Analysis of Trade with America’s Top Partners Exposes Flaws in Trade Theory

April 1, 2013

An analysis of trade with America’s top fifteen trade partners in 2012 once again reveals a powerful relationship between the population density of its trade partners and its balance of trade, and very little relationship between the balance of trade and the usual suspects blamed for imbalances – low wages and currency exchange rates.

Here are America’s top 15 trade partners in 2012, based upon total imports and exports:  Top 15 Trading Partners in 2012.  These fifteen nations (out of 228 nations in the world) account for 72% of all U.S. trade.  The top three nations – Canada, China and Mexico – account for 43% of all U.S. trade.  Saudi Arabia moved from 12th in 2011 to 9th.  Singapore dropped off the list, replaced by Italy.

It should come as no surprise that a few of our major sources of imported oil appear on the list – Canada, Mexico, Saudi Arabia and Venezuela.  It’s trade in manufactured goods that’s of greater interest, since it’s there that jobs are won or lost.  So let’s see how these nations stack up in terms of trade in manufactured goods.  Here’s the list:  Trade in Manfd Goods with Top 15 Partners.  Aside from China now edging out Canada for the top spot, the list doesn’t look terribly different.

Now, let’s test the results of U.S. trade with these nations against economists theories about trade – that trade deficits tend to be the result of low wages or perhaps currencies that are artificially low, and that trade deficits tend to shrink as wages rise in the surplus nation and as their currency grows stronger, making their exports more expensive and our exports more affordable.  And let’s test these results against my own hypothesis – that it’s actually disparities in population density that drive global trade imbalances while the above-mentioned factors so favored by economists actually have little or no impact.

Population Density

Let’s begin with the latter – the effect of population density – and look at a plot of per capita balance of trade in manufactured goods vs. population density.  (It’s important to express the balance of trade in per capita terms in order to remove the sheer size of a nation as a factor.  Here’s the chart:  Per Capita Balance of Trade vs. Pop Density.  (Some of the data points have been labeled with the nation’s name, some not, for the sake of legibility.)

This is a “scatter chart,” the purpose of which is to determine whether or not a correlation exists.  I had the computer generate and insert a “trend line” for the data, including the equation that defines the line and its “coefficient of determination.”  If such a chart yields a shotgun scattering of the data, then no correlation exists, and the coefficient of determination is close to zero.  On the other hand, if the data points tend to fall along a line – the trend line – then a correlation does exist and if all the data points fall perfectly along the line then we’d have a coefficient of determination of “1” – representing a perfect correlation.

As you can see, the data points do indeed tend to fall along a line – a lined defined by a logarithmic equation with a coefficient of determination of 0.51.  That’s a strong correlation.  Taking a closer look, we find the following:

  • There are four data points (nations) with a population density less than the United States, which is about 86 people per square mile.  They are Canada (10), Saudia Arabia (32), Brazil (61) and Venezuela (83).  The United State enjoys a surplus of trade in manufactured goods with all four of them.
  • There are eleven data points (nations) with a population denisty greater than the United States, and we have a trade deficit with all but one – by far the smallest – the Netherlands.  This isn’t surprising since the Netherlands is barely larger than the tiny city states which fall outside the boundaries of my theory (based on a rather arbitrary cut-off of 1,000 square miles).  They are excluded because cities represent incomplete economies.  They thrive primarily on services and are dependent on the surrounding countryside to complete their economies with resource production and manufacturing.  People who live in cities manufacture relatively little, since they lack the space required for manuacturing facilities.  The U.S. almost uniformly has a surplus of trade with city-states, regardless of their population density.
  • Our biggest surplus of trade in manufactured goods is with the least densely populated nation – Canada.
  • Our worse deficit (in per capita terms) is with Taiwan, which is also the most densely populated nation on the list.
  • Notice that, when expressed in per capita terms, our deficit with China no longer looks so abnormally large.  In fact, it falls right in line where you’d expect to find it.

It’s impossible to overstate the importance of this relationship between trade imbalance and population density.  Accurately predicting a surplus or deficit in 14 out of 15 cases is a very powerful correlation.  It puts into persepctive our very large trade deficit with China.  Of course it’s large; China is a very large country – one fifth of the world’s population.  It’s no wonder that we have a big deficit with China when we applied to them the same trade policy that produced the results we see with Germany, Japan, S. Korea, Taiwan, Mexico and other densely populated countries.  It’s exactly what we should have expected.

Low Wages

It’s impossible to gauge the effect of low wages directly, since the data on wages doesn’t exist and, if it did, it would vary industry-by-industry and even employer-by-employer, making the calculation of an average wage nation-by-nation a nightmare.  But the data is readily available for another factor – “purchasing power parity” (or “PPP”) – essentially a nation’s gross domestic product divided by its population – and it’s a good relative measure of how wages in one nation compare to another.  So let’s plot PPP vs. our per capita balance of trade in manufactured goods:  Per Capita Balance of Trade vs. PPP.

It’s immediately apparent that the data points are not randomly scattered, but tend to form a “V” shape, converging at a zero balance of trade as PPP falls toward zero.  The balance of trade tends to rise upward and outward – in either a positive (trade surplus) or negative (trade deficit) direction as wealth increases.  We already know that those nations on the surplus side (with the exception of the Netherlands) are all nations with population density lower than that of the U.S.  Those on the deficit side are more densely populated nations.

Because the data points fall on both the positive and negative side of the Y-axis, the computer is unable to generate an equation that describes the relationship that seems to be apparent in this chart.  But if we divide the data into two charts, it will be able to tell us the equation and just how strong the correlation may be.  So, first, here’s a chart for the data on the surplus side:  Per Capita Surplus of Trade vs. PPP.   There is a very strong correlation between our trade surplus in manufactured goods and a nation’s wealth.  As we deal with wealthier nations, our trade surplus (if we have a surplus) tends to be larger.  This makes sense.  Wealthier nations, where people earn higher wages, have more disposable income to spend on products both imported and produced domestically.  But, again, it’s important to note the role that population density has played here.  Not only can these people afford to buy more products, but they’re also able to utilize those products because they live in uncrowded conditions that foster high per capita consumption.

It’s also important to note there that, though these are wealthy nations, none are as wealthy (with wages as high) as the United States, with a PPP of $49,800 in 2012.  What this means is that every nation on this chart has a trade deficit with a nation (the U.S.) that actually has higher wages, not lower.  This debunks the notion that low wages cause trade deficits.

Now let’s look at the trade deficit side.  Here’s that side of the chart, with trade deficits now expressed as positive numbers so that a trend line equation can be calculated:  Per Capita Trade Deficit vs. PPP.  What we see here is exactly the same thing – that if we have a trade deficit with any given nation, it will tend to be larger if that nation is a wealthy nation.  While the correlation isn’t as strong – the coefficient of determination is .36 vs. .73 for the surplus nations – there’s still a fairly strong correlation.  Here it’s important to note that every nation on this side of the chart is more densely populated than the U.S. – most of them much more densely populated.  The per capita consumption of these people is stunted by overcrowding, leaving them incapable of consuming enough products to result in a trade surplus for the U.S.  Thus the trade deficit.  But why does the deficit tend to be larger for wealthier, densely populated nations?  These nations have grown wealthy because of their large trade surplus in manufactured goods, not just with the U.S. but with the whole world.  Poorer, densely populated nations are poor because of their overcrowding and because they haven’t been able to elevate their standard of living by manufacturing for export.

So far, it seems we can conclude that low wages don’t necessarily cause trade deficits.  And we can conclude that our trade imbalance (whether its a surplus or deficit) with poor, low wage nations tends to be small, but grows as partner nations become wealthier.  Whether the imbalance is a surplus or deficit seems determined not by wealth and incomes, but by population density relative to the United States.

Economists may argue that those deficits are due to some other factors – currency manipulation perhaps (and we’ll examine that one soon) – but as wages rise, our trade deficit will shrink as our exports become more affordable and their exports become more expensive for us.  Sounds logical, doesn’t it?  Alright, let’s see what the data says.  Let’s begin with a look at how the wealth of these fifteen nations (as measured by PPP) has changed relative to the U.S. since 2001.  U.S. PPP has risen by 38.1% during that period.  So, if a trade partner experiences the same increase in PPP, then their wealth relative to the U.S. hasn’t changed.  If it rises by 48.1%, then the wealth (and wages) in that nation have risen 10% relative to the U.S.  Using that methodology, here’s how the wealth of these nations has changed relative to the U.S. since 2001:  %Change in PPP Relative to U.S..

As you can see, eleven of our top fifteen trade partners experienced faster growing wealth (as measured by PPP) than the U.S., led by China with a growth rate of 210% in excess of the growth rate in the U.S.  On the other hand, four nations – all European nations – experienced a decline in wealth relative to the U.S., led by Italy with a decline of 20%.  If economists are right, then we should see an improvement in our balance of trade with nations that are growing more wealthy relative to the U.S., and a worsening of our trade balance with those nations where wealth (and wages) are declining.  Let’s take a look at the facts.  Here’s a chart that plots that change of wealth since 2001 vs. the change in our per capita balance of trade in manufactured goods:  %Change in PPP vs. %Change in Balance of Trade.

Here we see a shotgun-like scatter of data.  In trying to insert a computer-generated trend line, I got lines sloped in both directions depending on the type of line – linear, exponential, logarithmic and power.  To emphasize the randomness of the data, consider the following:

  • Eleven of these fifteen top trade partners grew in wealth (as measured by PPP) relative to the U.S., led by China with a growth of 210%.  Of these eleven, the U.S. experienced a worsening of its balance of trade with seven of them – the opposite of how economists say it should have responded to rising wages in those nations.
  • Four nations experienced a decline in wealth relative to the U.S., led by Italy with a 20% decline.  (The others are also European nations – the U.K., the Netherlands and France.  Germany was the only European nation among the five nations to experience an increase in wealth relative to the U.S.)
  • Of these four nations that experienced a decline in wealth (and wages) relative to the U.S., our balance of trade worsened with three of them.  It improved with the Netherlands.  This is in line with what economists predict should happen.
  • Overall, our balance of trade responded to changes in wealth among our top fifteen trade partners as economists would predict in only seven instances – less than 50% of the time.

From this data, we can conclude two things regarding the effect of wealth and wages among our trade partners: (1) The imbalance of trade – both surpluses and deficits – will tend to be larger with wealthier nations.  Whether the imbalance is a deficit or surplus has little to do with wages, but is determined by population density.  (2) Over the 12-year span studied, changes in wealth don’t predict which way our balance of trade will change.  Rising wealth is no more likely to improve our balance of trade than it is to erode it.

Currency Exchange Rates

Finally, let’s see what effect changes in currency exchange rates may have played in changing our balance of trade with these top fifteen trade partners.  Economists say that a stronger currency relative to the U.S. dollar should make a nation’s exports more expensive for American consumers and should make American products more affordable for consumers in that nation.  Thus, our balance of trade should improve.  Deficits should get smaller and surpluses should grow.

If we plot this on a bar chart, with two bars representing the percent change in balance of trade in manufactured goods and the percent change in currency, we should see both bars on the same side of the line, if economists are correct.  A positive change in the value of a nation’s currency should correspond with a positive change in our balance of trade with that nation.  So let’s see what really happened.  Here’s the chart:  %Change in Balance of Trade vs % Change in Currency.

Not only do economists seem to be wrong on this issue, the exact opposite seems to be true.  Growth in a nation’s currency exchange rate vs. the dollar is actually far more likely to correspond with a worsening of our balance of trade – not an improvement.  As you can see, economists’ prediction held true with only four nations – India, Canada, Brazil and the Netherlands.  In the case of India, a slightly weaker currency corresponds with a huge increase in our trade deficit.  In the other three cases, a strengthening of the currency corresponds with an improvement in our balance of trade.   In a 5th case – Saudi Arabia – our balance of trade in manufactured goods improved dramatically while the currency exchange rate held steady.  (The Saudi currency is pegged to the dollar.)  With ten of these fifteen nations, the change in our balance of trade was exactly the opposite of what economists predict.  The most blatant example is Venezuela.  In spite of their currency devaluing by 495% since 2001, our balance of trade in manufactured goods with them actually improved by 184%!  In the case of Italy, the Euro rose by 31% but our balance of trade with Italy worsened by 42%.  In the case of China, their currency rose by 24% while our balance of trade with them worsened by 319%!

Once again, economists are wrong and have the cause and effect relationship backwards.  Instead of currency rates affecting the balance of trade, what appears to be happening is that the balance of trade affects currency exchange rates.  If a nation has a trade surplus with the U.S., their currency strengthens.  If a nation has a trade deficit with the U.S., its currency tends to weaken.


An analysis of trade between the U.S. and its top fifteen trade partners, accounting for 72% of all American exports and imports, proves that the balance of trade in manufactured goods is determined by the population density of the nation in question.  Almost without fail, America experiences a trade deficit with nations more densely populated, and a trade surplus with nations less densely populated.

Claims that low wages cause trade deficits are false.  Our worst trade deficits are with densely populated, wealthy nations.  Densely populated nations that build their economies on manufacturing for export experience growth in their wealth and wages.

Claims that low currency values cause trade deficits are also clearly false.  The cause and effect is just the opposite.  A trade deficit with a particular nation tends to drive the value of that nation’s currency higher.  A trade surplus tends to drive that nation’s currency value down.

Those who claim that if we’re just patient enough, rising incomes and currencies will reverse our trade deficits, haven’t tested their theories against actual data.

Because the majority of the world’s population lives in densely populated conditions, the U.S. is doomed to a massive trade deficit in manufactured goods and a loss of manufacturing jobs as long as it places its faith in flawed free trade theory that fails to account for the role of population density in driving trade imbalances

Brazilian Company Awarded U.S. Defense Contract

January 17, 2012

As reported in the above-linked article on Reuters this morning, Brazilian aircraft manufacturer Embraer has been awarded a $355 million contract to provide small turbo-prop aircraft to the U.S. military. 

I’m using this report to illustrate how such trade deals would work under the trade policy I proposed in Five Short Blasts.  Those of you not familiar with that policy might assume that I, a real hawk on restoring a balance of trade, would be opposed to such a deal.  After all, here we have the federal government, led by a President who speaks incessantly about the need to create jobs, awarding a contract to a foreign company.  How can this be?  We should all be outraged!

Those who know me better know differently.  Under the trade policy I’ve proposed – one in which tariffs on manufactured products are indexed to population denisty, but one in which all other trade is essentially “free trade” – there would be no problem with this deal (assuming that there were no bidding irregularities).  With a relatively sparse population, and being rich in natural resources, Brazil is a perfect example of the kind of country we should engage in free trade.  And, in fact, Brazil is one of our best trade partners.  In 2010, our trade surplus in manufactured goods with Brazil soared to over $20 billion, shattering the record of $17 billion set in 2008. 

As one commenter to the Reuters article pointed out, these aircraft may actually be assembled at Embraer’s new plant in Melbourne, Florida.  But it’s likely that many, if not most parts, will be imported from Brazil and other countries.  But all of that is beside the point.  The point is that Brazil isn’t the problem when it comes to our trade deficit.  Neither is any other sparsely populated country. 

The real problem is attempting to trade freely with the likes of Japan, Germany, China, Mexico and other densely populated nations, desperate to find work for their badly bloated labor forces.  In November, the latest month for which trade data is available, we had a trade deficit of $20.9 billion.  Our trade deficit with China alone was far larger  – $26.9 billion.  Throw in Japan, Germany and Mexico and those three alone account for another $16.4 billion trade deficit.  Restore a balance of trade with those four nations and we swing to a $22.4 billion trade surplus.  That’s a swing of $43.3 billion per month, or $520 billion per year.  Assuming that 2/3 of that is labor, and assuming an average wage of about $50,000 per year, that’s 7 million high-paying manufacturing jobs.  And here I’ve limited this analysis to only four densely-populated countries in the interest of brevity.  I could have included two dozen more on the list and the numbers become even more staggering. 

As tempting as it is to malign free trade in general and to be critical of all trade deals like the one illustrated in this report, that would make no more sense than the blind application of free trade to every situation like we have now.  We don’t need to eliminate trade.  We need to apply trade policy that is aligned with economic realities, especially the inverse relationship between population density and per capita consumption, and its role in driving global trade imbalances.  We need to make smart use of tariffs where it makes sense and apply free trade principles where those make sense.

Current trade policy – confining ourselves to the extreme free trade end of the trade policy spectrum – makes no sense at all, and the results bear that out.

Why Obama Courts Brazil and Latin America

March 22, 2011

The above-linked article reports on Obama’s trip to Latin America and his call for a strengthened partnership with that region.  Why the emphasis on Latin America?  Why was he in Brazil just as military operations were being launched against Libya?  Here’s why:

Brazil Trade

Our trade surplus with Brazil, driven by an even larger surplus in manufactured goods, is rising fast.  This chart would look virtually the same for every country in South America.  Obama correctly sees Brazil and Latin America as key to his objective of doubling exports in five years to rebuild the manufacturing sector of our economy. 

What Obama doesn’t understand is why.  Why are we so successful in trade with these countries and with others, like Canada and Australia,  and such an abysmal failure in places like China, Japan and Germany, just to name a few?  We are successful in South America for the very same reason that China, Japan and Germany have trade surpluses with us and for the very same reason that Japan has a trade surplus with China.  We’re more densely populated than South American countries (and Canada and Australia), just as China is more densely populated than us, and Japan is more densely populated that China. 

It’s all about population density and its role in suppressing per capita consumption.  When nations share labor forces and markets through free trade arrangements, the results become lopsided when one has an over-sized labor force and a market stunted by over-crowding and low per capita consumption.

I’ve recently begun my annual updating of trade results for 2010 between the U.S. and each trading partner.  2010 saw a rebound in trade volumes from the recession-affected trough of 2009.  And, in virtually every case, what I’m finding is that our balance of trade in manufactured goods improved with less densely populated nations while it worsened with those more densely populated.  And, I might add, this is in spite of the fact that the dollar fell relentlessly in 2010.  It also holds true regardless of whether the nation in question is wealthy or poor.  (In other words, low wages aren’t a factor.)  I’ll share some selected charts with you for some of the more significant trade partners as I compile the data.  (I’m doing it alphabetically, so Brazil was among the first.) 

Obama would be well-advised to return home and focus not on our trade relations with Latin America.  Nothing he says or does there can work to America’s advantage any more than the disparity in population density does automatically.  Rather, he should focus his efforts on revising trade policy to stem the tide of imports from the overpopulated, export-dependent economies of the world.

Weaker Dollar No Solution to Joblessness

November 18, 2010

As I was postig the results of my study of currency exchange rates and trade imbalances – a study which showed that changes in currency exchange rate have no effect on trade – the above-linked article titled “Weaker Dollar Seen as Unlikely to Cure Joblessness” appeared on CNBC. 

The article begins by re-stating the now disproven economic theory:

A weakening currency traditionally helps a country raise its exports and create more jobs for its workers.

The article then goes on to make the case that a falling dollar won’t have this predicted effect.  I can add one more reason – the theory is flawed and not supported by the data.  A weaker dollar will have no effect on the prices of imports, as exporting nations will simply cut costs and subsidize their industries to hold the line on price, assuring that they maintain their share of the U.S. market.

Rising prices for imports certainly would bring manufacturing jobs back to the U.S.  But there’s only one way to make sure they rise sufficiently, and that’s for the U.S. to set the prices.  In other words, the U.S. needs to apply tariffs.  Tariffs must be applied to manufactured products, and the size of the tariffs must be proportional to the population density of the country of origin.  This would result in big tariffs on all products from China, even bigger tariffs on products from Germany, Japan and South Korea, but would leave products from nations like Canada, Australia, Saudi Arabia, Brazil and a host of others completely free of tariffs. 

The only problem we’d have then would be building factories fast enough to keep pace with demand.

Effect of Exchange Rates on Trade with Brazil and Colombia

July 25, 2010

Continuing our series examining the relationship between currency exchange rates and balance of trade, we now take a look at trade between the U.S. and two South American countries:  Brazil and Colombia.  In 2009, Brazil was America’s 11th largest trading partner.  And, given Colombia’s small size, they’re no slouch either.  So how has the balance of trade between the U.S. and these two countries been affected by the exchange rate between the dollar, the Brazilian real and the Colombian peso?  Has a stronger dollar led to declining trade balances, or a falling dollar resulted in improved trade balances for the U.S., as economists would predict?

Here are the charts.

$US-BRL Rate vs Balance of Trade $US-COP Rate vs Balance of Trade

We’re starting to see a pattern here.  These charts are virtually identical, and both are virtually the same as the charts for Australia and Canada that we looked at earlier.  There’s a good correlation between exchange rate and trade balance.  When the dollar rises, our trade balance gets worse and when it falls, the trade balance improves.

So much for my claim that population density disparities is what drives trade imbalances and not exchange rates, right?  Not so fast.  Once again, although the value of the dollar vs. the real has affected the balance of trade with Brazil, the balance has consistently been in favor of the U.S., as the population density disparity theory would predict.

In the case of Colombia, a nation slightly more densely populated than the U.S., they’re an oil exporter, leaving them flush with American dollars to spend on American manufactured products.  (Not to mention their illicit drug trade dollars.)

What we need here is a method for keeping score, so I devised the following spreadsheet:

Theory Correlation Score

For each country we’ve examined so far, I went year-by-year to see if the balance of trade followed the exchange rate theory.  That is, did our balance of trade improve in those years in which the dollar fell from the previous year, and did it worsen in those years in which the dollar rose?  I then totaled the number of years in which the trade balance behaved as predicted by the exchange rate, and divided by the total number of years.  For example, in the case of Canada, the trade balance behaved as predicted in 12 years out of 19, a 64% correlation.  (See the notes below the spreadsheet explaining the methodology.)

As you can see, for each of the less densely populated nations we’ve examined so far – Australia, Canada, Brazil and Colombia (where the population density ratio is less than or close to 1.0), there has been either a positive or strongly positive correlation between the balance of trade with the U.S. and the rise and fall of the exchange rate.  But in the case of China, more than four times as densely populated as the U.S., there is a strongly negative correlation.  That is, the U.S. balance of trade worsens as the dollar weakens.

In every case except Colombia, America’s balance of trade has been exactly as predicted by the population density disparity theory.  In the case of Colombia, where their population density is nearly identical to that of the U.S., the population density disparity theory barely applies at all, since there is very little disparity.  Nevertheless, since the U.S. is slightly less densely populated than Colombia but enjoys a surplus of trade in manufactured goods, I’ve scored it as not correlating with my theory.

All in all, this has been pretty inconclusive so far.  But stay tuned.  Next up is Japan.


Exchange rate data provided by