I am not a macro person (by nature at least- I don’t deal well with severe empirical limitations and unanswered questions I guess). That said, I enjoy this profusely:
What I think I don’t like about basic macroeconomics is that I feel like we (I mean instructors) don’t always do a great job explaining why things work the way that they do. For example, we introduce the concept of gross domestic product, or GDP (but even then are kind of murky on how goods with imported components get counted), and we give the “real” version of GDP a variable, namely Y. We then say that Y can represent aggregate production, expenditure, or income. Ok great- I guess it has to be true that the amount people spend on our stuff has to equal our income, but it would be nice to point that out explicitly. What is less obvious is why it must be true that the amount of stuff we produce has to equal the amount that people spend on the stuff we make- sure, that should be true in equilibrium, but what is stopping an economy from producing a whole bunch of stuff that goes into inventory? As it turns out, the extra output is counted as purchased by the company that made it, so we’re sort of forcing that part of the equality by redefining expenditure a bit. Sure, why not.
Then we get to the “expenditure categories of GDP”– you know, the thing in the bed:
Y = C + I + G + NX = C + I + G + (X – IM)
You see this all the time in macro, and what it means is that the spending on an economy’s output can be broken down into
- Consumption spending (C) – spending by households in an economy
- Investment spending (I) – spending by (mostly) businesses in an economy on stuff that makes more stuff
- Government purchases (G) – spending by the government
- Net Exports (NX, or X – IM) – the difference between foreign spending on domestic goods (exports, X) and domestic spending on foreign goods (imports, IM)
If you’re anything like me, this all makes perfect sense until you get to the net exports part, and then you’re like wait, what? Allow me to summarize a discussion that I think should happen in the classroom much more…
If we think about the different ways that stuff can be produced and consumed, we get something like this:
Since GDP, by definition, represents domestic production (regardless of where stuff is consumed), the area that should count in GDP looks like this:
But let’s think about the other GDP categories for a second- consumption, for example. Consumption represents the purchases by domestic households (other than newly constructed houses, technically speaking), and, if you’re anything like me, some of what you consume is produced in the U.S. and some of it is imported. As a result, the consumption area looks like this:
Therefore, the GDP identity needs to have a correction factor to turn the domestic consumption area (and, by similar logic to some degree, investment and government spending) into the domestic production area. Looking at the picture, it becomes pretty clear that we can do so by adding in exports and subtracting out imports. Funny thing- this is exactly what the net exports category of expenditure does!
Hopefully that helps the expenditure identity actually make sense as opposed to something you just memorize and try not to think too hard about. But it also highlights an important point- taking away imports, in and of itself, doesn’t increase GDP. Now, I get why people might think that, since looking at the basic Y = C + I + G + (X – IM), it certainly seems like Y goes up if you take away the thing that is subtracted out. The problem with this logic, as the pictures above illustrate, is that the IM part is just a correction factor, and you can’t take away a correction factor without also taking away the thing that you’re correcting! In other words, if you’re going to take away IM, you have to reduce, well, mainly C, and maybe some I and G, by a corresponding amount, at least in an accounting sense.
As a result, whenever anyone tells you that limiting or eliminating imports will increase GDP, they are making hidden assumptions about consumption (mainly along the lines that people will just buy domestic stuff instead and nothing else will change) that generally fall under the heading of assuming the conclusion. They are also potentially ignoring the fact that such an increase may not actually increase households’ standard of living if it makes their consumption decrease. (Taking away $100 of my imported stuff isn’t going to magically generate $100 of just as cool stuff for me to purchase from domestic producers- if this were true, I probably wouldn’t have been buying imported stuff in the first place.)
With me so far? Great, you’re farther along than Trump’s economic advisers in an important way. In reality, there are interconnections between the expenditure components that are not shown in the basic Y = C + I + G + NX identity, and these interconnections make it so that you can’t just look at this simple equation to analyze cause and effect.
Here, Greg Mankiw explains it a bit better, after mentioning that he agrees with Paul Krugman on the matter:
But of course you can’t model an economy just using the national income accounts identity. Even a freshman at the end of ec 10 knows that trade deficits go hand in hand with capital inflows. So an end to the trade deficit means an end to the capital inflow, which would affect interest rates, which in turn influence consumption and investment.
I suppose that their calculations might make sense in the simplest Keynesian Cross model, in which investment is exogenously fixed and consumption only depends on income. But that is surely not the right model for analyzing the impact of trade policy over the course of a decade.
(Mankiw provides more detail, but you have to acknowledge that Krugman wins the headline game.) I find it funny that people make it such a big deal when Mankiw and Krugman agree on anything…I mean, they agree on lots of stuff, namely basically everything in their respective textbooks. (Related: I know people who won’t use one of said textbooks bc of bias or whatever, and I find it hilarious since they are functionally identical for the most part.)