What Puerto Rico Tells Us About Global Trade

The United States compiles data about Puerto Rico’s trade with the world.

The largest supplier of imports to Puerto Rico?

Ireland.

The second largest?

Singapore.

Tax trumps gravity, it seems.

Incidentally, Switzerland jumped into third place in the 2016 league table, leaping past other exporters of chemicals and Puerto Rico’s suppliers of fuel oil, diesel, and the like.

It isn’t exactly hard to figure out what is going on here. Puerto Rico’s imports tend to be specialty organic chemicals and pharmaceutical products, and, well, they tend to be supplied from countries that are known to specialize in helping multinationals optimize their global tax bill.

And, setting aside trade with the fifty states for the moment, where are Puerto Rico’s biggest export markets?

Belgium and the Netherlands.

The big ports and distribution centers in northwest Europe.  Europe is almost certainly buying packaged pharmaceuticals—Puerto Rico’s biggest export to the world translates from trade jargon to “medicine, in measured doses, packaged for retail.”

It is a bit too simple to say Puerto Rico imports active pharmaceutical ingredients from low-tax jurisdictions and exports finished pharmaceuticals to high-tax jurisdictions. The imports from Ireland and Singapore could be for pharmaceuticals destined for the U.S. market, and the active ingredients for the finished goods exported to Rotterdam and Antwerp may be coming from the United States.

But the geography of Puerto Rico’s external trade is a good example of why I find the bilateral data useful. It helps identify when trade in tasks starts to look more like trade in tax.*

Puerto Rico also publishes its own trade data—with one release that splits trade with the United States from trade with the world, and another that offers more product detail but aggregates Puerto Rico’s “exports” to both the United States and the world.

This, to be clear, is a different definition of external trade than the U.S. census data uses. Puerto Rico counts the island’s “trade” with the fifty states as external trade. Territories are strange.

That data can help fill out some of the points that Bloomberg’s Justin Fox made in a very good recent column.

The numbers from Puerto Rico’s Government Development Bank (GDB) start in 2007, so the data series starts just after the infamous “936” tax break expired.

But to me they don’t suggest that pharmaceutical manufacturing has left the island in the last ten years (contrary to the received wisdom).

Pharmaceutical exports—counting “exports” to the rest of the United States—have been around $50 billion a year since 2007, and the pharmaceutical trade surplus (exports, net of imports, including chemical imports)** has been around $30 billion.

Puerto Rico’s "exports" of pharmaceuticals to the United States, in dollar terms, almost certainly exceed Ireland’s pharmaceutical exports.***

The real fall off in “exports” since 2009 has come in electronic manufacturing, not in pharmaceuticals. (At least not in nominal pharmaceutical exports).

If the nominal data doesn’t deceive, the fall in jobs in the pharmaceutical sector in Puerto Rico’s pharmaceutical sector has come in no small part from technological improvements and greater efficiency in production.

Though no doubt there also is a more complex story to be told.

In an industry with ongoing productivity gains, the number of jobs supported in the industry tend to fall over time—it typically takes a rise in production for the number of jobs to grow or stay constant. And that hasn’t happened in Puerto Rico, for a host of reasons. While Puerto Rico hasn’t seen an absolute decline in its pharma exports to the world or to the U.S., it has lost out a bit relative to other “offshore” sources of pharmaceuticals in the U.S. market. Puerto Rico is no longer attracting an ever rising share of the industry—it isn’t winning the competition among tax havens in the way it used to—and without that advantage, production has stalled and jobs have fallen.

But the pharmaceutical industry still has a big enough presence to basically drive Puerto Rico’s rather extraordinary balance of payments numbers (this, again, uses the Puerto Rican data that counts trade with the fifty states as external trade).

Income (dividends and the like) payments to off-island non-residents (the multinationals) are huge—resulting in a $30 billion or so annual outflow (for a sense of the relative size of this flow, Puerto Rico’s “domestic” economy—GNP—is now around $65 billion). That income deficit not surprisingly is roughly equal to the trade surplus in pharmaceuticals and chemicals. In other words, transfer pricing shifts profits to Puerto Rico, before dividend payments move the profits to yet another jurisdiction.

Excluding pharmaceuticals, Puerto Rico actually runs a significant trade deficit in goods. And that goods deficit in turn is roughly equal to Puerto Rico’s surplus in tourism and the (large) net transfer Puerto Rico receives from the federal government.

Transfers have gone up in part because of Obama-era policies—ACA funds and the like (see my paper with Greg Makoff). But also because the federal government provides help to poor Americans without jobs—and thus has helped to limit the social impact of Puerto Rico’s ten year slump.

It isn’t a pretty picture.

Puerto Rico is already a tax haven for multinationals and high-income individuals (income earned in Puerto Rico is exempt from federal corporate and income tax: See the Bloomberg explainer). It would have to be even more of a tax haven—not just an offshore jurisdiction but an onshore jurisdiction with a lower tax rate than elsewhere in the U.S.—to gain a further tax advantage. That runs against the desire to broaden the federal tax base.

And the current set of federal transfers haven’t done more than cushion Puerto Rico’s economic slide.

There aren’t any easy answers to Puerto Rico’s long-term economic decline. ****

But on one point I do have conviction: multinational companies, tax-wise, get a better deal in Puerto Rico than in the fifty states; low-wage Puerto Ricans who work on island, by contrast, get a worse deal from the federal government than they would get for work in the fifty states (thanks to the absence of the earned income tax credit). That really needs to change.  

* The geographical breakdown of services trade also is informative. The biggest market for U.S. exports of financial services? The UK Caribbean, which accounts for roughly a third of U.S. financial services exports. The biggest market for exports of research and development services? Ireland, with about a quarter of the total in 2014, followed by Switzerland. Ireland, Switzerland, Bermuda, and the UK Caribbean account for over half of total exports of research and development services (see the BEA interactive tables for trade in services, table 2.2). A big part of services trade is the physical movement of people and goods (travel and shipping). And at least for the U.S., a significant share of the rest looks to be related to the movement of intangible assets to low-tax jurisdictions. There are some very real exports in the services data (architectural services, e.g. skyscraper design and engineering, is one of my favorites). But some service exports look to be driven primarily by trade in tax.

** Puerto Rico’s trade data is based on the NAICS classification, which treats pharmaceuticals as a subcategory of organic chemicals. Puerto Rico provides a breakout though for pharmaceuticals (NAICS 3254, a subcategory of NAICS 325).

*** Puerto Rico’s exports of pharmaceuticals to all off island destinations can be compared to the roughly $15 billion—according to the U.S. data—that Puerto Rico exports to destinations other than the fifty states.  Or sum up the relevant categories in the (extremely detailed) U.S. data on trade with Puerto Rico.  By using the U.S. data on Puerto Rico's pharmaceutical exports to the world rather than Puerto Rico's export data, the New York Times inadvertently understated Puerto Rico's importance to pharmaceutical manufacturing.

**** Exempting Puerto Rico from the Jones Act would help a bit, but on its own in my view it is unlikely to be decisive.