I’m not done thinking about trade policy, especially after getting lots of interesting feedback from my NYT oped earlier this week. I’d put the responses into the following bins:
a) I agree! My family/community and I have been seriously hurt by trade and it’s about time the pundits and politicians started to get that.
b) I hope you’re right (that we won’t go back to fighting about FTAs anytime soon), but don’t be fooled by candidates who fake left and go right on this.
c) Stop complaining. Manufacturing jobs aren’t coming back and you and the other “protectionists” need to get over it.
d) You’re an idiot.
I won’t comment on ‘d.’ One hears that regularly in comment sections, and I can’t speak to its veracity, though I won’t say it’s wrong.
On ‘a,’ which was by far the most common response, there seems to be an outpouring of relief by people who feel that the reality of trade as a double-edged sword is finally getting elevated by the press, the punditry, and the candidates. Ed Porter’s piece today on this point (“On Trade, Angry Voters Have a Point”) is thoughtful and instructive, and Dean Baker adds important points as well, citing numerous aspects of the imbalanced trade problem, many of which I left out.
On ‘b,’ I’ve tried to hedge my language, because that of course may be correct.
On ‘c,’ we have a problem. First, as in Dan Drezner’s piece, there’s the silly “protectionist” name-calling. Clearly, my piece stressed not only the benefits of trade (along with the costs) but also that trade and globalization will proceed apace. It will do so, I predict, without FTAs. I could be wrong about that last part (see ‘b’) but not about the first part. I consider that–expanded trade–a positive development, both for the increased supply of goods here and the growth opportunities of emerging economies.
I also explicitly agree in the oped re lost jobs not coming back. Where I differ from point ‘c’ people is that I’m not convinced the sector can’t add jobs, perhaps in significant numbers. So let’s look a bit at the critics’ arguments and where they go wrong.
Myth 1: We’ve been losing manufacturing jobs for many decades, and that’s not going to change.
Actually, as the figure below shows, we lost a lot of manufacturing jobs in only one decade—the 2000s. It was in this period, after China joined the WTO and our trade deficit with them grew sharply (see Figure 1 here). In fact, our trade deficit as a share of GDP in the mid-2000s of about -6% was the largest on record. Deficits of this magnitude cause not only a severe drag on growth, but since they must be offset if we’re to avoid recession, they’ve contributed to damaging bubbles in recent years (it’s all here in Chapter 5).
Other than that, manufacturing employment was remarkably steady, holding at around 17-18 million from the 1970s to 2000. Since employment grew over this period, these jobs fell as a share of the total over that period, as is the case in every advanced economy, as expanded trade meant that more domestic demand for manufactured goods would be met through imports, especially lower-end goods like textiles.
But manufacturing jobs don’t always decline–including in recent years, as you see at the end of the figure.
Myth 2: We can’t add jobs in manufacturing because the sector is so productive.
This is the idea that even if we can expand manufacturing output, it won’t lead to jobs, because robotics are replacing workers in factories across the land. That’s wrong on two counts, the latter of which is most important.
First, the robot story implies an acceleration of manufacturing productivity (more output with fewer workers equals faster productivity growth), for which there is no evidence. The next figure plots manufacturing productivity growth (real output per hour), with a smooth trend through the bips and bops. It’s a noisy series, but there’s no trend there (I wouldn’t make too much of the dip at the end).
Second, since when was productivity a job killer? Economy-wide, both productivity and jobs have grown considerably over the years, with the intervening variable being demand, of course. In fact, the correlation between the annual percent change in manufacturing productivity and jobs in the sector is small but positive, and it grows over time (e.g., r=0.17 over full period, 1950-2015; r=0.30, since 1989 and 0.36 since 2000).
Our productivity edge is how we (or Germany, or other advanced economies) are able to compete globally despite our relatively high labor costs. But we can’t compete if the currency playing field is uneven, or when trading partners build up a “savings glut” that ends up importing consumer and labor demand from us (or, in Europe, from periphery countries to Germany, which is running trade surpluses of 8% of GDP). This is the trade deficit problem which I fret about and many of my critics fail to address.
Here are some results from a toy model. I definitely wouldn’t read much into this—the R-sq is a measly 0.17. But surely the “manufacturing productivity kills jobs” folks would expect a negative coefficient if we regressed manufacturing employment on productivity growth in the sector.
Not so, as the output below shows (see data note for details). Both productivity and net exports (both lagged) as a share of GDP load positively.
Like I said, it’s not much of a model, but what happens if we set net exports/GDP to zero starting in 1979 (which was about its average from 1950-79) instead of letting them go to the big negative place? The results are in the next figure. First, note that the baseline kinda roughly follows the actual levels of manufacturing employment (MAN). But once we disallow the trade deficit from going south, employment keeps growing, even as productivity growth continues apace.
I don’t really think we’d have 24 million factory jobs now instead of about 12 million had the trade deficit not grown. The factors driving manufacturing employment growth of course involve more than productivity and net exports. But I humbly submit that this humble model is telling us something important about the way forward, which is a) if we want to help our factory sector, we need to pursue more balanced trade, and b) productivity is not the problem.
My hope, as expressed in the oped, is that by getting the cards on the table about the real impacts of trade while taking a long break from divisive FTA fights, we may be making some progress towards a smarter trade debate. It would help if people didn’t label you protectionist the minute you start thinking about the problem.
Data note: The BLS quarterly manufacturing productivity series only goes back to 1987, so before then I used the annual manufacturing productivity data from the BLS international series. Net exports/GDP are from BEA/NIPA. Manufacturing employment is from the establishment survey. Manufacturing employment and productivity are entered as log growth rates.