The Power and the Money

Stating the obvious about the tariffs on Mexico

I am union-bound to say this, I suppose. The Mexican tariffs are stupid.

First, they will not achieve their stated goal, which is to get Mexico to stop Central Americans from arriving at the U.S. border. Consider what the Mexican government is currently doing:

1. Inviting Central Americans to stay in Mexico. The problem there has been that the Trump administration does not believe that Central Americans will stay in Mexico. Rather, it believes that legalizing their presence in Mexico only makes it easier to get to the border. Under pressure, the Mexican government suspended its visa program after 13,000 applied. But Mexico has allowed Central Americans to apply for asylum from Mexico, as the Trump administration wanted. 
2. Arresting and repatriating Central Americans: 29,660 in the first two months of 2019, about the same as the 29,239 average for those months in 2016, 2017, and 2018. President Trump actually thanked AMLO for this just last month. Through May, Mexico has detained 72,410 migrants and deported 53,371. That is less than the 340,944 Central Americans the CBP detained so far this year, but it is a substantial number. It is far from clear that Mexico could do better given the limited resources available to the Mexican government. 

Second, if the administration sticks to the plan, it will be an own-goal for President Trump. The tariffs on China are bad enough, but companies can adapt and the economy is strong enough to withstand them. Moreover, they are politically popular. Tariffs on Mexico are a different story. These are tariffs that will mess with supply chains. Alonso de Gortari (Princeton) recently calculated that 30% of the headline value of Mexican manufacturing exports to the U.S. consists American exports to Mexico. So they not only hits American consumers, they punch American producers as well. Plus, President Trump can kiss goodbye any chance of ratifying the USMCA. There are cheaper ways to keep immigration on the front burner for 2020, which I presume to be the political goal. 

Third, the tariffs benefit China by taking Mexico off the table as a potential site. Manufacturing has been moving to Mexico from China for years now. The trade war would have accelerated that, at least a little. It is a good thing for America to have industry move next door. But now, well, not so much.

Finally, here is something that is not a problem for this administration: the loss of any credibility as a negotiating partner. Yes, China and Europe will take notice of how the U.S. told Mexico that all it had to do was renegotiate NAFTA but then turned around and screwed it anyway. But that is a feature, not a bug. The current administration intends to impose tariffs forever. Sure, the President will tack back and forth according to political pressures, temporary advantages, and his own strange ego. But the goal is clear, as we reported back in October 2018. Blowing up your credibility just makes it all that much easier to insure that you cannot win the trade wars that you start. 

I have to admit, however, that I think the administration is going to back down on this one. The logic of point (2) above is pretty clear. So put that down as a prediction ... albeit not one that I am particularly confident in. 

Lessons from day care for President Trump

Bloomberg reports that President Trump is moving towards making good on at least one of his campaign promises: making a profit off American overseas deployments. 

The story is that he plans to introduce “Cost plus 50” as the standard: i.e., basing costs plus 50%. Now, this won’t earn the United States a whole lot of money. Basing costs just aren’t very high. The U.S. could conceivably gross about $15 billion from our allies, including a $5 billion markup. We would net less than $15 billion, however, considering that Japan currently kicks in about $1.8 billion and South Korea and Germany add another $800 million each. (South Korea just agreed to up that last to $1 billion.) But hey, $11.4 billion is still $11.4 billion. 

The thing is, the administration also plans to reduce the charge for countries that align their foreign policies with the United States. In other words, the plan would give governments concrete incentives to follow American foreign policy.

Before you get too excised about that, note that the United States already does exactly this. In 1983, the U.S. passed a law requiring the State Department to report how closely countries voted with the U.S. in the U.N. General Assembly. Since then, study after study has turned up the same pattern: countries that vote with the U.S. get either more U.S. financial support or more support from U.S.-supported institutions, like the IMF. 

Wait, the IMF is a tool of American foreign policy? Who would have imagined? Besides everybody, that is. (Specifically: Barro and Lee (2005) found that IMF loans are associated with similarity to U.S. voting patterns in the UN and economic ties with the United States. Stone (2008) and Eichengreen et al. (2006) found that countries that receive substantial amounts of U.S. foreign aid are more likely to be eligible for IMF financing.)

And of course the U.S. trades favors in even subtler ways. Germany and France support the U.S. at times when they otherwise wouldn’t because they know that the U.S. will support them even especially if the Russians invade. 

So who cares if the Trump administration turns subtext into text? Or even supertext, a sort of neon-green megawatt Bat Signal of our intentions, with a specific price tag attached? Heck, economists always say that certainty is good and prices are good. Trump will certainly bring more of both. Maybe we will get a big bang for our $11.4 billion bucks. 

Or maybe not. Sometimes, you see, it is better for quos and quids to remain implicit.

Consider what happened when an Israeli day-care chain introduced late charges. The number of late parents went up. By a factor of two. And if you are wondering, this was a controlled study, in which only some centers introduced the charges. 

The lesson? Before the centers introduced fines parents tried hard to arrive on time because it was the right thing to do. After the fine was introduced, arriving late became simply another commodity that you could buy, even if that price was called a “fine.” Be late, pay the fine, cleanse your conscience, free of any worries that other parents will disapprove. 

In other words, right not countries cooperate with the U.S. in part because they consider it the right thing to do. The U.S. might reward them, but only as part of an unspoken system of mutual favors between friends.

But in a Trumpian world, well, now U.S. support becomes just another commodity with a price tag on it. If the cost is too high, then a foreign government may as well just pay the price. Keep U.S. bases on your soil but follow whatever bloody foreign policy suits you, plus a small check to Washington.

In short, for an additional $11.4 billion per year (or less) the Trump administration will likely get a whole lot less cooperation while preserving most of our overseas commitments. Nice job, I suppose.

Of course, we would bet that this will not ultimately happen, the same way we predicted that NAFTA would not be abolished.

The method behind the madness of Trump’s trade policy

There is a recurring criticism of President Trump’s trade policy: it is badly designed to obtain its ostensible ends. The tariffs against China are supposed to pressure Beijing into changing its intellectual property policy. But the strategy is poorly designed for that end. Consider:

1. Instead of heavily taxing employment-heavy Chinese sectors, like toys and textiles, you moderately tax the intermediate and capital goods used by American industries;
2. Instead of building a coalition against China, you start trade conflicts with your allies ... once again over intermediate goods like steel and aluminum;
3. You further alienate our allies by threatening automobile tariffs* ... that won’t affect China;**
4. You don’t give the Chinese government any face-saving way out of the trade war. 

As a policy designed to change Chinese development policy, this is stupid. 

But as an exercise, my MBA and undergraduate students separately walked through this policy, and they came to a different conclusion: if you want to make sure that high tariffs on Chinese exports are hard to reverse, then the strategy makes sense. Hurting American industries and forcing them to rework their value chains is a feature, not a bug. Making it hard for the Chinese to back down is a feature, not a bug. And alienating the rest of the world, to further insure that the Chinese won’t compromise, is a feature, not a bug. 

It might not be the best way to insure that high tariffs stay around for the next few decades regardless of the electoral outcomes in 2018 and 2020, but it isn’t a terrible one either. Especially if causing collateral damage to American industries is the point.

In other words, there is method to the madness. High tariffs are not a tool, but the outcome. The trade war is not meant to be won, but to be continued forever.

* The new USMCA (oh God, do I have to change my NAFTA tag now?) is said to have a side agreement exempting Mexico and Canada from those potential tariffs. More on this when I have it. 

** Ford did announce that it was not going to start making cars in the People’s Republic, so it is not impossible to imagine that threatened auto tariffs might have some effect. But then you aim the car tariffs at China, not the world. 

Raising barriers to pay for a wall: we anticipate Trumpian logic!

Back in March of 2016, we suggested that a Trump administration could repeal NAFTA and claim that Mexico was paying for the wall via higher tariffs. A static back-of-envelope calculation implied that the U.S. would receive about $2 billion a year in additional revenue. Cutting that in half to account for falling trade volumes would give you about a billion. Thirty-year treasuries are currently yielding about 2.875%, so that billion-dollar revenue stream could plausibly be said to allow for $19.9 billion in borrowing to finance the wall, currently estimated to cost about $18 billion.

In reality, repealing NAFTA would not pay for the wall for two reasons. First, the economic dislocations would almost certainly reduce other tax revenues. Second, much (most?) of those tariffs would fall on American consumers more than Mexican producers. Third, inasmuch as Mexican producers will pay, many of those producers are American owned. Add that up, and repealing NAFTA will cost rather more than the additional tariff revenue.

But you could, in a static sense, claim that repealing NAFTA would pay for the wall. 

And now, in a fairly incoherent interview for the Wall Street Journal, the President has now jumped on the “logic” we laid out last March. “[Mexico] can pay for [the wall] indirectly through NAFTA. We make a good deal on NAFTA, and, say, I’m going to take a small percentage of that money and it’s going toward the wall. Guess what? Mexico’s paying.” 

The obvious read is that he intends to bring back tariffs on U.S.-Mexico trade (the small percentage) and will use that revenue to pay for the wall. If there was any consistency to this White House at all, I would take that as the death knell for the trade agreement.

But who knows what the President meant? Does he even know what he meant?

Three worthwhile Canadian puzzles

So, to continue with my recent interest in Canadiana, impelled by, well, being in Canada!

There is an interesting puzzle about Canadian economic development. Well, there are many many interesting puzzles, but what particularly grabs my attention at this moment is why Canadian productivity has persistently lagged the United States.

Below I show Canadian total factor productivity (TFP) as a percentage of American TFP between 1950 and 2014. Save for a few halcyon years in 1973-80, Canada has consistently lagged the United States. Moreover, the gap turned into a chasm in the 1980s and grew even larger after the Great Recession. 

The productivity gap predates WW2. After World War 2, Canada lagged the United States in total factor productivity. That is to say, a Canadian worker in the same industry with the same skills and working on the same machinery produced less than her American equivalent. In the early 20th century, the superficial gap was just as large: the average Canadian worker added only 81% as much value per hour as his counterpart south of the border.* (Table 6 on page 25.**) Canadian workers, however, used rather less machinery than their American counterparts, so TFP closed to about 94% of the U.S. level ... remarkably similar to the gap in 1950-67.

In fact, the gap widens if you go further back. In 1900, GDP per capita (the crudest measure) in Canada ranked only 80% of the United States. In 1870, it was 70%. Canada appears to have had more-or-less that income gap with the United States back to 1820. 

Heck, the gap goes back to even before the existence of either the United States or Canada! Vincent Geloso (Texas Tech) has a paper comparing 18th-century Quebec with New England and finds that living standards were about 31% lower in what eventually became Canada. (Remember, there was no settled European population in modern Ontario back then.) 

So what caused the gap? A seminal 1964 paper by H.E. English pinned the gap on American protectionism: Canadian businesses operated at too a small scale and tried to produce too wide a variety of products. In other words, Canada wasn’t big enough. (You can find a 1980 study making the same argument here.)

This seems plausible. And as theories go, it has the advantage of explaining why the United States has enjoyed such a lead for so long over so many places. Everywhere else is too small! There are just too many damn countries in the world. Those of you who have read my opinions about  secessionism will know that I would find this a very gratifying conclusion.  

But the scale explanation does not seem to fit with the big widening of the gap since free trade.

Or does it? It is true that trade costs between the U.S. and Canada are still pretty high. And it is true that interprovincial trade barriers place even bigger obstacles on Canadian firms than American protectionism.  And it is true that the returns to scale in Canadian businesses are pretty high. (Page 130 of the 2009 World Development Report.) Canada has lots of small firms compared to America (consistent with the scale economy argument) and those small firms tend to be less productive than their American counterparts (sort of maybe could-be consistent with the scale economy argument). 

Statistics Canada ran an interesting thought experiment to test the hypothesis. They found that if you altered the size mix of Canadian firms to look like American ones you would close the productivity gap by 5 percentage points while if you made Canadian small firms as productive as American ones you would close it by nine. (Table 6, counterfactuals one and two.) Fixing both effects would close two-thirds of the gap, call it 16 points.

But that is not the entire gap. In fact, it leaves an unexplained gap bigger than the total gap in the 1950s.

Or the 1870s, for that matter. Kris Inwood (University of Guelph) and Ian Keay (Queen's University) conducted a similar exercise for Canada in 1870. They could get Canadian TFP to rise from 93% of the U.S. level to 97%, which cuts most of the gap. (You can find an outline of their paper here; the main one is paywalled.) Strangely, they downplay their finding, which seems like a very Canadian thing to do. The punchline is not that a gap remained, the punchline is that they explained more than half of it!

So the fact that Canada had a protected small market explains a little more than half the 19th century gap and about two-thirds of the modern one. But that still leaves two puzzles. First, what explains the rest of the gap in both periods? And second, why has the gap grown so much larger recently, despite NAFTA and all the ways modern communications should make service exports easier?

It is a great pair of questions.

But I mentioned three in the title, no? What’s the third puzzle?

I will post on that later, but it’s even weirder than the productivity gap ... it’s that median incomes in Canada may have finally pulled even with the United States. That is to say, for the first time ever it matters not to most Canadians that they live in a fundamentally less-efficient economy than their southern neighbors. And that’s the biggest puzzle of all.  

* The gender shift is deliberate.

** John Baldwin and Alan Green, “The Productivity Differential Between the Canadian and U.S. Manufacturing Sectors: A Perspective Drawn from the Early 20th Century,” The Canadian Productivity Review (Statistics Canada: December 2008).

More evidence that Trump caused the peso to fall

First, pray for Miami. I have a love-hate relationship with that city, even deeper than my love-hate with Texas. I would not want to live in that entire metro, not in Miami-Dade, not in Fort Lauderdale, not in Palm Beach. But my family is there along with seven million other people. They have not evacuated. But my brother is sheltering the families of two of his daughters at his house outside the flood zone; the third is an R.N. and she is on call, along with her police officer husband. 

Second, allow me to present some evidence that Donald Trump was indeed responsible for most of the weakness in the Mexican peso since he arrived on the political scene in mid-2015. The Mexican peso has strengthened over the past year. When we were there this summer, the peso rose above 5½¢ and stayed there. It had plunged below 5¢ after the election. This was good for Mexicans who wanted to buy imported goods, and I have to admit that Mexico City is still very inexpensive, but still. 

The rise means that the peso is all the way back to its level of January 2016, before Donald Trump began winning any primaries. It is not quite back up to the 6¢ level of June 2015, when Trump rode down that escalator. That is probably due to residual uncertainty regarding the NAFTA negotiations: markets are partially pricing in the chance that the negotiations will fail, thereby exposing Mexican producers to WTO tariffs when exporting to the United States. 

And so, a prediction. If NAFTA comes out intact from the negotiations, then the real exchange rate of the peso will return back to the level it had back in 2015. 

Do note that candidate Trump could move the market with his words; President Trump much less so. Bets are being hedged, but nobody believes that White House words carry any information anymore. 

A new chapter in the NAFTA saga? Likely not

NAFTA renegotiation is upon us. The newspapers here have plastered their front pages with stories about the united front Canada and Mexico are putting up against the elimination of Chapter 19.

What’s Chapter 19? It’s the part of NAFTA that mandates binational panels review any anti-dumping actions imposed by one NAFTA country on another. Without it, NAFTA countries would be at the mercy of national courts if they contested anti-dumping actions aimed at their exports. You can find the details of all 118 NAFTA trade cases here; roughly 69% of all cases have been against the United States. (More stats at this link.) 

It turns out to be remarkably hard to calculate success rates. Who wins more in these panels? Someone would need to go through all 118 decisions and figure that out. (The headline decisions often split the difference, so it would be a judgment call.) That said, the data here let me do a quick-and-dirty estimate: in cases brought by Canada against the United States, the panel either made no decision or found in favor of the U.S. around 65% of the time. Winning!

There is a case to replace the ad hoc panels with a genuine international court, European Union style. But that is not on the cards. Another possibility would be to trust domestic courts: let American exporters put their faith in Mexican courts and Canadians trust American ones. But national court systems are slow and expensive, so Mexico and Canada have made it clear that they would rather junk NAFTA than rely on domestic courts to make decisions.

Now one can imagine all three countries setting up special domestic trade courts to review agency decisions on behalf of NAFTA plaintiffs in a timely manner. But the Trump administration has shown no sign of wanting to reform American courts along those lines and it unlikely that Canada and Mexico would accept such reforms regardless. It really looks like Chapter 19 or nothing.

Moreover, Mexico is putting its opposition on autopilot. The Third Commission of the Permanent Committee of the Mexican Congress (don’t ask, it gets complicated) has gone on record rejecting changes to Chapter 19. A full congressional resolution could tie the hands of the Mexican negotiators ... which might be, in fact, exactly what the Mexican government wants Congress to do. Consider two idiots playing chicken on Francis Louis Boulevard at 3am in Queens. One of them throws their steering wheel out the window. That idiot would certainly have an advantage in the resulting game ...

I would bet that the Trump administration quietly backs down, but with this man in office, you never really know. Still, FDI into Mexico is keeping pace, around $25 billion so far this year, similar to the $27.4 billion in 2016, $33.3 billion in 2015, and $27.7 billion in 2014. Similarly, the peso has recovered to around 17.8 per dollar, pre-Trump levels.

So, yeah, our terrible negotiator-in-chief is not likely to win this one, which is good.  

NAFTA and agriculture

There was some interest in this topic but not a whole lot. So I will put up a short post, which will be pretty straightforward. I will discuss U.S. exports today.

The key thing to keep in mind is that agriculture is a global market. So if third-countries displace American agricultural products in the Mexican market, then American producers will be able to export their products to the places the third-country used to sell to. But if Mexico just reduces its consumption of all agricultural imports, then U.S. producers will lose unambiguously.

Mexico takes 8% ($1.8 bn) of American soybean exports. It takes 21% ($2.4 bn) of U.S. corn exports. It takes 11% ($0.9 bn) of our wheat exports.  It turns out that 36% of U.S. corn travelled to Mexico by sea in 2007-10. See page 10. The share is even higher for other products: 55% for wheat and 38% for soybeans; see pages 20 and 23. I used USDA data to calculate costs per ton-mile for soybeans: subtracting out $230,000 per ship Panama Canal tolls for cargoes from the U.S. costs gives you an average cost per ton-mile of 49₵ from the U.S., 43₵ from Argentina, and 39₵ from Brazil; fixed costs come to $5.70 from the U.S., $7.20 from Argentina, and, inexplicably, only $2.70 from Brazil.

With that data in hand, let’s look at those three markets and imagine that Mexico slaps MFN tariffs on the U.S. while opening to the South Americans. 

Soybeans: Mexico could easily supply itself from Argentina and Brazil. Landed costs to Mexico are already cheaper for the Latin American nations, even with higher transport costs; only South American export taxes keep us competitive.

Slap a 4% tariff on American exports (while removing them on the South Americans) and Argentina and Brazil would drive us out of the market. But remember that soybeans are sold in a world market. So more South American exports to Mexico means less South American exports to somewhere else. American production will fill up the gap. When it all shakes out, you would have a small loss for the Americans, a small gain for the Argentines and Brazilians, and about a wash for the Mexicans. 

Corn: The U.S. and Argentina have more or less the same corn production costs. But transport costs from Argentina are quite a bit higher. If Mexico imposed MFN tariffs on corn and suspended them on South American exports, the landed cost of U.S. maize would still be about 13% less than Argentine corn. Americans would lose sales via a general cutback in Mexican purchases, not a switch to South America. 

Lose-lose, although more for Mexican consumers, since U.S. producers would presumably find alternative markets. That said, Mexican corn farmers would gain, and rural voters are a major constituency, about 30% of the electorate. (See page 10. Rural voters punch slightly above their weight in turnout terms: see page 9.)

Wheat: See soybeans. 

Upshot: ending NAFTA would not be catastrophic for American soy and wheat exporters, but both would have to discover new markets right quick or suffer some unpleasant transition costs. It would be a clear boon for Argentine and Brazilian farmers, who could capture a whole new market. Mexican corn farmers will gain; Mexican consumers will lose.

Still, NAFTA probably won’t end.

NAFTA, trade costs, and the difficulty of economic divorce

Market access can have a big impact even when the numbers are small. For example, extending normal trading relations to China did not dramatically increase access to the U.S. market. But it did increase certainty: previously, China’s access was subject to annual Congressional votes. Remove that uncertainty, and poof, suddenly firms were willing to make long-term investments aimed at the American market. (The paper on this topic is here.) 

The importance of permanence is why giant fluctuations in exchange rates should not be expected to have the same effects as small changes in tariffs. Exchange rates rise and fall subject to unpredictable economic forces. Tariff rates stay the same subject only to politics. Gabriel Mathy made that point here. 

So what about ending NAFTA? Would it undo economic integration between the U.S. and Mexico? 

Quite likely not.

Wait, what? Why would an end to NAFTA not benefit China and other low-wage manufacturers in Asia?

Well, simply put, they are too far away from America. In 2014, according to the Bureau of Transportation Statistics (see Table 10 and Table 11), the average container imported from Mexico carried goods worth $52,921. According to J.P. Morgan, the average cost of moving that container from Mexico was about $1,800. (The World Bank says $1,499, but that includes maritime and non-NAFTA shipments.)

So that is a cost  of roughly 3.4% ... compared to 15% from China. (The number of containers from China comes from the Maritime Administration; the value is from the Bureau of Transportation Statistics.)

This is back-of-the-envelope, but very little of the stuff that Mexico exports to the U.S. is set to move to Asia, Europe, or South America if NAFTA falls apart. Mexico is too close and its labor costs are too low.

Still, a NAFTA collapse could move production back to the USA, no? After all, it is not like U.S.-made goods don’t need to be transported to market. They do. Domestic truck and rail shipments came to $8.8 trillion in 2007. (To prevent link rot: BTS, National Transportation Statistics, Table 1-58: Freight Activity in the United States: 1993, 1997, 2002, 2007 and 2012. I used 2007 because 2012 trucking freight rate data is not available.) The total cost of moving all that stuff came to $262 billion. That comes to an average freight cost of about 3.0%. In other words, and not unexpectedly, the cost of moving goods around inside the United States is roughly equal to the cost of moving them from northern Mexico to the United States. So one would expect a rise in tariffs from zero back up to 3% to have an impact. 

But tariffs and transport costs are not the only trade costs! Total trade costs come to the difference in the markup between the factory gate and the final consumer between domestic goods and imported goods. When Dennis Novy (Warwick) used that approach (albeit indirectly) he found that NAFTA reduced the total trade cost from about 70% to around 33%. (Page 10)

How did that happen? The intuition is as follows. Signing NAFTA cut U.S. tariff rates from 3% to zero. Companies then established cross-border value chains. Those chains were not easy to set up, but once they were in place, it became far cheaper to export. Think of it this way: the first autoparts factory to open up in San Luis Potosí needed to figure out everything about selling to the United States. The fifteenth had a much easier time. Thus, NAFTA caused trade costs to drop by way more than the fall in tariffs.

The flip side of that, however, is that much of the fall in trade costs is not reversible. Abolishing NAFTA, therefore, might have remarkably little effect, taking trade costs up from 33% to 36%. Abolishing NAFTA plus a 20% border adjustment tax will reduce Mexican incomes (via depreciation) but would also have remarkably little effect on trade volumes. 

It would take a straight-up tariff in the 20%+ range to seriously undo integration in any reasonable time period, at least for manufacturing. And even then the impact would be slower and smaller than we think. In fact, most of the impact would come in the form of European cars becoming suddenly competitive in the American market. 

Agriculture would be a different story. I will discuss more if there is interest. 

Maybe NAFTA is safer than we think

It is true that President Trump has made threatening noises. And it is true that he has appointed people with protectionist bents to run Commerce and the USTR office. 

But said Secretary of Commerce, Wilbur Ross, founded a major Mexican autoparts manufacturer. International Automotive Components makes parts for doors, side panels, consoles, exteriors, floors and sound systems for BMW, Fiat Chrysler Automobiles, Ford, General Motors, Honda, Mercedes, Nissan, Toyota, Volkswagen, Hyundai/Kia, Volvo Truck and Freightliner. He also owns the International Textile Group, with plants in Morelos.  

Talk about irony!

Now, Secretary Ross will presumably liquidate those holdings. So he should not have a financial stake in preserving NAFTA. But he will viscerally understand the many companies that do. 

I now suspect that the much-ballyhooed NAFTA renegotiation is going to end with stricter rules of origin, say 80% instead of 62½. There may also be some weakening of the dispute-resolution mechanism and higher labor standards. But I increasingly think that the Trump administration will refrain from blowing it all up. 

Although I still have no idea what fakery they will use to pretend that Mexico is paying for the wall.

The Trump tariff and the American auto industry

I have no idea if the Trump administration plans to impose a 20% tariff on Mexico. It seems now that the surprise announcement had to do with plans for a border adjustment tax, which is a different thing. But who knows? This is the Trump administration! Why are you asking me to make predictions?

Anyway, yesterday I mentioned that any tariff would impose serious problems on the American auto industry. The big problem is twofold: first, Mexican suppliers provide a large chunk of the inputs into American auto production. Second, many of those parts in fact cross the border multiple times. The end result is that a tariff would raise the cost of producing domestic cars by quite a lot. 

Some figures: in 2012, the state of Michigan published a report on the Ford Motor Company’s Michigan Assembly Plant. The report found that the plant purchased about $3.1 billion from Tier I suppliers; of that, $831 million (or 27%) came from Mexico. (Roughly $3,800 per vehicle.) In 2014, the U.S. auto industry consumed $42 billion in intermediate inputs from Mexico (pages 10-11), approximately $3,500 per vehicle. So at first glance, the tariff would raise the cost of American-built small cars by $760 and all vehicles by about $710.

That is a lower bound, however. In the course of making a motor vehicle, auto parts cross the border as much as eight times. If the tax was applied at each crossing, the cost would go up commensurately. If you assume that components cross from Mexico four times, doubling in value-add each time, then the total cost per vehicle would rise from $710 to $1,300. 

The cost increase would hit small cars more than large ones ... but consumers would see the price hike in the large car sector. The reason is that auto companies in the United States need to sell a certain percentage of small cars in order to hit their fleet fuel economy standards. They will therefore hold down the price on small cars, transferring the impact to their larger vehicles.

The problem is that the U.S. producers more larger vehicles than small cars. 

How much would production shift back to the United States? I suspect less than the current studies predict. Why? Well, the Mexican peso would likely fall. The real exchange rate would go down and stay down. Mexican workers, who consume a lot of imports, would become poorer but they would also become cheaper. A 20% nominal decline in the peso means a drop from 22 per dollar to 26 per dollar. That is certainly conceivable. Hell, 30 per dollar is conceivable. A messy drop would potentially hurt the Mexican economy, if it turns out that Mexican companies have a lot of hidden dollar liabilities, but would nonetheless increase competitiveness.

And so, you would have a temporary increase in the price of large automobiles and light trucks, followed by a drop in Mexican costs. Some production would return to the U.S. (in order to avoid uncertainty) but in unsteady and unpredictable way. It would not be an economic apocalypse for the United States, but it would cause havoc across the Mexican economy with little (if any) employment gain in the USA and big losses for shareholders. And unless the tariff is extended to all countries, the big winners would be car manufacturers from Europe and Asia. 

That said, this is all theory. Who knows what the Administration will do?

A 20% tax on Mexican imports? Why, yes.

Now the Trump administration is floating the imposition of a 20% tax on Mexican imports to pay for the wall. Obviously, this would violate NAFTA, but the U.S. President likely has the unilateral authority to impose such a tax.

The International Emergency Economic Powers Act of 1977 gives the President the power to “investigate, block during the pendency of an investigation, regulate, direct and compel, nullify, void, prevent or prohibit, any acquisition, holding, withholding, use, transfer, withdrawal, transportation, importation or exportation of, or dealing in, or exercising any right, power, or privilege with respect to, or transactions involving, any property in which any foreign country or a national thereof has any interest by any person, or with respect to any property, subject to the jurisdiction of the United States.”

Presumably, the tax would be considered regulation of importation under the above paragraph.

Now, the Economic Powers Act also compels the President to “consult with the Congress,” but it is not clear that imposes a significant limit on his tax powers.

Alternatively, President Trump could use the Trading with the Enemy Act of 1917, which Richard Nixon invoked to impose a 10% import surcharge in 1971. But the Trump tax could be appealed to the  U.S. Court of International Trade USCIT). It turns out that there is some case law on this.  In United States v Yoshida International, USCIT’s predecessor ruled that the courts would not challenge the validity of a Presidential declaration of a state of war or national emergency but would check to see that the remedy imposed was appropriate to the nature of the emergency. I suppose that the emergency would be the existence of illegal immigration from Mexico and the remedy would be the wall. (Judging from U.S. v Yoshida, I do not think that the tax would stand under Trading with the Enemy but it might under the Economic Powers Act. Either way, it could be in place for years.)

The courts would have to review if this violates the NAFTA Implementation Act of 1994, but my guess is that the Implementation Act would not be held to supercede Presidential powers under the Trading with the Enemy Act or the Economic Powers Act. 

And even if it did, the Trump administration could announce that it was withdrawing from NAFTA. The logic would be that the NAFTA agreement allows withdrawal and the NAFTA Implementation Act gives the president the power to proclaim tariffs as laid out in the NAFTA. Poof, a rationale for unilateral Presidential powers. 

Of course, if you abrogated NAFTA, you could easily enough pay for the wall with the imposition of pre-NAFTA levies. It looks like the Trump administration is looking for more than that, however: they seem to want to damage U.S.-Mexico commerce as much as possible. 

Or at least drive down the value of the peso even further. 

The tax would also have to exempt autoparts. More on this later ... 

Peña-Trump meeting cancelled!

Over the past week, it seems to have sunk in to the Mexican commentariat that Trump is serious about building a wall and renegotiating NAFTA. Newspaper articles have gone from “Hey, maybe there is an upside here, we can modernize NAFTA!” to “We are at war.” 

Then President Peña cancels his meeting with the American president.

As I have pointed out, many Mexicans blame Peña for Trump’s election. They’re probably not wrong. The logic is that Peña could have humiliated Trump. Then he recently brings back Videgaray, who had to resign over the Trump summit, on the theory that Videgaray is down with the Trump administration. But then the American president orders construction on the wall and continues to insist that Mexico pay for it. It leaves Peña with the worst of all worlds. 

The signs are that the Mexican president is under great strain. His approval rating is down to 7%. “He stays late in his office and sleeps little,” says an administration insider. 

Ford is not moving jobs to Mexico

Head, hits desk. Eric Loomis (hat tip: Andrew Reeves) just repeated the Trump lie that Ford is moving jobs to Mexico. Ford is expanding production in Mexico, it is true. But Ford is keeping up production in the United States and Ford would be expanding production in Mexico even without NAFTA. 

The biggest piece of evidence for this is that Ford says so. But hell, don’t trust them. Look at what NAFTA changed.

Before NAFTA, the U.S. tariff on small cars was only 2.5%. That is not nothing, but nor is it enough to compensate for lower labor costs. Note that the real exchange rate of the peso has fallen by 5.4% since the beginning of the year. 2½ percent is barely a speed bump. Now, if you want to argue about NAFTA, you can argue about the 25% tariff that the United States imposed on trucks. That was a real tariff. But Ford isn’t moving truck production; it’s moving small cars. And NAFTA has fuck-all to do with that.

There is a discussion to be had about NAFTA. But I must wonder if Loomis is willing to have it in good faith.

Temper Tantrum Trump and trade policy

 OK, the fat crybaby from Queens becomes President and gets upset because Mexico looks at him funny, or something. What can he do?

Well, he can cancel NAFTA. It is not 100% clear that doing so will immediately cause tariffs to snap back. It is also possible that Mexico will regain its GSP preferences. The NAFTA Act of 1993 is silent on cancelling the agreement. But Fat Crybaby can do what he wants until a court decides. So boom, Mexico!

(Section 2 of the NAFTA Act implies that ... well, go read it and tell me. Canadians in particular.)

But that just hits them with America’s low existing tariffs. Fat Crybaby, he’s a mean mofo from Jamaica Estates. Gangsta lean straight from Queens, except whiny. So what else can he do without going full Mussolini? Well, step one is Section 122 of the Trade Act of 1974. Wham! 15% tariff surcharge plus quotas as tight as President Fat Crybaby wants to make them. (Presumably, tight enough to raise import prices 45%.) 

Could he go further? Sure! President Nixon did in 1971. Now, the Customs Court found against him, but keep two things in mind. First, we do not know how higher courts would have ruled. Second, Nixon imposed the tariff surcharge on  August 15, 1971. The court ruled on July 8, 1974. That is a long time.

But if gets better! Section 338 of the Tariff Act of 1930 allows retaliatory tariffs of up to 50%! Badabing! Now, in point of fact, Section 338 is aimed at tariff jurisdictions that discriminate against the United States compared to other countries. But hey, see above about how long it will take for legal challenges to say that the authority does not apply to target to be decided. 

And should President Fat Crybaby decide that China is the target, well, the International Trade Commission has already issued a finding under Section 337 of the same law. Whammo! In the nose!

And then there is the Emergency Economic Powers Act of 1977. It involves foreign exchange transactions, but it could be used to effectively impose quotas. So if a court ruled against the above, Fat Crybaby could pull out Section 203 of the Economic Powers Act and wait for the courts to decide that one. 

In short, if Fat Crybaby throws a temper tantrum in office, then he can pretty much do whatever he wants, at least for several years. Sure, the WTO will rule, but whatever. Winning!

This is not going to get Mexico to pay for the wall

 Well, now we know Trump’s big leverage over Mexico: force people to show ID in order to make money transfers.

Really, that’s it. That’s the big idea. So let’s briefly unpack it.

First, it would take some time. The Trump administration would need to go through the regulatory process. So this “Day 3” bullshit it just that. 

Second, it would then immediately hit the courts under multiple grounds, not least NAFTA. So more delays.

Third, let’s say it happens. Suddenly, undocumented immigrants can’t make money transfers via Western Union. Well, they would immediately move towards other means. And illegal transfer organizations would spring up. Bitcoin? Well, yes, but as this World Bank report from 2005 discusses, also “ethnic stores, travel agencies, moneychangers, hawala-type systems, courier services, hand-delivery, etc.” Some of which would, in 2017, use Bitcoin to actually send the money. 

This would all raise costs for the migrants sending money home, but it would not stop the flow. So the Trump administration would have to go after them. That can be done, but more delays. (Here Carlos will mention that the administration will be facing massive bureaucratic pushback, slowing things further.) 

At this point the Mexican government has been chuckling at Washington’s blundering for about two year. But finally, somehow, the screws go in. Then what?

Well, to figure out what the Mexican government will do, you need to look at Mexican politics. Half of all remittances go to eight states. Two of them are the D.F. and Mexico State. Those are rich places; no elected officials are going to come under pressure if remittances dry up. Of the top-ten states, few households receive remittances: 10% in Michoacán and Guanajuato; 7% in Guerrero. These households would take roughly a 15% hit to their incomes, with some knock-on effects for local economies. (Page 115.) One good counterfactual might be the Great Recession, which put a big brake on remittances from the United States.  This Banxico paper shows the results. They were unpleasant.

But would they be so unpleasant as to get the Mexican government to pay to avoid them? Uh ... no. First, it is not like Mexican governments are bending over backwards to address rural poverty now. Second, even if all the affected households decided to vote against the incumbent government, the swing would be less than 5%. Third, the swing by people not affected will be more, since, you know, they would be the ones paying the taxes to build the border wall they don’t want.

Note that the above completely avoids any Mexican nationalism. Which is, in fact, a thing and means that no government would pay. So we are back to cancelling NAFTA and ... yes, this is in the proposal ... raising visa fees. Because “even a small increase in visa fees would pay for the wall.”

OK. The U.S. issued 1.5 million non-immigrant visas to Mexicans (including border crossing cards) and 82,476 immigrant visas. Let’s say the U.S. issued $17 billion in 30-year bonds to pay for the wall. You could finance the resulting $804 million annual expense by raising visa fees a mere $502 per visa. A small increase, indeed. I mean, it is possible that demand might go down a tad if the modal visa fee went from $160 to $662.

I am wasting my time, obviously.

And Mexico throws down a gauntlet for Drumpf

President Peña: “There have been episodes in human history, unfortunately, where these expressions of strident rhetoric have led to very ominous situations. That how Mussolini got in, that’s how Hitler got in: they took advantage of a situation, a problem perhaps, which humanity was going through at the time, after an economic crisis.” 

It’s pretty clear that the Mexican president would have preferred to have stayed silent, but the Donald is so unpopular south of the border that silence was no longer politically viable. 

Which means one of three things should the GOP frontrunner become the 45th president of the United States. (1) He will publicly lose to Mexico when it comes to paying for the wall. (2) He will repeal NAFTA and use the resulting revenue to claim that Mexico “paid” for the wall. (3) In complete contravention of American law, he will start confiscating wire transfers sent by putative Mexican nationals to their compatriots south of the border and use that to claim that Mexico “paid” for the wall. 

Could repealing NAFTA pay for a border wall?

 It is hard to come up with counterfactual tariff revenue if NAFTA is repealed. You need to go through the U.S. uniform tariff schedule, apply those tariffs to the current duty-free trade with Mexico, and then estimate how much trade volumes would decrease if tariffs were reimposed. 

Here is a back-of-the-envelope estimate. In 1991, tariffs on imports from Mexico came to around $567 million on $39 billion of imports. Imports from Mexico are now around $294 million. That would imply revenues around $4.5 billion. A slightly more (but not much) sophisticated estimate uses the tariff rates from 1991 and adjusts for the share of imports using USITC data. At the two-digit level (which is a blunderbuss) that would get you revenue of $3.9 billion today.  

Of course, the Uruguay Round then cut tariffs by about half, so assuming that Mexico would retain normal trading relations after the U.S. repealed NAFTA, then those revenues would fall to something around $2.0 billion. 

Even if you cut that by half again for falling trade volumes, you get $1 billion per year. Enough for the wall?

One engineer estimated the amount of materials needed. A second estimated the cost. Cut that materials cost in half (for only 1000 miles of wall) and you have $8.5 billion. Double that again for total cost: $17 billion. 

Given low U.S. borrowing costs, the Drumpf administration could plausibly claim that ending NAFTA did indeed force Mexico to pay for that wall, even if some of that tariff incidence fell on American consumers. (Once NAFTA is finished, President Trump could try to get Congress to remove normal trading relations from Mexico and raise tariffs further, or impose taxes on wire transfers from the United States, but that is not something he could legally do on his own.) Over two decades or so, those revenues could finance a wall. If you squint really hard. And Donald sure is squinty-eyed.

The things I think about these days.

The rogue state of Canada?

Who’s the bad boy of international arbitration? Well, the Argentine Republic, of course. And then the Bolivarian Republic. Followed to nobody’s surprise by Ecuador. But the country formerly known as Dominion is in the running: number six most sued according to the United Nations. That is deucedly odd, when you think about it.

What the hell? Canada? Well, Canada is a member of NAFTA, and Chapter 11 of NAFTA protects property rights. So when any level of government in Canada does something that an American or Mexican investor can credibly claim is tantamount to expropriation, they become vulnerable to arbitration.

Now Canada is about to ratify the Washington Convention, which would open up ICSID to its companies. In the way of Canada’s hyperfederalism, the hold-up is with provincial ratification. (In many ways, Canada is the European Union with elections and an army.) Given the country’s experience with NAFTA, there is some worry about the possibility that ICSID will overrule the country’s sovereignty.

These worries are overblown. The Washington Convention does not automatically open the way for foreign investors to go to arbitration. That requires either domestic enabling legislation or an investment treaty. What ICSID does do, is allow companies to enforce their judgments against third countries in Canadian courts. That is something that Canada’s myriad mining and petroleum companies will embrace.

Now, it also means that NAFTA arbitration decisions can be enforced around the globe ... but the fact of the matter is that Canada is a small country right next to the United States, and NAFTA decisions could already be enforced there. Giving French courts the same privilege is not going to change a whole lot. It also removes the ability of national courts to review arbitration decisions, but as a practical matter NAFTA had already removed that. Note that the Ontario Court of Appeal’s recent decision had to do with a case against Mexico.

In short, for those of you who worry about this sort of thing, Canada is already in deep with the big legal web of international property rights protection. (I like to call it the Empire, because it is cool and ironic.) Joining ICSID won’t change anything.

¿Sufragio efectivo?

  Democracy and central banks.  Can they be reconciled?  I argued yes, but it is possible that the Banco de México has a different opinion.

It seems as though somebody at the central bank somehow screwed up quite possibly the most famous saying in Mexican history.  “Sufragio efectivo, no reelección” somehow became “Sufragio electivo y no reelección.”

I am very curious as to what non-elective suffrage would in fact be.

Of course, Porfirio Díaz, who used the slogan in his 1876 campaign, went on to be “re-elected” in 1884, 1888, 1892, 1896, 1900, 1904, and 1910.  Apparently he misplaced the comma the first time around.

Anyway, central banks aren't the only perplexing institution puzzling modern democracies:  the same can be said for investor-state arbitration.  NAFTA has an entire chapter devoted to investor protection, with ultimate enforcement relying on trade sanctions.   Pretty tough stuff!  “Sufragio electivo y no expropriación,” you might say.

Problem is, of course, in defining expropriation.  Mexican tax law currently allows different companies with the same owner to consolidate results.  In other words, losses at one subsidiary can be used to offset profits at a second.  Congress plans to change that, as part of a broader tax hike.  (Yes, Mexico needs to raise taxes.  No, it doesn't need to do so in the middle of a recession.  Don't get me started about Hooverism south of the border.) 

The catch?  Well, the tax hike might be called expropriation under NAFTA.  Now, generally for a case to be made under NAFTA or ICSID rules, creeping expropriation requires both evidence of discrimination and more than one action.  So I doubt that there would be much of a case here. 

Yet ... arbitration law is bizarre.  I must say that the judicialization of investor-state disputes has been a great thing for world peace.  Consider the risks that the Brazilian government has taken in defending its investors abroad!  But it remains confusing and (unlike independent central banks) not-entirely-legitimate in the eyes of the relevant electorates.  At some point, the world will need to find a better way, or the system will collapse under its own weight.

In this case, however, I don't think that a straw will be added to weight bringing down the world's current imperially non-imperial means of dealing with investor-state disputes.  But we shall see.  Any lawyers in the house?

Keep on trucking

  A while ago, I wrote that I was fairly sanguine about the Buy-USA provisions in the stimulus bill. I prognosticated that they would probably not be applied to our NAFTA partners, and even if they were, the NAFTA dispute-resolution mechanism would limit the damage. Now along comes another dispute, this time over trucking, that poses a greater threat to trade ... although I am still confident that it will end well.

What's the story? In 1994, NAFTA passed the U.S. Congress. It mandated that all three countries open up their long-distance trucking markets to the others. The U.S., however, punted when it came to Mexico, claiming that Mexican trucks weren't safe. In 1998, Mexico took the U.S. to NAFTA arbitration.  On February 6, 2001, the U.S. lost ... partially. The panel ruled that the U.S. couldn't issue a blanket ban on Mexican trucks, but that it could assess their safety on a case-by-case basis. The Bush administration pushed Congress to issue enabling legislation that would bring the U.S. into compliance with the ruling, and with the political capital gained from 9-11, finally succeeded in November 2001. (Not everything the Bush Administration did was wrong, although much was botched ... keep reading.)   Congress authorized the FMCSA to inspect Mexican trucks, and the Bush administration issued the required regulations to implement the law. 

Problem solved! Not quite. The Bush regulations had some, well, holes big enough to drive a truck through. More below the fold.

Yet another modest proposal, in which I outrage my friends by advocating protectionism but please Canadians and Mexicans by saying it shouldn't be aimed at them

  The new stimulus bill requires that all steel be purchased domestically. Brad Delong is unhappy:

“If there's little excess capacity in the U.S. steel industry — so that the price of steel is high enough to induce people to look outside for suppliers — then a stimulus won't be much needed. If there's a lot of excess capacity so that a stimulus is needed, then steel customers should be able to bargain prices down to marginal cost — in which case foreign producers will have an extremely difficult time competing on price given that steel is heavy and distances are great.”

Well, maybe. Marginal cost can vary a lot, and there are two steel-producing nations whose companies enjoy almost as good access to American markets as their American competitors: Canada and Mexico. The price of the same product imported from the two countries varies between 14% and 65%.  (December price data below the fold.) Meanwhile, China has achieved significant import penetration, despite prices which exceed Canadian ones by 40% on average, and by 11% in the chief product category China exports to the United States. The implication is that American firms are high cost.

Normally, the fact that U.S. firms are high-cost wouldn't change anything. It would make sense to buy the cheapest steel from wherever. American consumers would save by going to the lowest cost producer, and Americans as a whole would be better off, even if American steel producers would suffer. 

Under depression conditions, however, I'm not sure the above logic still holds. Under depression conditions, using public money to buy from domestic high-cost producers has three advantages that don't mean anything in normal times. First, domestic high-cost producers will spend more of their revenues, not park them as saved profits ... after all, they have high costs! Second, the workers and suppliers of American steel companies (who make up the chief costs of the domestic producers) are far more likely to spend their wages and earnings on American goods and services than their foreign counterparts, keeping the multiplier higher than it would be otherwise. Finally, having steel mills shut down in American towns and American workers thrown out of work will have psychological costs for American consumers and businesses. They could worsen the downturn.  Such costs will be much less if the workers being tossed on the dole are in Ontario or Hebei. In short, when American resources are unemployed and their unemployment looks set to last, buy-USA provisions can make the country better off as a whole.   

More below the fold.