Monday, April 25, 2016

Blinder on Trade

Alan Blinder has an excellent op-ed in the WSJ on trade. It's hard to excerpt as every bit is good.
1. Most job losses are not due to international trade. Every month roughly five million new jobs are created in the U.S. and almost that many are destroyed, leaving a small net increment. International trade accounts for only a minor share of that staggering job churn. ...

2. Trade is more about efficiency—and hence wages—than about the number of jobs. You probably don’t sew your own clothes or grow your own food. Instead, you buy these things from others, using the wages you earn doing something you do better.  ...
3. Bilateral trade imbalances are inevitable and mostly uninteresting. Each month I run a trade deficit with Public Service Electric & Gas. They sell me gas and electricity; I sell them nothing....

4. Running an overall trade deficit does not make us “losers.”...

5. Trade agreements barely affect a nation’s trade balance. ..a nation’s overall trade balance is determined by its domestic decisions, not by trade deals... America’s chronic trade deficits stem from the dollar’s international role and from Americans’ decisions not to save much, not from trade deals. Trade deficits are not a major cause of either job losses or job gains. makes American workers more productive and, presumably, better paid.
One could say much more. Trade is not a "competition," for example. But,  having done this sort of thing, I'm sure lots of other good bits are on the cutting room floor.

Alan is more sympathetic to government "help" to trade losers, which I agree sounds nice if it were run by the benevolent and omniscient transfer payment planner, but I think works out poorly in practice when we look at the success or failure of actual trade adjustment programs. But that is a small nitpick.

Alan closes by wishing that Bernie Sanders and Donald Trump understood these simple facts a bit better. I think his list of politicians needing enlightenment could be a little longer. But he's courageous enough for speaking the kind of heretical truth that will come back to haunt him should he ever want a government job.


  1. Dr. Blinder is confused. An economist of his stature should know better.

    GDP = C + I + G +(Ex - Im). I will bet they still teach that in Macro classes at Princeton.

    Trade deficits do indeed result in job losses in the USA. And domestic multiplier effects make trade deficit job losses even worse.

    1. I'm not sure I get what you're arguing. The equation is correct but why do you assume that means a trade deficit causes job losses?

    2. And that is just a static analysis. On a dynamic basis those goods that are purchased as part of a trade deficit depreciate over time. Meanwhile the debt incurred paying for the imbalance does not.

    3. Just because you put a variable on the right hand side of an identity does not mean that it causes changes in the other variables. You might as well write

      Ex-Im = GDP - C - I - G

      to conclude that more government spending reduces exports.

    4. Good grief,

      You do realize that an increase in G will increase GDP (Nominal anyway) by an equal amount?

    5. Always and everywhere? By 1.000? The equation provided allows that an increase in G will decrease I by an equal amount instead. How do you know it goes your way?

      The point: One equation, an identity, does not determine which item causes which item. And just putting your favorite one on the right hand side doesn't do that either.

    6. Fair point.

      The trade deficit could fall and personal consumption / government expenditures / or investment could fall right with it - no net effect on GDP.

      Likewise an increase / decrease in government expenditures could have no effect on trade balance if it is offset by an opposite change in either consumption or investment.

      And so the trick is to increase one variable (increased personal consumption, increased government expenditures, increased investment) without decreasing the others - we can agree on that?

      I don't agree that bilateral trade imbalances on the scale of countries are inevitable.

    7. Was it necessary to add 'all other components remaining the same'?

      An increase in Im (or decrease in Ex) -- ALL OTHER COMPONENTS REMAINING THE SAME -- will cause a decline in GDP. A decline in GDP (for the USA) will result in job losses in the USA. Job losses in export-sensitive industries in the USA will have multiplier effects causing more job losses through other sectors of the economy.

      Does that make it more clear to everyone??

    8. Ed,

      I think it was. We are talking about nominal GDP here and so any decrease in trade deficit could be offset by a decrease in government expenditures. No net effect on nominal GDP would be realized.

      However when we consider Real GDP, only the price of domestic consumption (local production and imports that are consumed) is counted towards the domestic price level and inflation rate.

      Real Consumption = ( C - Im ) / ( 1 + PCE Deflator )

      I am not sure how the inflation portion of I (Investment in capital goods), G (Government services), and Ex (Exports) is calculated.

      We can use the general form:

      Real GDP = [ C - Im + I + G + Ex ] / ( 1 + GDP Deflator )

      but aside from C and Im, we don't have individual inflation measures for each of the other terms ( I, G, and Ex).

      My guess would be that an increase in Exports would have a net positive effect on Real GDP even if the net effect on nominal GDP is totally offset somewhere else (decreased G, decreased C, etc.).

  2. I would not contest the bulk of what you have written above, it is pretty much economic orthodoxy, but could you elaborate a bit more on how we should deal with the losers of free trade?

    Given Autor, Dorn and Hanson's recent and excellent paper, there's substantial evidence that these losses, which theory has always predicted to be significant, are heavily concentrated on a small segment of workers, such that they bear significant losses in terms of employment and lifetime income, i.e., as opposed to having these losses be spread out in society.

    You might disagree whether or not these workers 'deserve' (loosely speaking) to be compensated when there's a free trade agreement, although I would argue that they do, as I would find it hard to believe that anyone has the required foresight to anticipate the effects possible future trades would have on their long term employment prospects when looking for jobs, especially taking as a given the question of interregional mobility.

    Even if you do disagree with me on this, and I am honestly not sure if you will, I think that the success of Sanders and Trump, and similar rises of free trade sceptics politicians in the EU, in addition to surveys that show similar increases in scepticism towards free trade in many countries, suggests that politically, future free trade agreements may depend on being able to help out those lose in free trade agreements.

    So, to summarize, given that there are losses from trade agreements, that these losses are significant and heavily concentrated, and, either for 'moral' or for political reasons, there is a necessity to help workers subjected to these losses, what would you recommend be done to help these workers?

    I am especially curious since you express, as I would expect, a great deal of scepticism towards government assistance in this area.

    If I were to guess, I would say that you would likely propose that less regulation and more flexibility in general would help to cope with these changes, and I wouldn't necessarily disagree (although I would question by how much more it would help), but, if this is the case, would that mean you would want to condition your support for future trade deals such that they would require more 'flexibility reforms'?

    1. I think focusing on "losers from trade" is a bad idea. There are "losers" all over the place. When Uber comes in, taxi drivers and their companies lose. When an auto company moves from Michigan to Tennessee, the Michigan workers lose, just as much as if the plant moved to Mexico. When washing machines came in, laundries and wash workers lost.

      Far more is "lost" due to domestic innovation and churn than that which occurs across borders. And far more is gained as well -- just as somebody gains when the Chinese manufacturer turns around and spends his dollars, as he must.

      The focus on borders appeals to unpleasant nationalistic and xenophobic feelings. The economic phenomenon cares little where the innovator resides.

    2. Good point, and I agree in many ways. It is far too easy to frame these issues from a nationalistic point of view, as you say, and ignore that any modern economy has an immense amount of churning and technological change.

      To the extent that these matter more, good policy would indicate that we should care about making sure that any worker can adapt and adjust to these changes, and we shouldn't focus on mitigating changes stemming from trade losses.

      I would say, however, that still leaves the political framing of the issue, though. I might very well wrong in my interpretation of Sanders/Trump and politicians and surveys in the EU, but assuming I am not, I would argue that we, as a profession, should still have some good answer to the question: "How can we help those who lose in a trade agreement?". Otherwise, I am concerned that opposition might increasingly grow.

      And not just having a good answer, as in rhetorically, but an answer that can be implemented and be effective.

      And, being completely honest, I am not sure what the right answer is. I believe that deregulation in general and taking steps that allow for more interregional mobility are certainly an important element (and one that has more general benefits), but given the findings of Autor et al., I suspect that wouldn't be enough.

      Perhaps a minimum income policy, that would potentially allow for more innovation (by providing a safety net) could radically change our perceptions of the problem, although perhaps not.

      And, then again, I might be overly concerned about this opposition, and I would be happy to be convinced otherwise.

    3. John,

      GDP = C + I + G + (Ex - Im)

      "Ex - Im" is now some nationalistic, isolationist, xenophobic plot?

  3. NAFTA was a disaster because of Mexico, not because of Canada. It's a basic tenet of international trade that trade is generally beneficial when conducted among nations at equal development levels.

    Even Paul Krugman writes that by allowing illegal aliens "we are importing poverty".

  4. I understand free trade theory. But deep in the night, late in the bar, liquored-up, I will tell you free trade probably doesn't work in the real world for the United States.

    Ronald Reagan was the Great Protectionist according to Cato Institute. Milton Friedman said Ronald Reagan made Smoot-Hawley look benign.

    What Donald Trump has proposed is dwarfed by what Reagan implemented!
    Trump wants a 45% tarriff against one nation and that as a bargaining chip. Reagan erected tariffs against dozens of nations, sometimes up to 100%.

    Yet, we reflect on the Reagan years as ones of prosperity.

    This business of mounting debts to foreign holders, in the hundreds of billions of dollars a year. What will be the impact on inflation?

    And do we say that free trade works for the United States, but only due to the structural impediment that the U.S. dollar is the reserve currency?


  5. Yes, trade is about efficiency and wages; but, not just wages and efficiency – it is about the allocation of scarce resources, so productivity as well as efficiency are considerations. And our productivity (as measured by per capita real GDP) has been in decline since the 1960s. Since the 1990s we have engaged in trade agreements that have off-shored productive sectors of our economy and traded them for service sectors that do not have the same multipliers or effect on a broad segment of the different sectors of our economy. The result is fiscal policy today cannot work like it did in the 1960s; and, monetary policy has its limits.

    Why should a country like the United States, the most advanced economy in the world by far, have a trade deficit with an under-developed country like China and with Chinese money coming back into the U.S. for investment? Shouldn’t it be the other way around? China is the one that needs the investment to develop, much like the U.S. did during the second half of the 19th Century. Where is this Chinese investment in the U.S. going? It feeds our finance sector – which by the way has grown as a percentage of our GDP coincidentally with the trade imbalance. Is this economics?...or political influence? And with our slowing productivity growth over the past twenty-five years, it makes you wonder if the investment in the U.S. is getting the best return than if we had more of our productive capacity on-shore rather than off-shore and an economy that was less centered around our service sector and more centered around industrial and commercial enterprises. When the U.S. was going through its industrialization during the mid-18th Century it was a net-importer of capital investment from Europe – we did have a tariff that affected the trade in goods. Why shouldn’t China be the same? They are a net-exporter because they have an industrial policy to be a net-exporter rather than ramp up domestic development any faster than they have.

    1. "And our productivity (as measured by per capita real GDP) has been in decline since the 1960s"

      I think that's a bit of a stretch. The Kennedy/Johnson years did have by far the highest per capita growth rate of any post-WWII administration, but the second and third highest were Reagan ('81-'88) and Clinton ('93-'00).

      It looks roughly the same if you look at full business cycles- a big surge in the 60's, followed by a gradual slowdown into the early 80's, followed by stronger and more stable growth through the 2001 recession, and then much slower growth since then.

    2. A stretch?
      Per capita Real GDP – 2009 $s; per capita Real GDP factors out the effects of inflation and population growth

      1. 1940s 40%
      2. 1960s 35%
      3. 1790s 33%
      4. 1870s 28%
      5. 1980s 25%
      6. 1970s 25%
      7. 1920s 24%
      8. 1850s 24%
      9. 1880s 23%
      10. 1990s 21%
      11. 1950s 19%
      12. 1840s 15%
      13. 1910s 14%
      14. 1930s 13%
      15. 1830s 10%
      16. 2010s 5% for 5 years
      17. 1890s 9.0%
      18. 1820s 8.0%
      19. 1860s 7.6%
      20. 2000s 5.5%
      21. 1900s 5.0%
      22. 1800s 3.0%
      23. 1810s 1.0%
      [Source: BEA Historical Tables]
      Zack -- that's an interactive chart you referenced. You can pull the numbers down from there -- and don't forget the '70s.

    3. I don't see how lumping together decades as your unit of measurement makes sense. Decades don't correlate with business cycles or even presidential administrations- the 1980's, for example, is a combination of Reagan plus 1/4 of Carter's term plus 1/4 of Bush's.

      Looking at business cycle peaks/troughs according to the NBER:

      Following the end of that 1960's boom, the next recession was in '73-'75. If my math is correct, per capita growth during that cycle averaged 2.61% peak to peak and 1.76% trough to trough.

      The next recessions came in '80 and '81-'82. Growth during those two cycles averaged just 1.55% peak to peak and 1.62% trough to trough.

      The next recession was in '90-'91. Growth jumped to 2.41% peak to peak, 2.86% trough to trough during that cycle.

      The next recession was in 2001. Average growth rate: 2.08% peak to peak, 2.19% trough to trough.

      If one wanted to simply look at presidential administrations instead, the trend is similar. The average growth rate from 1981-2000 was considerably higher than from 1969-1980.

      So yes, I do think it's a stretch to say per capita growth "has been in decline since the 1960s."

    4. Any grouping of the data is arbitrary. Let's just take your referenced chart and instead of looking at the overall increase, break it down by annual increases over the prior year.

      It looks like a downward trend in producivity growth to me.

    5. How is it arbitrary to look at full business cycles? That seems like the least arbitrary way of measuring things.

      "It looks like a downward trend in producivity growth to me."

      How so? There was clearly a lot more volatility from the late 60's to early 80's with sharp ups and downs every few years. That doesn't mean that growth overall was lower, just that the booms and busts were more frequent and severe.

      Again, look at the rates for those business cycles or even presidential terms above. Do you dispute those numbers?

    6. If you want to measure the 1960s productivity (what NBER calls an expansion) NBER style -- like you said -- trough to peak:
      The trough - 1960/Q4 = 16,938; The peak - 1969/Q3 = 23,332. That's 4.19%.

    7. Comparing expansions, NBER style (trough to peak)
      1960s I already gave you (from 1960 to 1969) = 4.19%
      1970s -- I am skipping, too many recessions (but there was good productivity gains)
      1980s -- 1982/Q4 to 1990/Q3 = 3.59%
      1990s -- 1991/Q1 to 2000/Q4 = 2.72%
      2000s -- 2001/Q3 to 2007/Q4 = 1.60%

      To me, that looks like its going down. Your turn.

    8. Zack -- We have the data broken out by decade and by NBER methodology. If the decade data shows the 1960s to be 40% better than the 1980s and 70% better than the 1990s, why do you think by rearranging the time period you will get a different result when the 1960s is head and shoulders above any other peacetime decade (and period) in U.S. history? I have lots of opinions; BUT, that is a FACT!

    9. 1. I already said in my original comment that growth in the 60's was higher than any other time period. So what? That has nothing to do with whether it "has been in decline" since then.

      2. You can't only look at expansions and just exclude the recessions! Of course growth will look higher when you completely remove all of the worst quarters!

      3. You claimed growth has been in decline since the 60's, but now you're saying you can just skip the cycles of the 1970's because there were "too many recessions?"

      As I said before, look at those two long cycles from the '81-'82 recession to the '01 recession. Whether you measure peak to peak or trough to trough, growth during those complete cycles was stronger than the average of the previous three cycles, which started with the '69 recession. For presidential terms, growth under Clinton- last year in office, 2000- was significantly higher than Nixon, Ford, or Carter. Again, are you disputing any of this?

    10. [I don't know if this comes through.]
      Chart: GDP (percapita real GDP) rates of growth from 1960 to 2015. Including contractions and expansions -- a straight line trendline goes from year 1 (1960) from 3% to year 55 (2015) to 1%. That's downward sloping, isn't it?

    11. Again, in my very first comment I said that there was a major boom during the Kennedy/Johnson years and also that growth has been much lower after about 2000. Drawing a straight trend line by starting in 1960 and ending in 2015 doesn't refute either of those things and it doesn't mean that there has been a continuous downward trend after the 1960s.

      I don't know how many ways I can try to make the same point- if you look at entire business cycles, the growth rates during the long Reagan/Bush and Bush/Clinton cycles, which covered almost 20 years, were higher than the previous three cycles by a solid margin:

      Peak to peak: 2.23% per capita from Q3 1981- Q1 2001 vs. 1.91% from Q4 1969- Q3 1981.

      Trough to trough: 2.48% per capita from Q4 1982- Q4 2001 vs. 1.67% from Q4 1970- Q4 1982.

      Unless I've made a huge mathematical error somewhere, I don't see how this can be described as being "in decline since the 1960s."

    12. And if you draw a straight line trendline from 1970, the trough year, the negative GDP year, the year of the recession -- 1969 is positive and 1971 is positive -- from 1970 to 2015 -- the straight line trendline is negative slope: from about 2.25% in 1970 to about 1.25% in 2015.

      So, if we are quibbling about words "since the '60s", where I include the 1960s and it seems you want to start after the 1960s -- the GDP is still declining.

      Basically, the post-WWII years to present, the GDP is in a negative trendline -- with one big exception; 1946 the GDP was down 12.5%, so actually the long-term trendline is just slightly positive. When you take the trendline from 1947 to 2015 it is distinctively negative, like from the '50s, the '60s, the '70s.

    13. "from 1970 to 2015"

      This is getting tiring. I've already said multiple times -including in my last comment- that growth has been much slower since around 2000. And no, drawing a graph with a "straight line trendline" from one point to another, several decades apart, does not mean that there has been a continuous decline (or increase) for the entire period in between.

      I've been making the same point from the start, regarding various business cycles. I'll make it one more time. Growth during the two long cycles from the 1981-82 to 2001 recessions was well above the average of the previous three cycles from the late 60's through early 80's.

      For that matter, the average growth rate during those two long periods looks like it was also stronger than the average of the cycles that preceded that 60s boom.

      Again, if you don't dispute the numbers I've been posting, I still don't see what there is to argue about here.

    14. Quibbling about words --
      Even going from the 1980s to the 1990s the rate of GDP growth was in decline: the '80s cycle, trough (1982/Q4) to trough (1991/Q1) = 3.08% per annum; the '90s cycle, trough (1991/Q1) to trough (2001/Q4) = 2.63%.

      I said it was arbitrary grouping data by decade. Using NBER does not change that. You want to be arbitrary in the time span you pick. The original statement that you object to is "in decline since the 1960s" -- I used that statement to be inclusive of the '60s.

      To do a sensitivity analysis to reverse that downward trend you have to chop off a good bit of the last 15 years as the ending point; and, chop off a good bit of the '60s for the beginning point -- that is the only way you will get a positive trendline.

      Arbitrary is as arbitrary does.

    15. "Even going from the 1980s to the 1990s the rate of GDP growth was in decline: the '80s cycle, trough (1982/Q4) to trough (1991/Q1) = 3.08% per annum; the '90s cycle, trough (1991/Q1) to trough (2001/Q4) = 2.63%"

      I think those numbers are off, but in any case you're moving the goal posts again. You claimed growth "has been in decline since the 1960s" so I posted the growth rates for full business cycles since the 1960s. These numbers show that at most, it's been in decline since the 1990-91 recession. That's a big difference.

      The rest of your post is just repeating stuff I've already responded to.

  6. In addition to the trade imbalance we are also running up huge federal budget deficits that also suck up foreign investments in our treasuries. Is it any wonder that the growth of our trade imbalances and our budget deficits are coincidental and linked to our own declining economic growth and whether we are growing at near full employment (late ‘60s and late ‘90s) or not?

    In addition the overall ‘balance’ or surplus/deficit affects our exchange rate and the dollar status as the reserve currency. Right now the Fed wants to raise interest rates and when it made a micro-minuscule change recently ramifications went world-wide and into China with a strengthening dollar and weaker economies overseas.

  7. Also, whose interests are being served when something as obscure as copyright and trademark provisions to minor detail get into trade agreements but something as ubiquitous as labor agreements, health and safety provisions, and environmental concerns are glanced over?

    The free trade argument is nothing more than a rhetorical exercise that has no basis of factual reality and only exists in academia. Where has free trade ever existed?: not between any countries, not the European Common Market (there are fees within the Value-Added Tax, subsidies to protected industries), not between our states (tax abatements for property tax and other subsidies and inducements to move companies); and not even in Free Trade Zones nor Duty-Free Shops, which are artificial constructs.

    In none of the examples of a Ricardian comparative advantage do you get a full discussion of opportunity costs (trading manufacturing jobs and related sectors for service sector jobs); externalities (tax policies, currency manipulation); costs shifts from the firm to another entity (the fissured workplace); or societal costs (help for displaced workers). If economics is the study of the allocation of scarce resources then total costs and total benefits must come into the equation of “free” trade to make it fair trade. So, are we talking about economics or political influence when it comes to free trade?


Comments are welcome. Keep it short, polite, and on topic.

Thanks to a few abusers I am now moderating comments. I welcome thoughtful disagreement. I will block comments with insulting or abusive language. I'm also blocking totally inane comments. Try to make some sense. I am much more likely to allow critical comments if you have the honesty and courage to use your real name.