by Pierre Lemieux
Peter Navarro’s film raises one interesting question: How can somebody be so wrong?
In the wake of the new tariffs on steel, a Wall Street Journal story reminds us of the steel industry’s influence over the U.S. government (“Navarro’s Ties to Nucor Highlight Trump Advisers’ Steel-Industry Connections,” March 16, 2018). Many of the senior advisors to the President have worked in or for the steel industry. Peter Navarro, a senior trade advisor, received $1 million from Nucor Corp., a steel manufacturer, to produce a 2012 documentary film, Death by China, against imports in general and from China in particular. A book with the same title was published in 2011, co-authored with journalist Greg Autry.
These facts confirm one of the conclusions of public-choice theory and the theory of collective action: special interests, especially producers’ interests, will capture the government to obtain interventions to their benefit. This is called rent-seeking, of which steel tariffs are just an example.
The poor economics of Navarro’s documentary has been well criticized by Daniel Griswold of the Mercatus Center. The film is pure propaganda. So is the book, but I will focus on the film here. Peter Navarro’s film raises one interesting question: How can somebody be so wrong?
Among the errors, one is particularly glaring. It is worth watching Peter Navarro at 1:08:12 in the film. With great certitude and professorial confidence, he says:
And the argument is pretty simple. The gross domestic product grows with only four things: consumption, investment, government spending, and net exports. And when you run a trade deficit, the simple math of that is that that’s a negative in terms of growth.
This is simply and demonstrably false. As I explained in a previous post here (“Misleading Bureaucratese“), GDP is not equal to the sum of consumption, investment, government expenditures, and net exports (exports minus imports). It is instead the sum of consumption (right!), investment (right!), government expenditures (right!), and exports (not net exports!). GDP is the acronym of gross domestic product, and only expenditures on domestic production are included.
The reason why, as a pure accounting operation, the national statisticians of the Bureau of Economic Analysis (BEA) subtract imports is that they are already included in the measure they have of consumption, investment, and government expenditures. The value of GDP they thus calculate does not include imports in any way–and specifically not as a subtraction–because GDP cannot include them by definition.
In short, imports do not reduce GDP or its growth in any way.
A good introductory textbook of macroeconomics will explain this. Moreover, one can easily check it in the official explanation of the nature and measurement of GDP, the BEA’s Concepts and Methods of the U.S. National Income and Product Accounts (November 2017). I quote from p. 2-9:
Thus, GDP is equal to personal consumption expenditures (PCE) plus gross private domestic fixed investment plus change in private inventories plus government consumption expenditures and gross investment plus exports minus imports. In this calculation, imports offset the non-U.S. production that is included in the other final-expenditure components. For example, PCE includes expenditures on imported cars as well on domestically produced cars; thus, in order to properly measure domestic production, the sales of foreign-produced cars that are included in PCE are offset by a comparable entry in the imports of these cars.
All statistical agencies in the world follow the same methodology.
How can somebody like Peter Navarro, who after all has a Ph.D. in economics from Harvard University and used to teach business at the University of California at Irvine, defend an interpretation that contradicts a fact that even college economics students must understand? Perhaps he simply ignores what GDP is, because he touched so many other fields of inquiry and action, and has been teaching business, not economics? This is not a very satisfactory answer because he knows, or at some point has known, basic national accounting, and the sources of information are easily available to him. He could just check the BEA’s Concepts and Methods.
Another hypothesis is that his opinion was bought off by the steel industry, but I don’t think this is satisfactory either. We must assume, at least for the purpose of discussion, that, wherever the money comes from, the proponent of an opinion maintains basic intellectual honesty. The opinion must be discussed on its own merits.
Here is another hypothesis, more acceptable I think. Navarro may have once heard, or perhaps even read (although a written source would not be easy to find), a different interpretation of the accounting identity referred to above. Perhaps he just quickly glanced at the accounting identity in a macroeconomics textbook without carefully reading the explanation. This immediately confirmed his previous political preferences–his prejudice–and he did not further investigate the matter. Psychologists and behavioral economists call this a “confirmation bias” (see, for example, the discussion of psychologist Raymond Nickerson in the Review of General Psychology, 1998).
It is as if Professor Navarro’s intellectual wheels had got stuck in ruts and he had no good reason to get his chuck wagon out of them. He had no good reason because the wagon moved in the direction he wanted to go. That some people would pay him to stay in the ruts, he might have seen as further confirmation that his peculiar theory was true.
Of course, we are all subject to fall in intellectual ruts, to accept facts or ideas simply because they confirm our prior opinions and values. It takes effort to avoid confirmation bias. It’s difficult to admit that one was wrong. But trying to avoid confirmation bias and be ready to admit error would seem to be a necessary condition for any intellectual (and not purely rhetorical) enterprise. Truth, even if the quest is difficult, must be the ultimate criteria. I like how, in Freedom and Reform, Frank Knight set the standard:
And the obligation to believe what is true because it is true, rather than to believe anything else or for any other reason, is the universal and supreme imperative for the critical consciousness.
READER COMMENTS
michael pettengill
Mar 19 2018 at 3:57pm
Don’t make the same mistake with “investment”.
Google:
Andrew_FL
Mar 19 2018 at 6:02pm
I think you’re overly optimistic about economics textbooks and undergraduate introductory macro classes. In the introductory macro class I took, this aspect of GDP calculation was indeed incorrectly explained in exactly this manner and I would venture to guess many no econ majors who took an introductory macro class would attest to the same.
Swimmy
Mar 19 2018 at 6:36pm
I am flummoxed by how ridiculous this particular argument is.
Imagine a business hires me as a consultant. Before they can even show me the states of various projects, I proclaim:
“I know how to raise profits. Profits is simply Revenue minus Costs, right? And employees are a cost in this equation. So all we have to is fire all the employees, and we recover all of those costs as increased profit! Bam! Done.”
I am fired immediately.
Later, I go to the Republican candidate for president and give him an argument of exactly the same quality. He promises me a position in his cabinet. Half of the country is ok with this.
Radford Neal
Mar 19 2018 at 8:31pm
An even more basic problem with Navarro’s statement is that even if he were right about the definition of GDP (which of course can be whatever we choose), a positive value for net exports is NOT a good thing!
Higher GDP is of course not necessarily a good thing in general – such as when it goes up due to repair work after a disaster. And the component of GDP due to gross exports (or net exports, if we redefined it that way) is always going to be a bad thing.
Believing exports are good in themselves is the basic fallacy of protectionism. This is all too rarely pointed out in public discussions, even by free traders. It is good, GOOD, GOOD! if we import more than we export! The only cause for worry in this situation is that we might suspect that the party can’t last forever – that sooner or later, people in other countries will figure out that it’s stupid to send us nice, valuable stuff while they get nothing in return…
This worry is captured in the current account balance, which is in deficit for the US at the moment, but not unusually so.
Mark Z
Mar 19 2018 at 8:39pm
“And the argument is pretty simple. The gross domestic product grows with only four things: consumption, investment, government spending, and net exports. And when you run a trade deficit, the simple math of that is that that’s a negative in terms of growth.”
I have a few other genius ideas: per this ‘simple math’, why not just raise government spending by 20 trillion dollars? Or, to think we could increase GDP growth merely by increasing our consumption! If everyone went out and blew all their savings on consumer goods tomorrow, prosperity would ensue!
That each of the factors that make up GDP are not independent of each other from one year to the next is does not seem like something that it should require a PhD in economics to see.
Peter Gerdes
Mar 20 2018 at 7:47am
I think you are being a little unfair here when you attack his definition of GDP. He didn’t say *domestic* consumption and its reasonable to assume that when he says consumption he means total consumer spending. Indeed, that is the way non-experts would naturally understand the word.
And he is right that in the short term it really is true that a sudden shift to buying foreign imports rather than domestic products (a popular imported toy comes out at Christmas) reduces GDP in nominal dollars. It’s just that provided we don’t start exporting more and don’t stop manufacturing stuff our money gains purchasing power (we now get the foreign imports plus all the stuff we were getting domestically for the same price).
But most of all this is just an indictment of GDP as an important benchmark. Ultimately we value comfort and consumption and want to do less back breaking labor so if we stopped making as much stuff because other countries insisted on giving it to us then yay for us.
Pierre Lemieux
Mar 20 2018 at 12:35pm
@Peter Gerdes: Thanks for your comment. I am not sure I have made my argument sufficiently clear. (My previous post and the references therein are more exhaustive.) Let me try to recast it.
My argument is a purely accounting one. That Navarro takes consumption expenditures as including imports is the very reason that makes it necessary to deduct imports from the formula, for imports are not part of the GD(omestic)P by definition. Imports must be taken out precisely to offset their inclusion in consumption expenditures, as the BEA explains. If imports increase by $100, the value of GDP does not change a iota because the new imports increase consumption by $100 but $100 more is deducted from the total (in the supposed “net export†term).
Your second paragraph, then, is not relevant to my argument. A sudden shift to increasing imports does not change GDP at all. It may be that, for reasons different from accounting ones, imports eventually lead to a reduction of domestic production (that is, GDP); as you say, “we†get more for less, and that would be good. But this point is outside my accounting argument.
Here is an analogy. The weight of 1 litre of water is 1 kilogram. The weight of a one-litre bottle of water is equal to the weight of the water (1 kilogram) plus the weight of the plastic bottle (20 grams). This is the same as saying that the weight of the water in our bottle (1 kilogram) is equal to the weight of the bottle of water (1.02 kilogram) minus the weight of the plastic bottle (0.02 kilogram): this is our identity. It is obviously a mistake to say that if you decrease the weight of the plastic bottle, this decreases the weight of the water: that is exactly the mistake I am criticizing.
I agree with your last sentence, but it does not contradict anything I said nor does it conflict with the definition of GDP. Remember that GDP is production, which is equal to income, which is a means to consumption (of either domestic or foreign goods, for either current or future consumption).
dede
Mar 20 2018 at 10:06pm
The litre bottle analogy is great!
Pierre Lemieux
Mar 20 2018 at 11:58pm
Thanks @dede. I had been thinking of such an analogy and @Pierre Gerdes’s comment (thanks to him too) gave me the opportunity to fine-tune the idea.
Pierre Lemieux
Mar 21 2018 at 1:17am
Oops! I meant “@Peter Gerdes” of course.
OH Anarcho-Capitalist
Mar 21 2018 at 8:34am
Well, since the topic is not just shameless mercantalist propaganda but Keynesian constructs used to “measure economic output”, I’d argue that government “spending” which consists mostly of transfer payments and production of non-private sector goods and services using current capital resources, government consumption should either be ignored as duplicative ala imports or subtracted from investment!
Warren Platts
Mar 21 2018 at 11:21am
It is unfair to say Navarro is suffering from confirmation bias. There is a grain of truth to what he says. In the GDP equation, C is actually a combination of Cd and Cf, as follows:
Y = Cd + Cf + I + G + X – M
where Cd is consumption of domestic production and Cf is consumption of foreign production (for simplicity, we will ignore Id, If, etc.)
So suppose in period 1, Cd=55 and everything else equals 15. In that case, C=70, and Y=100.
Now suppose the country decides to live within its means but increases Cf by 5 to 20. Assuming I,G,X=15, then Cd must be reduced to 50. C remains unchanged, but NX is now -5, and Y=95. GDP is down.
On the other hand, suppose period 2 is December, and people break out their credit cards to buy Xmas presents, so C increases to 75. Suppose further that domestic production satisfies the increase in consumer demand. Then Cd = 60, C=75, everything else equals 15, then Y=105. GDP is up.
But suppose instead that the entire increase in demand goes for imports. Now, Cf=20, M=20, C=75, and Y=100. And now, all of a sudden, the GDP equation is an accounting identity: “In short, imports do not reduce GDP or its growth in any way,” it is said.
There is no free lunch, however. GDP remains unchanged, AND consumption goes up. What’s not to love? Well, of course, the increase in consumption is financed by debt. Or what is practically the same, by the sale of assets. Either way, the bottom line on our balance sheet is reduced. Our stock of wealth declines by 5.
And it is inaccurate to say rate of growth is unaffected IMO because the rate of growth is relative to the previous counterfactual case where the debt was used to finance the purchase of domestic production. That is, even though GDP is unaffected, the rate of growth is less than it otherwise would be if debt had been spent on domestic production.
All this is not to say that if we could magically close the trade deficit tomorrow, the annual GDP growth rate would jump up by 3% per year. But it would probably go up by 1% per year–and when added up over 30 years, that is huge. We can see that in the past 30 years, average growth was only 2.5%, compared to 3.5% for the 30 years previous to that (57-87), prior to the era when “unilateral free trade” really took off with NAFTA, WTO, China.
Yeah, yeah, you say: What about taxes, what about regulations, what about employment? Well, Trump has lowered taxes, relaxed regulations, and we have low unemployment. The economy is seemingly doing great, but the GDP growth rate is not increasing proportionately. Instead, the trade deficit just keeps getting bigger. Instead of recycling our gains within our own economy, we spend much of it on imports. It’s getting to be a pattern: these days, whenever we get a decent boom going on, the trade deficit just balloons, and GDP growth rate doesn’t change much.
That free trade would cause a headwind for US GDP growth was totally predictable according to modern theory by the end of WWII at the latest (although any 19th century protectionist would say the same), according to Paul Samuelson (2004). As he put it: “Insidiously, and later dramatically, a catch-up nation so to speak often throws out an adverse head wind, slowing down the growth rate of the lead U.S. bicycle racer. This is realism and not just captious pessimism.”
Pierre Lemieux
Mar 21 2018 at 12:50pm
@OH Anarcho-Capitalist: You are right to say that the inclusion of government production at cost in GDP is very questionable. Simon Kuznets, one of the main builders of national accounting, opposed doing so. It is wrong, however, to suggest that government transfers enter into GDP; only government expenditures of goods and services, of course, are included. Reading my article “What You Always Wanted to Know about GDP But Were Afraid to Ask†(Regulation, Winter 2006-2007) may help understand what GDP is and how it is measured; it will also tell you why the decision to include government expenditures on goods and services at cost was made.
Pierre Lemieux
Mar 21 2018 at 1:00pm
@Warren Platts: I think the main error in your reasoning is that, in your fourth (not counting the equation) paragraph, you arbitrarily decrease GDP by 5. Of course, if one decreases the GDP figure (from 55 to 50), it will decrease (by 5). But–and that is the point–an increase of imports by 5 will not do this. It will increase C (that is, the Cf part of C) but decrease M by the same amount, so there will be no change. Consider that the equation (properly an accounting identity) you give describes the expenditure side of GDP, that is, to whom domestic production (100 in your example) goes. The expenditure side of GDP is by definition equal to its production/value-added side. (Exercise: Where do we see in the identity the part of production that is exchanged for imports?) It is also important not to confuse production with assets. You may find a few useful explanations in my article “What You Always Wanted to Know about GDP But Were Afraid to Ask,†Regulation, Winter 2006-2007. The analogy I gave in the discussion above might also be useful.
David Cushman
Mar 21 2018 at 7:58pm
The fundamental problem with Navarro’s statement is that it ignores the relationship of changes in NX (net exports) to the other variables in the GDP = C + I + G + NX equation. First, in the short run, the effect on GDP of a decrease in NX from a rise in imports (M) depends on the extent to which the increased imports reflect a fall in spending on domestic goods versus a fall in saving. To the extent the rise in imports reflects a fall in spending on domestic goods, then GDP does fall. I guess this is what Navarro is thinking of. This is old-style short-run Keynesian stuff, where GDP is solely determined by spending.
But I think the more glaring error is his failure to see the long-run situation. If saving (S) in the long run is not affected by the composition of consumption (domestic vs. foreign), a reasonable assumption, then changes in imports will instead be offset by changes in investment or exports, not in GDP, because S = I + X – M. And if I (domestic investment) is not affected by the trade balance (I know of no reason why it would be), then a tariff-induced fall in imports leads to a fall in exports through an appreciation of the dollar. So, a Navarro-Trump import restriction causes a fall in exports, while leaving the trade balance unchanged.
Great stuff: Navarro-Trump plan kills exports.
Warren Platts
Mar 22 2018 at 12:56pm
Hi Pierre: Thanks for the link to your informative Cato article on GDP (that I shared on twitter)! However, I still maintain that increasing imports could slow down GDP growth, though this is not necessary.
As I see it, an increase in imports can be can be paid for in one of three ways:
I made a spreadsheet to show what I mean:
t Y C Cd Cf I G X M NX
0 100 70 55 15 15 15 15 -15 0
1 105 75 55 20 15 15 20 -20 0
2 100 75 55 20 15 15 15 -20 -5
3 95 70 50 20 15 15 15 -20 -5
If you can read that, there is an initial situation where trade is balanced, and foreign consumption (Cf) is 15. In the three following scenarios Cf goes up by 5 to 20. In (1) X also goes up by 5 to 20, GDP goes up by 5; in (2) Cf goes up by 5, but it is paid for by a sale of assets, so GDP is unchanged (although the national balance sheet would show a reduction in assets or increase in liabilities of 5); and in (3) the increase in Cf is paid for by a decrease in Cd.
It is the number (3) case that is worrying. For example, there were recent headlines that consumer spending last December increased by 5%, but since much of this went for imports, that is the reason that was given for the fact that GDP growth for Q4 did not exceed 3%.
Of course GDP still grew, but just not as fast as it otherwise might have, had more consumer spending been directed towards domestic production rather than foreign production. I think that is Navarro’s point. Am I wrong?
Pierre Lemieux
Mar 22 2018 at 1:02pm
@David Crushman’s comment (and some others) remind us of something important: We must distinguish between a purely accounting argument (what Navarro is obviously making and what I criticized) and an economic argument (which explains what adjusts in the real world in order for an accounting identity to hold true). If domestic production is to replace imports (if GDP is to increase), investment or exports must decrease to liberate the necessary resources. If exports decrease with imports, the substitute consumption goods will be more expensive because comparative advantage is negated.
Pierre Lemieux
Mar 23 2018 at 8:47pm
@Warren Platts: (1) In your first case, I don’t understand why GDP could go up. This would not be possible anyway under full employment. But it may not be really relevant to your argument. (2) Note that foreign investment can compensate for, or create, a trade deficit. (3) Your third case assumes import substitution, and moves from an accounting argument (the topic of my post) to an economic argument. I suggest the correct economic argument would be the following. If a lower GDP and more imports do not reduce total consumption compared to t=0, this means that the terms of trade have improved: through imports, Americans get more goods for less work. Everybody would be happy. (4) Listening to Navarro, it seems pretty obvious that he was making a purely (invalid) accounting argument. If not, he does not understand that consuming more by producing less (ceteris paribus) is good, not bad; that would be an economic error.
Joe munson
Mar 24 2018 at 6:01pm
“We must assume, at least for the purpose of discussion, that, wherever the money comes from, the proponent of an opinion maintains basic intellectual honesty.”
I think this is misguided, of course, a percentage of people are going to be dishonest for material gain, and if we simply must assume everybody is being honest we are far more likely to be conned.
Emerich
Mar 24 2018 at 7:11pm
Pierre Lemieux, your posts are always well written. True too, no doubt, but I decided to mention that your posts are always readable.
Pierre Lemieux
Mar 25 2018 at 1:28pm
@Emerich: Thanks!
Comments are closed.