Hai hai R(ab)BA
[Curious about the title of the piece? Scroll to the bottom for an explanation or, even better, read all the way through!]
So that was a heavy week. I had thought that if there was one institution we can rely on for reason in this time of uncertainty, it would be the boffins at the RBA. Surely we could count on them to deliver the goods - “the goods” being that sweet-as general equilibrium thinking that us immigration economists yearn for (I’ll come back to this). Sadly we were mistaken. For those not in the know (read: have a life) the RBA Governor Phil Lowe made a speech to the Queensland Economics Society, where he said, among mostly sensible things, that immigration had lowered aggregate wages in the economy. This obviously sent quite a few of us for a spin. Immigration/labour economists were a bit puzzled. Had the RBA suddenly lost access to its journal articles catalogue? Maybe the internet cut out for a day.
There are a couple of major studies that show this link between immigration and wages to be false. I talk about some of these in a presentation to the Monash University’s WOMBAT (Workshop Organised by the Monash Business Analytics Team) conference and will elaborate more on this slide below.
I also highly recommend the Ace Attorney meme format below. Basically you could just watch the clip and understand more about this literature than most economists.
Right, you’re persistent. Read on then.
Mariel Boatlift
So let’s start with the big one - Card 1990. This paper looked at the Mariel Boatlift. In April 1980 Castro opened the ports of Mariel In Cuba and anyone who wanted to leave the country could do so. About 125,000 Cubans found their way to Miami, Florida raising the labour force by about 8 percent.
David Card looked at the outcomes of native workers in Miami and compared this to four other comparison cities - Atlanta, Los Angeles, Houston, and Tampa-St. Petersburg. He found that - the Mariel influx appears to have had virtually no effect on the wages or unemployment rates of less-skilled workers, even among Cubans who had immigrated earlier. Card reckons that part of this null effect was because Miami had experience in absorbing large waves of migrants in the past. Also migrants supply labour, but they also consume goods and services, and in doing so they stimulate demand
And no one ever debated the effects of migration on local workers ever again. The end.
Yeah, nah.
There was another possible mini boatlift in 1994 with tens of thousand cubans trying to make it to Miami but the Clinton administration intervened and the boats were instead diverted to a base in Guantanamo bay. So Angrist and Krueger decided to see what the employment metrics in Miami and the four comparison cities were in response to this non-shock - “the Mariel Boatlift that did not happen”.
For Whites and Hispanics, the unemployment rate fell in Miami and fell even more in the comparison cities between the pre (1993) and post periods(1995), though the difference between these two changes is not significant. This is consistent with a causal interpretation of Card's (1990) results, which attributes the difference-in-differences to the effect of immigration. For blacks, however, the unemployment rate rose by 3.6 percentage points in Miami between 1993 and 1995, while it fell by 2.7 points in the comparison cities. The 6.3 point difference-in-differences estimate is on the margin of statistical significance (t = 1.70), and would have made it look like the immigrant flow had a negative impact on Blacks in Miami in a DD study. Since there was no immigration shock in 1994, this illustrates that different labor market trends can generate spurious findings in research of this type.
So Card - 0; Angrist and Krueger -1 ? Not so much.
Peri and Yasenov then released another paper (I know this is getting tiring, bear with me for a few more paragraphs) that used a more technical synthetic control method than Card 1990, and used tighter measures of labour market outcomes hourly wages rather than retrospective yearly wages. They found support for Card’s original paper.
Enter Borjas. Borjas was one of the more prominent economists to study immigration and apply an economics lens to it. Sadly many of his models have been superseded. To give you a bit of a taste of where things are at the moment with ol’mate George. The moniker, “doing a Borjas” is synonymous for holding restrictionist views on immigration. Anyway, his paper in 2017 showed that once you looked at very small pockets of labour markets - dropouts by race, which he claimed were the groups most similar to the Cuban migrants entering Miami - you see negative effects from migration. There was one significant problem with this - Borjas samples were too small and you just couldn’t run the analysis he did on them without getting incredibly biased results, and there were compositional changes to the sample in the pre and post wave data (see Clemens and Hunt).
Natural experiments don’t come around very often. There’s a list of well known ones here but for the most part, it’s unusual. So how do you estimate the labour market impacts of migrants when there is no natural experiment?
Shift share instruments
That Mariel Boatlift section was the easiest to explain. Slightly more difficult is the shift share instruments (in this case, Skill-share instruments) that are used to estimate the effects of immigration .They’re useful because well, you don’t need to engineer something like the Mariel Boatlift to draw estimates of the impacts of migration. You just need a decent enough data set with a good read on the number of migrants coming into different sections of the labour market (Where sections are defined as education categories, and experience cents (1-40 years of experience in 5 year increments). Turns out it’s not that easy to do with Australian data but a few of us have had a decent crack at it.
The first paper is Breunig, Deutscher and Tho 2015 which was written for the PC’s report on the migrant intake. They found no evidence to suggest that migration had harmed local workers - what about incumbents. I redid this research and applied it to new data in 2019 and found similar results to Breunig et al. That migrants didn’t take away the jobs of local workers, and that there was no evidence that they depressed wages in the economy either.
There’s another excellent paper by Crown et al which links together temporary skilled migration data from DHA and analyses HILDA data. The study suggests that in response to waves of temporary skilled migration, wages of native workers increased, and it also induced native workers to take on occupation that had “higher intensity of communication and cognitive skills”.
Now to that General Equilibrium point I mentioned in my introduction. As economists, we often estimate things holding something constant to see the effects - ceteris paribus - and this is called a partial equilibrium analysis. It only looks at how things are affected in one aspect of the economy. In this instance, it’s the labour market. But obviously migrants impact other parts of the economy as well. Migrants supply labour but they also demand goods and services. Migrants in Australia are younger and help with Australia’s ageing population. Migrants also pay more in taxes than they take out with primary skilled migrants in particular contributing 319,000 dollars over the course of their lifetime, and adding 4.2 million dollars in GDP. Migrants are also a part of our customer base - international education has ballooned as a sector and is an important part of the fabric of our society. Many international students (though low as a percentage of all international students-~16%) decide to stay on in Australia after they finish their degree. The author of this piece is one of them.
Not satisfied yet? Well a lot of those studies regarding the US usually feature lower skilled migrant workers and they find no effects. What’s more the US decided to look at this comprehensively through this bumper study by the National Academy of Sciences and found similar results.
Perhaps a Nobel Prize Winning economist might convince you, then?
Coming back to the RBA
But despite the above we get the following quote from the Governor
A flat supply curve means that a shift in demand has only a small effect on prices, or in this case wages. In my view, this is one of the factors that has contributed to wages being less sensitive to shifts in demand than was once the case.
It didn’t take long of course for some of the economics commentariat, who had been waiting a bit too eagerly for this moment, to run their takes to the presses.
The most prescient among us, Michael Potter, understanding what would follow in the days (hours even) after this speech, asked the Governor a clarifying question (paraphrasing here)- “your comments today might be misconstrued as being critical of the use of overseas workers. Would you like to clarify them?”
The Governor responded as one might expect - that he was not criticising the use of overseas workers. Allowing workers into Australia increased the productive capital base in Australia and we wouldn’t have been able to have investment fundamental to Australia’s prosperity without them. He added that he was looking forward to the day migration returned because it was important for the economy.
But sadly it was too late and these messages of support for immigration were ignored by journalists (and their editors) whose headlines screamed that economists had “admitted” that immigration lowered wages in some appreciable way.
If only it were that simple.
[Rabba means God in Hindi/ Urdu/ Arabic; Hai Rabba - Oh God]
Some more links:
Article by Greg Jericho on this in the Guardian
A survey of natural experiments in immigration econ and How mass migrants affect weak institutions