The Trump administration is apparently in the middle of crafting yet another executive order, this one dealing with H1B visas. While nothing has been made public, it purports to address abuses. That might be ok, in principle. There are certainly examples of companies circumventing laws to get cheaper employees in the USA.
A big talking point around H1B is whether we need them at all. Many complain there's a STEM shortage, while others point to articles like this that "prove" there isn't. I'd like to address a few things within.
As the article rightly points out, a lack of definitive data or even definition of what STEM is makes this a potentially difficult web to untangle. The author concludes there can't be a STEM shortage because:
1. There are more STEM graduates than there are open STEM positions
2. If there truly were a shortage, wages wouldn't stagnate. Apparently they have?
This is a gross oversimplification and fails to acknowledge how H1B fits into the STEM fields and that not all graduates are qualified to do these jobs (consider those barely passing or going through a for-profit diploma mill). Let's look at some relevant numbers, because the article above left lots to be desired.
There is a cap of 65,000 H1B each year, and 20,000 more for those with advanced degrees (85,000 total). Compare that to a total near 160M across the USA. Each year, we're at most giving an H1B to 1 in 1870 workers. Granted, these are "good jobs" and probably matter more economically, but it's still a fraction of a tenth of a percent. It would literally not move the unemployment numbers as reported to a tenth of a percent.
The literal majority of H1B go to people in the computer-related occupations (see Table 8a). Per that document, 375,038 out of 598,781 (63%) fell into that category in CY 2013+2014.
And therein lies the answer: The CEOs shouting about STEM shortages are likely running software companies. Being in that field myself, I can confirm that my own (and my peers') incomes have not stagnated**. I don't want to get into the numbers too exactly, but our typical wage growth is well above COLA. Table 10 shows the 25th (64,000), mean (84,000), median (75,000) and 75th (96,000) percentile compensation ranges as well. Because the median is a lot lower than the mean, there must be a tight cluster on the low end and a longer range on the high end (meaning probably 20% or so are above 100k per year). If they're getting lowballed at these wages, those are still well above the American median and us software engineers should only complain so much, right?
If the article focused on the niche within STEM that actually consumes the bulk of the H1B visas, it may have come to a very different conclusion. Software development is singularly poised to be most hurt by limits to the H1B program.
The comments section is rampant with people complaining about how Americans aren't even given a shot. This is blatantly false in our industry. One guy makes the correct point: while there is a never-ending stream of applicants, few meet a quality bar. What is that bar? That they be able to work autonomously, solve problems and create reliable software. Others jumped on him, that perhaps his bar was unrealistic. I suppose those same people should just live with paying the same for an inferior product (insert joke about XFinity cable boxes). I can confirm from my time giving interviews that people I really want to work with are not appearing left and right. For anyone wanting an analogy, imagine someone shows up for a trucking job and doesn't even know what a clutch or a steering wheel is. That level of interviewee is all too common, and that lack of quality is common across domestic and foreign applicants. It's not like being foreign makes someone a magical unicorn candidate. They have to be good too.
** - This does depend on employer, but most large firms (ie: "the ones you've heard of") are in constant competition with each other, constantly wanting to hire away each others' talent. Why? Because they'd rather pay extra for someone who's proven they can cut it than try to wade through hundreds more resumes for an unknown.
A big talking point around H1B is whether we need them at all. Many complain there's a STEM shortage, while others point to articles like this that "prove" there isn't. I'd like to address a few things within.
As the article rightly points out, a lack of definitive data or even definition of what STEM is makes this a potentially difficult web to untangle. The author concludes there can't be a STEM shortage because:
1. There are more STEM graduates than there are open STEM positions
2. If there truly were a shortage, wages wouldn't stagnate. Apparently they have?
This is a gross oversimplification and fails to acknowledge how H1B fits into the STEM fields and that not all graduates are qualified to do these jobs (consider those barely passing or going through a for-profit diploma mill). Let's look at some relevant numbers, because the article above left lots to be desired.
There is a cap of 65,000 H1B each year, and 20,000 more for those with advanced degrees (85,000 total). Compare that to a total near 160M across the USA. Each year, we're at most giving an H1B to 1 in 1870 workers. Granted, these are "good jobs" and probably matter more economically, but it's still a fraction of a tenth of a percent. It would literally not move the unemployment numbers as reported to a tenth of a percent.
The literal majority of H1B go to people in the computer-related occupations (see Table 8a). Per that document, 375,038 out of 598,781 (63%) fell into that category in CY 2013+2014.
And therein lies the answer: The CEOs shouting about STEM shortages are likely running software companies. Being in that field myself, I can confirm that my own (and my peers') incomes have not stagnated**. I don't want to get into the numbers too exactly, but our typical wage growth is well above COLA. Table 10 shows the 25th (64,000), mean (84,000), median (75,000) and 75th (96,000) percentile compensation ranges as well. Because the median is a lot lower than the mean, there must be a tight cluster on the low end and a longer range on the high end (meaning probably 20% or so are above 100k per year). If they're getting lowballed at these wages, those are still well above the American median and us software engineers should only complain so much, right?
If the article focused on the niche within STEM that actually consumes the bulk of the H1B visas, it may have come to a very different conclusion. Software development is singularly poised to be most hurt by limits to the H1B program.
The comments section is rampant with people complaining about how Americans aren't even given a shot. This is blatantly false in our industry. One guy makes the correct point: while there is a never-ending stream of applicants, few meet a quality bar. What is that bar? That they be able to work autonomously, solve problems and create reliable software. Others jumped on him, that perhaps his bar was unrealistic. I suppose those same people should just live with paying the same for an inferior product (insert joke about XFinity cable boxes). I can confirm from my time giving interviews that people I really want to work with are not appearing left and right. For anyone wanting an analogy, imagine someone shows up for a trucking job and doesn't even know what a clutch or a steering wheel is. That level of interviewee is all too common, and that lack of quality is common across domestic and foreign applicants. It's not like being foreign makes someone a magical unicorn candidate. They have to be good too.
** - This does depend on employer, but most large firms (ie: "the ones you've heard of") are in constant competition with each other, constantly wanting to hire away each others' talent. Why? Because they'd rather pay extra for someone who's proven they can cut it than try to wade through hundreds more resumes for an unknown.
No comments:
Post a Comment