Economics 15/08/2009: Migrants in Ireland - a recent study

Here is a quote from the recently published paper on immigration to Ireland, Alan Barrett (ESRI) "EU Enlargement and Ireland’s Labour Market" IZA DP No. 4260, June 2009 that actually getting me going, folks (emphasis is mine):

"In order to get a sense of the educational profile of EU10 immigrants, we need to
draw on earlier research. Barrett and Duffy (2008) show the education levels of EU10
immigrants, along with those of other immigrants based on data from 2005. In Table
5, we present their figures.

The first point to be taken from the table is that Ireland’s immigrants, in general, are relatively highly educated. We know from Barrett et al (2006) that about 30 percent of the Irish labour force have third level qualifications. Hence, the proportion of immigrant with third level qualifications, at over 40 percent, points to a high-skilled inflow.

As regards immigrants from the EU10, although they have the lowest proportion of highly educated across the immigrants groups, they still compare favourably with the domestic labour force in terms of skill levels."

Now, there are several things going on in this statement. Here is the table Alan actually refers to:
Obviously, disregard the USA part - there are only 28 observations in the entire sample of American respondents in 2005 data set (couldn't find any Americans, folks?). But the rest is pretty much in line with what we do know... And here are the issues:

• 6.4% of the EU10 migrants report only a primary level or less - the highest number for all migrants. For all Irish residents, this figure was around 14% in 2008 and it was 16% in 2005. The fact that the EU10 citizens come from the countries with compulsory primary and secondary education helps, and the quality of their primary and secondary education should be pretty comparable to that of their Western European, American and Irish counterparts.
• 9.3% have lower secondary education and 37.8% have upper secondary education. Seems like EU10 workers are more educated. They are not - these percentages refer to the highest level of education achieved. So 53.5% of the EU10 residents in Ireland have attained, as their highest level of education only 'at or below higher secondary level'. This is more than any of the other migrant groups: 47.3% for UK, 25.7% EU13, 33.4% for Other and 39.2% for the entire migrant population. In other words, more EU10 migrants stopped their education at higher secondary level than any other group of migrants. For the overall Irish labour force this number was 65% - meaning only 35% of Irish resident labour force has moved on to reach higher educational levels.
• Of course, any education below a full third level degree means little in terms of skills - Alan should know this. For anyone without a third level degree, their skills are determined largely by the length of tenure and work-related training. I wrote about huge returns to tenure in Ireland, relative to education and this surely explains much of wages differentials for migrants. It is slightly surprising seeing a good economist, such as Alan, occasionally mixing up the two concepts. Here, of course, EU10 migrants lose, for they are (a) relative new comers (have not enough tenure to acquire requisite skills); and (b) might have greater language barriers to absorbing on the job training as efficiently as their UK, US and Irish counterparts. One must also add that Irish companies are not really as well advanced in formal and structured on-the-job training and that leaving work training to FAS (as many do) is a veritable disaster. All of this likely reduces actual skills of the lower (below college degree level) educated migrants.
  
• Now, consider those who actually finish their third level degree or progressed above it. EU10 = 19.2%, Irish labour force average = 16% (recall we are comparing 2005 figures here). A small differential, which is likely to be statistically significant only if the data is measured accurately (it is not - see below). But when it comes to comparing EU10 nationals to other migrants - well, they are not in the running, are they?

So Alan's conclusion that EU10 migrants are relatively well educated only holds water when compared against the native workers, but not against other migrants, and if we are to assume that data for the EU10 migrants is free of self-reporting biases and errors, and if their degrees are actually fully comparable to those in the rest of the world.

Here is a different look at the same data (augmented with the latest CSO stats):There are more fundamental difficulties in making these comparisons that really are not Alan's fault, but do distort analysis.

One simply cannot bunch those EU10 migrants who arrived before the Accession 2004 (many) and those who arrived after (even more). In econo-speak, there are cohort effects.

These cohort effects distort 2004-2005 data, because in those years, majority of EU10 citizens residing in Ireland were most likely the same residents who lived here before 2004. Few years back I published a paper on the topic and showed some evidence that the pre-2004 group - selected under meritocratic migration policies - was indeed of a better quality than those that followed them post-2004. Not surprisingly, given that back pre-2004 they had to prove that they can compete against Americans, Europeans, Asians and the rest. Post-2004, this requirement was removed - the EU10 states were simply encouraged by this Government to displace workers from elsewhere. Of course, such displacement took place in the sectors where skills are less important than brawn - construction, retail, hospitality. Hence, not surprisingly, many of post-2004 workers were elected (and elected themselves) into lower paying domestic economy jobs, for:
(a) the path of least resistance made them move into jobs available to them, and
(b) they actually might have had difficulty proving to productivity-focused MNCs and a handful of externally trading domestic firms that they have better skills than, say, Indian software engineers, American finance specialists and so on.

Third problem is in the data itself. Virtually every taxi driver in NY is a self-reported 'medical doctor', 'dentist', 'pharmacist', 'physics professor', 'engineer' or a 'lawyer'. And yet none work in their fields, despite the US operating an extremely meritocratic system of qualification examinations to confirm medical degrees, for example, or to pass your bar exams, has comprehensive degree recognition culture and requires no specific certification for many fields. This is the issue with self-reported data when it comes to educational attainment questions - it is often of extremely poor quality.

Workers from the UK or US or EU13 might have fewer reasons to embellish their qualifications - they do not perceive this labour market to be discriminatory toward them. And, if you hold a degree from an internationally ranked university, you have nothing to prove to anyone. Those from the countries identified by their compatriots or Irish media or trade unions as being at risk of discrimination will have an incentive to gold-plate their qualifications. And if the university from which you obtained your degree does not rank in top 100-200 in the world, well - you just might add that extra claim to your qualifications, to strengthen your CV. Not all will take such steps, but some (how many?) will.

These, in my view, are very interesting areas for inquiry. I certainly wish Alan would have explored at least some of them. And I certainly hope he will update his data sources, for 2005, hmmm - that is ancient history by today's norms.