You can see this machinery at work in Gregory Mankiw’s latest opinion piece for the New York Times, which instructs us, “Don’t Worry About the Trade Deficit”. It patiently explains that the money leaving the country on the current account (via trade) equals the money entering the country on the capital account (via financial flows), and that the latter drives the former. The key paragraph reads:
The trade deficit is inextricably linked to this capital inflow. When foreigners decide to move their assets into the United States, they have to convert their local currencies into American dollars. As they supply foreign currencies and demand dollars in the markets for currency exchange, they cause the dollar to appreciate. A stronger dollar makes American exports more expensive and imports cheapers, which in turn pushes the trade balance toward deficit.So here we have a theory of the relationship between international capital flows and trade: a shift in net financial inflows causes an appreciation of the dollar, which then causes a greater trade deficit. Where does this theory come from?
The ultimate source is the national income and product identity, augmented by the accounting relationship between savings and consumption. From NIPA we have
Y ≡ C + I + G + NX
where Y is national income, C is consumption, I is private investment, G government spending (purchases of goods and services by government) and NX is the trade balance. That “≡” sign means “is identical to”, which is subtly different from simple equality, “=”. With equality, the stuff on the left hand side has the same value as the stuff on the right hand side. With identity, the stuff on the left hand side is the same stuff as on the right hand side, just expressed differently. If A = B, A and B could be different but for some reason aren’t. If A ≡ B, A and B are two different ways of writing the same thing, so it is impossible for them to be different, even for an instant.
Now let’s add a second identity that represents the truism that all income is either consumed, publicly or privately, or saved or paid in taxes:
Y ≡ C + S + T
with S being savings and T being taxes. Putting them together, we see that
C + I + G + NX ≡ C + S + T
Subtracting C from both sides and rearranging gives us
NX ≡ (S-I) + (T-G)
Read this as “the trade balance (net exports) is identical to the sum of net savings (savings minus investment) and the government’s budget surplus or deficit (taxes minus spending).” This is not a theory of what causes what; it’s an identity, different ways of expressing the same “what”.
How can this be? An important source of identity relationships in economics is the two-sided nature of economic transactions. If I pay you $10 for something, this can be seen as $10 of spending (mine) as well as $10 of income (yours). My spending is not equal to your income; it is your income: one transaction, two ways of recording it. Add up these identities over a whole economy, subdivide income or spending into different categories, and you get a set of macroeconomic identities.
So now consider what an export or import means. An export signifies the receipt of income by people in the exporting country but not the corresponding spending. An import signifies spending without the corresponding income. So a trade deficit has to take the form of national spending exceeding income, which means net borrowing. This borrowing can occur either in the private sector (less saving relative to investment) or in the public sector (government borrowing). And net borrowing on a national scale is what we mean by the capital account, so there we have it.
All of this is in the realm of identity. There is no causation at work here. Net exports does not cause net borrowing, nor is it the other way around; they are identical to one another, different ways of recording the same economic events. Any measured discrepancy would be a problem of measurement, not a problem of whether the identity is true. (This is an oversimplification, since it leaves out transfers, asset incomes and currency flows. The main adjustment would be to add transfer payments and net asset income to the current account. A further adjustment would need to be made for changes in holdings of currency, the reserve account, which is a non-borrowing method of finance. That would add more algebraic symbols, but it wouldn’t alter the essential logic.)
So now reread the above paragraph from Mankiw. This is telling a story of causation over time, not identity. First we have increased inflows of finance, then the dollar appreciates, then the trade deficit widens. The first process could be virtually instantaneous, because currency markets are comprised of exactly these demands for and supplies of foreign exchange. The second process, however, takes place over months, depending on the shape of the J-curve that depicts the change in trade balances over time in response to an exchange rate shock. But we know from the identities that the trade balance is identitical to net borrowing at each moment; that’s what it means to be an identity.
The only valid theory of trade deficits and capital inflows would be one that looked at all the determinants of this identity simultaneously. It would have to consider the factors that determine what goods a country produces, what it exports and imports, the terms of trade, differences in international growth rates and the demand and supply of financial assets. One way of looking at the current account-capital account identity does not cause the other; the international position of the US economy, which can be measured either way, is caused by the balance of all the influences on trade and finance. And yes, that includes the choices made by firms in the US to increase or offshore their productive capacity, along with everything else. The relative importance of these factors is an empirical question, not a theoretical one.
The flaws in Mankiw’s analysis do not come from inaccurate data or a faulty assumption here or there, but a basic misunderstanding of Econ 101. The guy needs to take an intro class. Of course, if the textbook isn't very good it won't help him much.