native-born citizen in any of the small taquerias, food trucks, or retail establishments in such areas.

    Many construction crews are also overwhelmingly Latino in Northern and Southern California, and I was driven through a large job site in Texas that was 100-percent Latino. I have several family members who worked “in the trades” 3 decades ago. The workers were white, black, Latino, and an occasional Asian-American. I’m told by relatives in the midwest that when there is a large natural disaster, the roofing crews that come in to re-roof damaged houses are 100-percent Latino. (It seems as if plumbing and electrical workers are still often native citizens.)

    While the general contractor might be American, the experienced men running the job site are Latino, and Spanish is the primary language spoken on the job. I also read recently that while the population in Los Angeles has continued to increase, the number of taxpayers is decreasing, which seems to point to a growing of the underground economy. Another topic, but related and with significant long-term repercussions.


    • Actually, we often patronize the Latino businesses on Monument as well Love those taquerias! The supermercado we go to is in that same mall as Ranch 99. But you’re right; I can’t recall ever seeing any non-Latinos there. Even when I went to the cell phone store for a battery, it was all-Latino clientele and staff (though the latter had good English). All this makes me feel quite comfortable, reminding me of my formative years in East LA. Arriba Soto Street!

      Such cultural richness is wonderful; I revel in it. But as Mario Cuomo said, there are costs — and not just fiscal. People are being squeezed out of “the trades,” not just blue collar but also professional jobs such as software development.


      • I enjoy the “richness” as well, hence why I go there frequently. My waistline is proof. I’ve been invited to one illegal back-of-the-patio “restaurant” in the back of a duplex, and am told there are several such establishments serving the local community. My Spanish is poor, but I give it a shot. The same attempts in San Francisco are met with scorn.

        Thank you for not calling me a racist. Really. I will admit when I see factoids that run counter to my assertion. For example, the only time I have seen a construction site with African American workers recently was a recent article on covering a long-planned development in south eastern San Francisco.

        Some assume that our American values will be adopted by these new groups, but given the historically large infusion, I don’t think that is the case. Are we no longer a melting pot? Does Paris give us something to consider?

        I went on one business call in the South Bay a decade ago and a whole large department was East Indian, I think Sikh. I didn’t ask, it wasn’t relevant, but as I walked through the office it was like I was on parade (I’m not big cheese), and there was zero diversity. There were probably upwards of 100 developers in sight. I also don’t see or hear the need with my friends or acquaintances who are Central or South American to hire a “diverse” workforce. They want the business, period. And I have mentioned before that Asian companies I have worked for have hired zero white, black, or Latino programmers or testers, and only hired a handful of American-Chinese workers. They lived on H1B Visa compliant workers.

        My last visit to the South Bay I felt quite old. Free high-end snacks in the lobby, kids walking around with their heads in their laptop or iPhone, and I didn’t see anyone who appeared to be over 28 years of age.


  10. Interesting analogy at the end of your blog, Norm. How does “knowing a particular computer language” compare to “knowing a particular language” — that is, in your example, Chinese? I wonder of there is a parallel in terms of driving out native Cobol speakers, for example.


    • I assume you are referring to my comment on the hiring practices of the Ranch 99 Chinese supermarket chain. Since the customers are almost all Chinese-speaking, the market employees are either all or almost all Chinese-speaking as well. In fact, it’s not just enough to speak Mandarin; all the ones I’ve seen speak Cantonese too.

      The industry lobbyists like to dismiss programmers who are having trouble finding work as COBOL people, but I know almost none who are in that category. But certainly the lack of background in some particular computer language or OS is the standard excuse for rejecting the older job applicants. It’s usually a phony excuse, because the same employers are also rejected older workers who do know the new-fangled language or whatever.


  11. Two observations about “virtually every credible economist.”

    First, is the Pope catholic? If he is not and wants to keep his job he needs to keep quiet about it. Likewise, any economist who publicly questions free trade or immigration is going to find his job prospects severely limited. This applies to both government and private jobs as well as consulting fees. Those who hire economist are unlikely to hire someone advocating anything interfering with profits.

    Second, most economist in prominent positions today were trained in a time when an increase in the Gross National Product was the same as an increase in the standard of living of the vast majority of the American people. Free trade and increased immigration have broken that link. While GNP and corporate profits have had steady growth, for roughly 80 percent of American workers real wages (wages adjusted for inflation) peaked in 1973. That peak was itself below the 1946-1973 wage trend so the departure from wage growth started before then.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s