Scott Sumner, Trade Deficits, and Dark Matter
Scott Sumner at EconLog pounced at the red meat I waved in front of him. (I had sent him Trump’s ludicrous tweet on trade wars.) One of Scott’s points was very interesting, and I want to make sure readers understand the claim.
Specifically, Scott wrote, “4. Most people assume that the US runs a persistent trade deficit. But if the trade deficit really were persistently negative, properly measured, then this time series should also be negative and falling. Instead, the US surplus on investment income is strongly positive and rapidly increasing. If this is what it means to be a “debtor nation”, then lets have lots more debt!”
He then presented a chart, but I’ll give a more comprehensive one below:
(So what Scott showed was more-or-less the green line in my chart above, although it’s not exact–perhaps his wasn’t seasonally adjusted?)
Another clarification: I think technically Scott should’ve been talking about the current account deficit, not the trade deficit.
Anyway, the basic idea is this: Over time, the flow of annual income that Americans earn from their foreign assets held abroad, keeps growing more than the flow of annual payments that Americans must make to foreigners on their U.S.-based (but foreign-owned) assets. This is not what you might have expected, since the U.S. has been running large and persistent current account deficits for decades. A current account deficit means that the amount Americans earn on their foreign assets PLUS their sales of goods and services to foreigners (i.e. exports), is less than the amount Americans spend on imports and on making payments to foreigners for their US-held assets.
So what Scott is saying is that there must be something screwy with how we measure current account (sic) deficits.
This is the “dark matter” thesis of Hausmann and Sturzenegger, from 2005. They argued that superior US know-how explains why American investors can earn higher returns on Foreign Direct Investment abroad, than foreigners can earn abroad. So effectively the foreign savers invest in the US at a low yield, and Americans then invest abroad at a higher yield, earning the spread. That seems fine when hedge funds or commercial banks do it, but in terms of the conventional trade accounting it looks “unsustainable.” Here’s an analogy I made back in 2007 (in an essay in which I made some unfortunate and snarky predictions at the end):
Suppose a stock speculator spies a company that is undervalued. He invests $100,000 in the company, waits one week, then sells the shares for $105,000. Then he uses the profit to buy his wife a necklace.
His wife…is quite furious. “How much did you earn selling your labor or merchandise last week?!” she demands. He thinks for a moment and realizes the answer is “None.”
The speculator’s wife continues with the interrogation: “And how much did you spend on assets this week?” The husband replies that he spent $100,000 on stock shares.
She then demands, “And how much did outsiders spend on assets that you owned during this same time frame?” After a pause—for he’s not used to thinking in these terms—the husband informs her that outsiders invested a total of $105,000 in assets that he owned.
The wife starts to tremble with rage. “Do you mean to tell me that you financed this $5,000 bauble purely by increasing your liabilities to people outside of this household??”
So to be clear, I’m not now saying, “Sumner is right, it’s dark matter, everything is fine.” I’m just clarifying the argument.
For example, I could flip things by imagining a guy telling his wife, “I’m a smart investor, babe. I borrowed $10,000 from credit cards at an introductory APR of 1.99%, and I invested $9,000 of it into 30-year bonds yielding 3%. Then I spent the remaining $1,000 getting you some new shoes and myself a new sport coat. Some people call this ‘living beyond my means’ but I call it leveraging dark matter. I mean, I’m only paying $199 in finance charges, while I’m earning $270 in interest on my Treasuries. It’s all under control.”
“So what Scott showed was more-or-less the green line in my chart above, although it’s not exact–perhaps his wasn’t seasonally adjusted?”
His is annual, yours is quarterly. You can duplicate his graph by taking the series BOPXRA and BOPMIA and adding them together
Thanks Andrew. Well, I *did* add the two series together to get the green line. So the difference must be the annual/quarterly distinction.
Yeah, neat thing about FRED’s series ID numbers-that “A” on the end denotes that those series are the annual counterparts to the series you used (BOPXR and BOPMI)
As another note, if you wanted to turn your series into his you could also click under “modify frequency” and then select “annual”-in this particular case you’ll want to choose under aggregation method “sum” because for some reason your series given quarterly totally not annualized total for the quarter (i.e. how GDP is reported, as a “seasonally adjusted annual rate”) Or you could multiple your series by 4.