I just returned from cruising the Caribbean on Anthem of the SeasMaybe you’ve heard of it? Fortunately, no coronavirus panic marred our vacation, and the concluding scare at the dock turned out to be a false alarm.  Though I’d seen a little of the Caribbean before, this trip was a heavy dose: after a stop at San Juan, Puerto Rico, we sailed on to St. Maarten, Antigua, St. Lucia, and St. Kitts.  My social science reflections:

1. I’ve been writing about Puerto Rico for years.  Its great open borders experiment doesn’t just put an upper bound on the dangers of “brain drain”; it also shows that liberalization makes migration serenely “snowball” rather than frighteningly “flood.” I was excited, then, to finally see Puerto Rico with my own eyes.  San Juan, at least, met my high expectations.  It’s a rich and beautiful city – and I didn’t merely visit the tourist areas.  The country is packed with Walmarts and other big-box retail.  Uber works well.  I saw near-zero remnants of Hurricane Maria, and – through my bilingual sons – had two fun chats with jovial Uber drivers.  Next time, I’d really like to drive the Ruta Panorámica.  No doubt I’d witness some serious poverty, but I doubt it’s more disturbing than driving around rural Mississippi or West Virginia.

When I invoke Puerto Rico, immigration skeptics often scoff.  Hurricane Maria aside, they deem Puerto Rico a disaster zone.  After all, if it were a U.S. state, Puerto Rico (with a per-capita GDP of just $32,000) would be even poorer than Mississippi, right?

I call this a willfully misanthropic comparison.  To grasp the effects of open borders on Puerto Rico, you have to ask, “How would Puerto Ricans be doing if they didn’t enjoy free migration to the U.S.?”  To answer that question, you’ve got to look not as Mississippi, but at other Caribbean islands.  Which I then proceeded to do.

2. Our next stop was St. Maarten (the Dutch side).  Official estimates of its per-capita GDP vary very widely.  While the CIA absurdly sets it at $66,800 (PPP), Wikipedia provides only a 2008 estimate of $15,400 (nominal).  Compared to Puerto Rico, in any case, the island looks quite poor.  A good chunk of this admittedly stems from low population; 42,000 people plus tourists aren’t enough to sustain more than a few restaurants or entertainment venues.  Even taking population into account, however, living standards look low.  Desperate peddling of tourist wares is a common job.  Roads are bumpy.  The shiniest business we saw was a KFC.  The biggest grocery store we found wasn’t bad, but about a third of the refrigerated shelves were empty.  There’s no Uber, but since there are plenty of taxis, I blame regulation.  After SNUBA diving (awesome), we taxied to the local tropical zoo, which sadly turned out to be shuttered since Maria.

What explains the gap between the official economic statistics and what we saw?  The simplest story is that a few super-rich expats drive up the average, but it’s hard to believe that’s close to the whole story.  The next explanation is that I’m such a spoiled American that almost everywhere on Earth looks impoverished to me.  Another is that the statistics are fake; but wouldn’t countries want to overstate their poverty to get extra foreign aid?  Last, CPI bias is plausibly astronomically unfavorable in small islands where there’s not much to spend your money on.  (As I told Tyler, there are odd parallels between small Swiss towns and these Caribbean islands; in both places, even the rich have little to buy).

While we’re on the subject of CPI bias, the Internet has clearly been a nearly unmeasured godsend for the whole region.  In 1990, islanders would have been cut off from 99% of humanity’s cultural bounty.  Today, the curious can sample and savor this bounty for modest connection fees.

3. Then we sailed on to Antigua (a subset of Antigua and Barbuda), with recent per capita GDP estimates of $17,500 nominal and $28,000 PPP.  It did indeed look a little richer than St. Maarten, though that too could be confounded with higher population.  The downtown was fun to see, but the roads were bumpy and even the main sidewalks poorly maintained.  While shuttling to snorkeling, we saw a huge sports stadium (10,000 seats!) largely funded by the government of mainland China.  (Other islands, in contrast, seemed oriented toward Taiwan).  There were fewer desperate peddlers, but almost no businesses even in the historic downtown.  As Richard Scarry famously inquired, “What do people do all day?”

4. Next, we saw St. Lucia.  Geographically, it was the most beautiful of the islands.  The Pitons are splendid, and we passed some scenic harbors and resorts.  Economically, though, St. Lucia looked the worst.  This fits with official statistics, which put its per-capita GDP at $10,000 nominal and $15,000 PPP.  Even though it has roughly twice the population of Antigua and Barbuda, the KFC was again the shiniest business we saw.  The main downtown church was closed, and the nearby park contained about a dozen apparently homeless men, though perhaps they were just relaxing and drinking alone.  Desperate peddling was intense.  The local police seemed to be one of the main employers.

5. Our last stop was on St. Kitts (a subset of St. Kitts and Nevis), whose per capital GDP of $19,000 nominal and $31,000 PPP make it the richest island we saw after Puerto Rico.  Since we spent six hours hiking Mount Liamuiga, the local volcano, we never walked the town.  Yet we did get to see a long stretch of one of the main coastal highways, and the country did indeed look marginally richer than Antigua.

My hiking guide described himself as “fascinated by economics,” and we had a good amount of time to chat.  He suffered from severe pessimistic bias; I tried in vain to calm his fears that U.S. agro-business faced imminent crisis.  When he playfully accused me of having naive faith in mankind, I told him, “No, I just believe in business.”  He mentioned his Netflix subscription, but I didn’t have time to rhetorically build on that foundation.

My guide knew a handful of islanders who worked in the UK, but viewed his countrymen as deeply provincial.  Cruise ships dock all the time in St. Kitts, but when I asked him if he knew anyone who worked on such a ship, he insisted, “It’s not something they would ever think about as a possibility.”  This surprised me, because workers of Caribbean origin were fairly common on my ship, especially relative to their countries’ populations.

6. Are the latter four islands the ideal comparison group for Puerto Rico?  Not really.  Antigua, St. Lucia, and St. Kitts are all former British colonies, and St. Maarten’s is so anglophone that it might as well be.  A better approach is to compare Puerto Rico with other former Spanish colonies; the Dominican Republic is the most obvious counterpart.  Since the latter’s per-capita GDP is only $9000, Puerto Rico’s open borders experiment look even better.

7. I’ve repeatedly heard people claim that open borders would turn the U.S. into Haiti.  On this journey, I was struck by the fact that almost nowhere in the Caribbean is remotely as awful as Haiti.  St. Maarten, Antigua, St. Lucia, and St. Kitts all have roughly the same demographics has Haiti – all are 90%+ black.   They all have roughly the same history of hellish slave plantations.  Furthermore, according to the least-bad estimates, their national average IQs are all extremely low.  St. Lucia comes in second-to-last on Earth, with an average national IQ of 62 (versus Haiti’s 67).  Despite these parallels, St. Lucia roughly matches average global per-capita income, and St. Maarten, Antigua, and and St. Kitts are comfortably above this average.

8. With the able assistance of Nathaniel Bechhofer, I’ve pointed out that “Deep Roots” theories of national development are highly sensitive to population-weighting.  If you count China and India as two data points, the empirics say that national ancestry matters a lot.  When you weigh countries by their populations, however, national ancestry barely matters at all, because the two most-populous countries on Earth have done poorly in modern times despite their illustrious histories.  Critics have pushed back; each country should count as a separate “experiment,” so we should base our worldviews on the unweighted results.

Yet in that case, each and every tiny upper-middle-income Caribbean country should statistically count as much as China and India.  I just checked Putterman and Weil‘s data, and found that none of my last four islands is actually in their sample.  (Haiti and Jamaica are, but even the Bahamas fell through the cracks).  If we re-did Deep Roots estimates with ten more Caribbean data points, I predict that their results would markedly attenuate.  So would Garett’s main findings in Hive Mind.

Personally, I continue to think that population-weighting is the way to go in cross-country regressions.  If you disagree, though, you’d really better add the island nations of the Caribbean to your sample and watch what happens.