How we calculate US job losses from currency management

May 11th, 2015 at 12:02 pm

[This post was written jointly with Ben Spielberg]

In this PostEverything piece, Bernstein and Baker quote estimates of jobs lost in the U.S. due to higher trade deficits induced by currency management. In this memo, we explain how we derived the numbers cited in that piece.

The analysis is “back-of-the-envelope” and is thus a rough estimate of the job impacts of currency management, but it relies on careful research and provides a useful context for evaluating our claim that the failure to address the currency problem in the Trans-Pacific Partnership trade agreement is a significant omission. This point is increasingly germane given the arguments from proponents of the agreement, up to the President himself, that opponents simply fail to understand the way the TPP will work and how much it will help working people.

Fred Bergsten and Joseph E. Gagnon from the Peterson Institute for International Economics have done some of the most important work on this question and our approach, similar to that in various papers by other economists, begins with their key insight: when our competitors use their trade surpluses to buy dollars, thus depressing the value of their currencies relative to the dollar, that raises the US trade deficit. Our jobs estimates, like those of others we cite below, are driven by this exchange rate mechanism.

Figure 1, pasted in from a Joe Gagnon paper, shows a striking correlation between trade surpluses and the purchase of foreign exchange reserves by countries with large holdings of foreign assets. Gagnon’s statistical analysis formalizes this relationship, indicating that the buying and selling of foreign assets, most frequently the US dollar, has a large influence on trade surpluses and deficits.

Gagnon Fig1

In another paper, Bergsten and Gagnon’s regressions suggest that 70 to 90 cents of every dollar spent on foreign reserve currency shows up in trade balances – in surpluses for the countries doing the purchasing and in deficits for the countries supplying the reserve currency.

Using data from the World Bank, we first determine the average annual increase between 2008 and 2013 in a country’s holdings of foreign reserves. We multiply each country’s annual average increase in foreign reserves by 62 percent to obtain the approximate share of that increase due to purchases of US dollars, as opposed to other reserves (according to currency composition of foreign exchange reserve or COFER data from the International Monetary Fund, the US dollar made up about 62 percent of the world’s foreign reserve currencies between the second quarter of 2011 and the end of 2014).  We then multiply the resultant numbers by 70 percent for a conservative estimate and by 90 percent for a higher-end estimate of the effect on the US trade deficit, per Bergsten and Gagnon’s results.

Finally, to get our jobs numbers, we convert these values to percentages of GDP using the five-year average of US GDP between 2008 and 2013.  The final job figures are derived using Robert Scott’s estimate (based on previous work by Josh Bivens) that a 1 percent increase in the U.S. trade deficit as a share of GDP is equivalent to the loss of 1.2 million jobs.

For example, China’s average annual increase in total foreign reserve holdings between 2008 and 2013 was $383 billion.  By multiplying by .62, we estimate that $237 billion of those holdings were in US dollars.  We then multiply $237 billion by .7 to get a low-end estimate of $165 billion of impact on the US trade deficit, and by .9 to get our high-end estimate of $212 billion.  These numbers are between 1.06% and 1.36% of the annual average of GDP between 2008 and 2013 (approximately $15.5 trillion).  Finally, we multiply by 1.2 million per percentage point of GDP to obtain our final estimate that currency management by China costs between 1.3 and 1.6 million jobs annually.

Of the countries involved in TPP negotiations, Bergsten and Gagnon have identified three currency managers: Japan, Malaysia, and Singapore.  As shown below, purchases of foreign reserves by these three countries is estimated to have cost the US between 250,000 and 320,000 jobs annually between 2008 and 2013.

According to the White House, however, the TPP “is explicitly designed to allow others to join in the future.”  It is therefore appropriate to consider potential job loss from currency management in all twenty countries Bergsten and Gagnon identify as past managers.  Our estimates suggest that the potential consequences of failing to address currency management could range from 2.5 to 3 million jobs lost annually.

Currency ManagerTable

The numbers above are sensitive to a variety of assumptions.  However, there are three reasons to believe the estimates are conservative.  First, by solely focusing on new purchases of foreign reserve currencies, we ignore the fact that purchases countries have already made exert an influence on the US trade deficit.  China’s approximately $4 trillion in foreign reserves, even if held constant, would likely impact our trade deficit, but our analysis does not account for this fact.  Bergsten/Gagnon write that “China’s heavy past intervention has lingering effects on the level of the exchange rate, keeping it considerably lower than it would otherwise be.” Our method ignores this effect.

Second, Scott in his analysis multiplies the projected change in the trade deficit by 1.6, a multiplier that has been used for infrastructure spending, to estimate its impact on GDP.  We, however, use no multiplier.

Our estimates are within similar ranges of past work. Bergsten/Gagnon argue that currency manipulation costs the US 1-5 million jobs. Ken Austin, in an important recent paper in the Journal of Post-Keynesian Economics, estimates the cost of dollar reserve accumulation at 6 million jobs (note: Austin includes the third effect that we ignore: when a dollar accumulator runs a bilateral trade surplus that doesn’t include the U.S., that too will strengthen the dollar and increase our trade deficit).

Again, these calculations provide very rough estimates.  But they certainly provide evidence, even outside of many other legitimate concerns about leaked contents of the TPP, that the agreement is not nearly as benign from the perspective of American workers as its proponents claim.

[We thank EPI’s Josh Bivens for helping us with this work; any mistakes are our own.]

Print Friendly, PDF & Email

4 comments in reply to "How we calculate US job losses from currency management"

  1. Kevin Rica says:

    A technical note: The Austin calculation of 6 million lost US jobs was for 2008 when “net U.S. reserve sales peaked at $550 billion or 3.9 percent of U.S. GDP.” The current rate of US job losses due to reserve sales should be much lower.

  2. Smith says:

    Speaking of trade deficits, there is something not considered which is very important.

    China accumulates dollars in order to goose exports which enrich elites at the expense of ordinary workers. Because money is not spent either by Chinese elites or Chinese workers on U.S. products, it accumulates. Even ignoring the effect of the extra income gained through unnaturally augmented export levels, not buying products means what? It means the money will probably be spent buying American businesses. Even if all currency manipulation ended immediately, the already enormous accumulation of dollars can realistically only result in this outcome. Unlike Russian oligarchs it won’t be just Manhattan real estate, penthouse apartments. Unlike Japanese capitalists in the 1980s, it won’t be just Manhattan real estate, marquee buildings.

    If you don’t think this is significant, then you have never worked for a company that was owned by another country (a business private or publicly held). Also China is not Britain, Germany, nor even Japan.

  3. purple says:

    The plan is to eventually include or force China to join TPP, once everyone around them does. So China should be considered a possible TPP country.