Brad Setser

Follow the Money

Cross border flows, with a bit of macroeconomics

Print Print Cite Cite
Style: MLA APA Chicago Close


The Dollar Trumps Carrier

by Brad Setser
November 30, 2016


Here is a bit of ballpark math.

Carrier has announced it will keep open a factory in Indiana, retaining around 1,000 jobs.

Since the election, the broad dollar has appreciated by about 4%, presumably because of the impact of an expected loosening of fiscal policy. Let’s assume the change in the nominal dollar is equal to the change in the real dollar, and that the rise in the real dollar persists (e.g. it isn’t eroded by inflation differentials)

Using a rule of thumb I learned from the Peterson Institute’s Ted Truman (hence, the accumulated wisdom of the Federal Reserve’s International Staff in the 1990s) a 10 percent move in the dollar changes the U.S. trade balance by about 1 percent of GDP.

So the 4 percent dollar appreciation should raise the U.S. real trade deficit by about 0.4% of GDP.

Nominal U.S. GDP is about $18.65 trillion now (q3 data).

0.4% of GDP is currently about $75 billion.

The Commerce Department estimated that a billion dollar of (goods) exports supports about 5,300 (5,279) jobs in 2015.*

To be conservative, let’s use that number for import competing and service jobs too. That works out to a very rough estimated loss of 390,000 jobs in export and import competing sectors from the stronger dollar (job losses that play out over time, as the exchange rate has an impact with a long lag).

Just saying …

Discussions about trade often tend to focus on trade policy, not the actual trade flows. But trade flows—and the trade deficit—are what matters for total jobs in the tradables sector, and changes in the value of the currency have a big effect on the level of exports and thus the size of the trade deficit.** One that is very hard to counter with sector or firm specific measures.

Trade agreements tend to get the most press, but changes in the dollar tend—at least in my view—to have a bigger impact on trade flows. Trade agreements should raise both imports and exports (or, in some cases, the offshore income of U.S. multinationals—but that is a more complex story). A stronger dollar by contrast raises imports and reduces exports. It encourages multinationals of all stripes to locate less of their global production in the U.S.

And while the value of the dollar fluctuates, a persistent increase in the dollar – say from looser fiscal policy in the U.S. than in its peers – would have a persistent effect on the trade balance.

The numbers I used are arguably conservative.

The IMF – in a large study that explicitly looked at the impact of global supply chains (published as chapter 3 of the fall 2015 WEO; summarized here) – found that a 10% currency move generally changes the trade balance by about 1.5% of GDP.

And one of the models the Fed uses also suggests a 10% dollar move has an impact of around 1.5% of GDP. Joseph Gruber, Andrew McCallum, and Robert Vigfusson summarized the Fed staff’s current trade model; the fall in real GDP they mention comes primarily from a deterioration in the real trade balance.***

“The equations for goods and services exports predict that a 10 percent appreciation of the real dollar would reduce the level of overall real exports by about 7 percent after three years. After three years, the model predicts that real imports would be almost 4 percent higher. In total, therefore, a 10 percent dollar appreciation would reduce the level of real GDP by about 1-1/2 percent relative to its baseline path after three years”

And I could have estimated the impact on goods and services separately, as the Commerce Department estimates that a billion dollars of services exports supports more jobs than a billion dollars of goods exports. 7,000 for services versus 5,300 for goods in 2015 (with the result that a billion dollars of goods and services exports supports close to 6,000 jobs, given the relative shares of goods and services in a billion dollars of exports).

The higher number of jobs supported by a billion dollars of service exports shouldn’t be a surprise. Some services exports derive from high end design and engineering (intellectual property rights show up in the balance of payments either as service export or as offshore investment income), which generate large profits and rents but do not directly support all that many jobs. But the “high-end” image of service exports is a bit deceptive. To be blunt, the typical service exporter is a hotel that caters heavily to international travelers.

Using the higher end estimate of the impact of the dollar and the weighted average for goods and services for the loss of export jobs (and assuming that the jobs impact of imports is similar, e.g. a dollar in lost exports/ higher imports leads to a loss of around 6,000 jobs in the associated sectors) would raise the loss of tradable sector jobs to well over 600,000.

To be clear, the loss of jobs in the tradables sector isn’t the loss of jobs in the economy overall, not when the economy is operating at full employment. A fiscal expansion that leads to a monetary tightening that pushes the dollar up will generate additional jobs in the non-traded parts of the economy.

But the work of David Autor, David Dorn, and Gordon H. Hanson also should by now have generated a consensus that the transition out of manufacturing and tradeables jobs isn’t seemless; the expanding sectors won’t be in the same place, geographically speaking, as the shrinking sectors. (See Tim Duy)

* The Commerce Department found that a dollar of exports supported more jobs in 2015 than in 2014—as employment in the export and export supported sector went up even though exports didn’t. In 2014, a billion dollars of goods exports “only” supported 5,000 jobs. So much for always rising productivity in the export sector. I would emphasize the numbers are estimates … the results are plausible, but there is an error band.
** China is an exception, as its peg and then managed currency have been a policy focus. Some trade policy types though argue that too much emphasis is placed on the currency relative to China’s other commercial policies that impede U.S. firms.
*** The Fed’s model has predicted the impact of the dollars 2014/15 rise on exports relatively well. Imports have increased by less than the model implies, for reasons that have been discussed previously (inventories, the impact of the fall in investment and the like—though some of the gap remains a mystery).


  • Posted by SAH

    Great piece, Brad.

    In your experience, how stable has the relationship between the real exchange rate and the trade deficit been in the US?

    To what extent do you think other factors, specifically domestic factors/policies in high surplus countries that contribute to high national savings, can increase the propensity for US fiscal spending to result in a larger US trade deficit?

  • Posted by Pierre

    Brad, aren’t you off by a factor of 4? That 18 trillion GDP is at an annual rate…
    Other way to frame it: quarterly gross job losses from the Business Employment Dynamics (BLS) are roughly 6.5 million. Funnily enough, job gains are about the same. That’s the churn in the labor market. Translates into 500K gross losses (and gains) per week. Looks like Trump’s little finger in the dike didn’t do much…

  • Posted by Brad Setser

    Pierre – $18.65 trillion is annualized, but that is what i needed for the calculation. Estimated full year 16 GDP will be similar. I am not a fan of the churn argument (e.g. saying the loss of a factory doesn’t matter, look at how small it is next to the natural churn in the economy) in this context. For small towns, the loss of a major employer leads to one way churn (jobs lost, no jobs created) not the usual churn among retail, fast food jobs. The autor,dorn and hanson results suggest that the impact of china on the labor market was much larger than the “it is just another form of churn” arguments imply,

  • Posted by Brad Setser

    SAH. Dollar/ trade relationship has generally been stable. The strongest element is the estimated fall in exports from the strong dollar. That has been very very consistent. Saw it strongly from 98 to 02. And — as the fed paper notes — the recent fall in exports is consistent with the model. There is some evidence that the rise in exports from 03 to 07 was a bit smaller that might have been predicted; i will have more on that later. The export effect generally is stronger than the import effect. On the import side, recently imports have been weaker than implied by the model (as noted). In general though the trade/ dollar relationship is very robust. you cannot get a good fit with the trade balance without it.

  • Posted by Brad Setser

    SAH as for your second argument, high national savings leads to weak internal demand (absent really high investment), looser monetary policy, and generally a weaker exchange rate — so the discrepancy in savings influences trade in part through the exchange rate. This is part of Krugman’s argument that there is no “immaculate” transfer — changes in the trade balance do not happen simply through changes in savings and investment, with no change in relative prices. I do though think that the rise in the US deficit from a fiscal relaxation would be much smaller if other, high savings US trading partners were also relaxing fiscal policy. Their fiscal expansions would spillover into demand for US exports, and, more importantly, there fiscal policies would change expectations about monetary policy and thus support their currencies.

  • Posted by John Green

    Rookie question that this article brings up: Why would looser fiscal policy STRENGTHEN the dollar – I would think that much like looser monetary policy, it would weaken the dollar (looser fiscal policy usually is associated with higher debt).

    Thanks in advance!

  • Posted by Brad Setser

    John — a looser monetary policy by definition lowers interest rates, and thus the return on holding dollars. A looser fiscal policy tends to raise interest rates (directly, because the treasury issues more, and indirectly, because it stimulates the economy and that leads the central bank to tighten monetary policy). Higher interest rates increase the return on holding dollar assets. So in a conventional neo-Keynesian macro model, fiscal expansion strengthens the currency.

    the overarching assumption is that the government remains solvent, and the market is willing to finance the government for a higher interest rate. for the US and most other advanced economies that issue debt in their own currency, i think that is the right assumption

    hope this helps

  • Posted by ellen1910

    Thanks for your response to John’s question. I, too, was confused. I would have guessed that expansionary fiscal policy would lead to increased imports and a lower dollar.

    Are you saying that the FRB’s unwillingness to buy the new government debt and its tendency to raise interest rates at the hint of economic expansion overwhelms the effects I thought would obtain?

    Or am I just wrong ab initio?

  • Posted by David

    Having a hard time with your logic. QE ended late 2014 and the dollar went up 20% in short order and has bounced around that area since. Tighten money supply and it went up, now because Trump was elected and his promise of massive spending to loosen fiscal policy is making it go up. Can’t have it both ways. The US data and presumptive interest rate hikes that are coming give demand to the dollar as it’s the only game in town right now. Europe and Japan continue to dump money into their economic systems thus devaluing their currencies, which adds to the unintended consequences of a stronger dollar.

  • Posted by Brad Setser

    Ellen — your intuition is half right. fiscal expansion leads to more imports. but because it (in a standard model) pushes up interest rates, is also leads to more financial inflows, and a stronger dollar. And yes, the FRB by raising rates plays an important role in this process. if there is a fiscal expansion at a point in time when the fed is at the zero bound (eg it wants to reduce rates but cannot) then the fed would not respond to the fiscal stimulus and there would much less of an impact on the currency.

    David. I see 20214 and now as consistent with the same basic factor — the dollar rises on expectations of tighter us monetary policy (when others are loosening). in 2014 there was a sense that the us economy was now strong enough for the US to raise rates, while EA/ Japan were for different reasons easing more. Now there is an expectation that a “late cycle” fiscal expansion will add to us demand and the fed will respond with higher interest rates. so both 14 and 16 dollar appreciations driven by expected fed tightening (on going easing elsewhere). That at least is how i see it