What I Didn't Know About Macro

I have a solid understanding of microeconomics, but I don't know much about macroeconomics.

Last month I read 24 of the 38 chapters in Tabarrok and Cowen's Modern Principles of Economics to remedy this. Here's the most important stuff I learned.


You compute GDP by adding up a nation's final products. That includes consumer spending (George buys a car or dinner at a fancy restaurant), businesses investing (Toyota buys factory robots), government spending (the US military buys a new airplane), and exports (Qatar Airlines buys a new American plane). You then subtract out anything imported (Toyota's robots, if they're foreign-made).

An upside of this GDP calculation: it's invariant to companies splitting apart. Toyota splitting off its door manufacturing factory into a separate company and then buying those doors shouldn't (and doesn't) increase GDP.

A downside of this GDP calculation: it's sensitive to whether trades are priced. If I paint your fence and you paint my roof, we contribute \$0 to GDP. If you pay me \$200 to paint your fence, and then I pay you \$200 to paint my roof, we've contributed \$400 to GDP.

A memorable framing (reproduced by GPT):

Two economists are walking through a field when they discover a pile of horse dung. The first economist turns to the second and says, "I'll pay you \$100 to eat that." The second economist thinks it over, decides that a hundred bucks is a hundred bucks, and eats the dung. He collects his money, and they continue on their way.

After a while, they come across another pile of horse dung. The second economist turns to the first and says, "Now, I'll pay YOU \$100 to eat that." The first economist pauses, considers the proposition, then eats the dung. He gets his \$100 and they continue walking.

After they've walked a bit further, the second economist turns to the first and says, "You know, we've both eaten horse dung, and yet neither of us is any richer than before." The first economist nods and says, "True, but you're missing the bigger picture. We've increased the GDP by \$200."

Wage differences for equivalent jobs

Why do US truck drivers make more than Nigerian truck drivers, even when they do the same job? Because the US economy has more capital per worker. If your truck drivers have large new trucks (lots of capital), each driver moves more goods and produces more value, and you can justify hiring drivers at a higher marginal wage. If your drivers have small old trucks (low capital), your drivers produce much less value each, and so you can't justify paying as much for the marginal one.

Why does the US have more capital per worker than Nigeria? T & C claim it's because the US has better institutions, and therefore attracts more investment. For example there's less incentive to buy nice trucks if the government might nationalize your business and take them at any time. Strong and stable rule of law enables investment, which yield high capital per worker, which yield high wages.

Predictable Inflation is OK

Unpredictable inflation (and deflation) is bad, whereas predictable inflation (and deflation) is mostly [0] ok. Consider 5% predictable inflation.

  • Your Ubers become 5% more expensive, but your salary increases by 5%. You anticipate this and so buy just as many Ubers.
  • The principle of any loans your bank has written decreases by 5%, but they anticipated this and so charged you more interest when they initially wrote the loan.
  • The government (who printed the additional money) effectively taxes any cash you have. This tax is very hard to avoid; even kids and drug smugglers pay it. The government anticipated this and so decreases other taxes to fund the same budget.
  • The effective capital gains tax rate increases, since capital gains are paid on nominal earnings. The government anticipated this and set the nominal capital gains rate somewhat lower.

None of these seem that bad. Basically, everyone anticipates the effects and the market quickly equilibrates. Perfectly expected deflation should equilibrate the same way.

Now consider an unpredictable 5% increase in inflation (you thought the inflation rate would be 5%, but it turned out to be 10%).

  • Your business does \$115k against expectations of \$110k. You perceive excess demand and increase supply. This involves ordering more stuff from your suppliers, who now face real increased demand in addition to inflation-hallucinated over-demand.
  • Your bank loses money on its loans, whose repayment is worth less than they expected. If this happens too much, your bank gets angry and stops writing long-term loans.
  • You get a get-out-of-jail free card on your mortgage.

The case for unpredictable deflation is even worse:

  • Your business did \$100k in revenue this year, but did only \$105k this year against expectations of \$110k. You perceive a lack of demand and cut supply. This includes ordering less material from your suppliers, who now face real under-demand in addition to inflation-hallucinated under-demand.
  • You agreed to give your workers a 10% raise, but now money is tight and you need to either make pay cuts or do layoffs.
  • The burden of any debt you owe increases, since you originally priced it to compensate for 10% expected inflation.

As you can see, unlike predictable inflation, unexpected inflation causes confusion. You can't tell whether a change in revenue is due to a change in demand or just inflation, and so it's hard to shift supply correctly in response.

Additionally, volatility in inflation makes it harder to write long-term loans, since banks don't want to assume the risk of huge changed in the inflation rate over the term of the loan.

Some economists (Milton Friedman) think we should stick the inflation rate at 3% and keep it there. This is highly predictable and corresponds with the natural growth rate of the economy, so the value of a dollar would be stable year-to-year.


Events that temporarily change demand for something are called "shocks."

There are two fundamentally kinds of shocks. In a real shock, something happens that increases or decreases the underlying productivity of a sector. Examples include earthquakes, technology breakthroughs, change in gas cost, new regulation or deregulation, and pandemics. In an aggregate demand shock, demand goes up or down, but fundamental productivity is unchanged. Examples include consumers feeling fearful and spending less, or businesses feeling confident and spending more.

T & C list five specific ways these shocks can spread through the economy. My previous understanding was something like "everyone gets scared," but it turns out that these mechanisms are rational. Here they are:

  • Uncertainty about the future economy may cause you not to lock-in long term investments, like building office space.
  • Banks are more finicky about lending to businesses who don't have lots of spare money. Without a supply of capital it's hard to get loans.
  • If you feel your business is already nearly underwater, you might take huge risks or otherwise be irresponsible, since there's potential upside and no downside.
  • It's more profitable to invest when others are investing, and nobody wants to invest when nobody else is investing. It's better to build office space when companies are hiring and new startups are getting founded.
  • If it's hard to get money now, you want to save your marginal dollar for risk aversion instead of investing it.

There are two government responses to shocks: fiscal policy (increases or decreases in taxes and government spending) and monetary policy (the fed printing money or changing the discount rate).

Fiscal Policy

Fiscal policy is when the government tries to increase (or decrease) spending, usually to compensate for a shock. It generally does this by cutting taxes (so consumers have more to spend), or by spending itself. Fiscal policy often works, but it has a few notable failure modes.

The first failure mode is "crowding out." In this case, the government spends in place of private firms, and so doesn't increase total spending at all. For example, if the government hires 100 factory workers to build widgets, and they all quit GE to work for the government, total employment hasn't changed at all. The government has "crowded out GE." Likewise, if the government taxes away the money that GE was spending on tooling and then spends it on building a military factory, the government hasn't actually increased total spend.

The second is more subtle. Aggressive fiscal policy can fail catastrophically if the government spends so much that it's at risk of default. In this case, people try to sell their dollars to buy foreign currencies, further driving down the price of dollars. Then the government is at higher risk of default, on and on.

Fiscal policy works well after an aggregate demand shock. The government can replace spending from scared businesses or consumers and get the economy back on track. It works less well after real shocks; a small amount of government spending can't really encourage construction projects when gas is at five dollars a gallon.

Monetary Policy

In the US, the Fed manages monetary policy. The Fed can do two major things: buy and sell bonds via open market operations and lend money at the discount window.

The idea behind open market operations is that the Fed wants to put money into the economy. They do that by buying stuff. They could buy MacBooks, but instead they buy treasury bonds from banks like Goldman. Now Goldman has more cash and fewer treasury bonds, and is incentivized to lend that cash to make a profit.

Usually the government does this until banks have so much money that they're willing to lend it to each other at a target "federal funds" rate. This is the interest rate you hear quoted in the news, but don't be confused: the fed manipulates it only indirectly. The mechanism is similar to how you or I would set the price of carrots to \$10/pound by buying carrots until the price equilibrates there.

If you're wondering, quantitative easing is closely related to open market operations. QE is like the turbocharged version of open market operations, where the fed buys not only short-term treasury bonds, but also long-term treasury bonds, corporate bonds, stocks, etc. The government might be forced into QE if it wants to stimulate the economy, but the funds rate is already at zero. (Some countries, like Japan, choose to push interest rates negative instead.)

The other thing the fed can do is set the rate that they lend to banks at the discount window. It spooks the market if you take a loan from the discount window (is your bank going to collapse?), so everyone tries not to use it. That means that usually the discount rate doesn't matter much. But the fact that banks can borrow from the discount window means that banks are less likely to get hosed when they lend to other banks, and so more willing to lend.

If you're curious, SVB was apparently trying to get a loan from the discount window, but went into receivership before the paperwork went through.

Et Cetera

  • Ideas are underproduced, because the inventor can't capture their full value. Once the idea is out there, other people can free ride on it.
  • Argentina used to be one of the richest countries in the world!
  • Ricardian Equivalence: rational taxpayers know that the government sending them cheques in the mail this year means higher taxes sometime in the future, and might just save the current stimulus to pay off that future cost. If all taxpayers behave like this, cash infusions won't stimulate the economy at all.
  • Exchange rate pegs are super dangerous. You peg 100 rupees to the dollar. When your economy looks weak and your peg looks suspect, everyone wants 1 dollar instead of their "equally valuable" 100 rupees. Speculators sell rupees short and buy dollars. Your government spends dollars to buy up rupees and maintain the peg. Eventually your government runs out of dollars, and the peg collapses. This is how George Soros got rich; he effectively sold tons of pounds and broke a peg to the Deutsche Mark.
  • "Who ultimately pays a tax does not depend on who writes the check." FDIC insurance is widely lauded as being a tax on the banks, not the taxpayer. But since Americans are mostly inelastic consumers of banking services, the price of banking goes up slightly to cover the cost, and people with bank accounts pay in the end. On the other hand, in situations where the taxed good is elastic (e.g. has close untaxed competitors), the sellers will decrease their prices to absorb the tax and remain competitive. In general, whether the buyer or seller pays a tax goes with the elasticity of the good. This concept is called tax incidence.
  • Elasticity is escape. The more elastic party is the one with more escape options (e.g. a close alternative, or they can forgo the item).
  • Best-price guarantees ("we'll beat our closest competitor's price by 10%") are consumer-hostile. Normally, you want your business to have lower prices than your competitors, so consumers buy from you. But if you have a best-price guarantee, this dynamic shifts. Now you want your business to have higher prices that your competitors, so customers are incentivized to use your best-price guarantee to buy your product at slightly under your competitor's price.


[0] "Mostly" because you can imagine problems even with enough predictable inflation (people might not want to save) or deflation (people might always want to wait for tomorrow to buy stuff, when prices are cheaper).