Expectations Matter
By now we’ve gone through several rounds of the Trump tariff craze. We’ve seen the markets drop and rise like rollercoasters, but many have missed some interesting nuances of what has transpired.
First of all, a brief glance at markets gives us a curious picture of what occurred. As Trump began speaking and argued for “reciprocal” tariffs, the stock market had a slight increase. This is because markets had already priced in the possibility of something like 10% across-the-board tariffs. That’s what they were already expecting. Reciprocal tariffs would actually be a much more pro-free-trade system when compared to a flat 10% rate, so the markets began to rise. It was only when Trump announced his understanding of reciprocality that markets crashed.
In other words, expectations matter. If markets expect you to behave like a dumb-dumb and you do something that’s only mildly stupid, then stock prices will increase. Likewise, if they think of you as a tech genius and you demonstrate that you don’t understand SQL, they will fall.
This is true for a wide variety of economic variables. Will inflation fall (relative to the counterfactual) if the Fed increases interest rates? Depends on what the market expected the Fed to do to begin with! Will inflation raise unemployment as the Phillips Curve teaches us? Only if the inflation was unexpected.
After you begin indexing to expectations, things never really look quite the same.
Causation Does not Imply Correlation
I’m sure you understand that correlation does not imply causation, but do you understand that causation does not imply correlation? One variable could be causing another, but you may never find it in the data!
Here’s a simple example: imagine yourself driving on a road with a speed limit of 60mph. Suppose further that you’re in a bit of a rush and are quite a good driver, so you’ll make sure to maintain 60mph constantly. What happens if you come across an uphill portion in the road? You’ll step on the gas to maintain 60mph. And what will happen if you find the car going downhill? You’ll take your foot off the gas and might even apply the brakes to stop yourself from violating the speed limit.
Now imagine if a rogue statistician were to stumble upon a dataset showing when you pressed the gas pedal and the velocity of your vehicle. He would discover that as gas usage varied, velocity remained constant. There was no correlation. The statistician concludes that neither gas nor brake pedals affect speed. The next time he sits down to drive a car, he dies.
The morale of the story is that causation may sometimes exist without a single hint of it in the data. In fact, causation may cause no correlation. A fun macroeconomics factoid is that though the money supply has a very strong correlation with inflation over longer timeframes in different countries, there is little to no correlation in countries that adopt inflation targeting.
It makes sense if you think about it. There’s different shocks to the macroeconomy. Central banks adjust monetary policy to make sure they hit their 2% inflation target. This leads to variation in the money supply, but no variation in inflation. Thus, we are left with causation and no correlation.
Income vs Substitution Effects
People often get into arguments when a certain change induces two different effects that push in opposite directions. A common example in economics is when the price of a good changes. Will an increase in the price of water increase or decrease the demand for iced tea? On the one hand, people may start substituting iced tea for water, leading to an increase. On the other hand, when they buy bottled water, they’ll be spending more money and have less left over for iced tea, leading to a decrease in demand. The final change in demand is unclear.
This former phenomenon is known as a “substitution” effect, while the latter is often referred to as an “income” or “wealth” effect. They are difficult to unsee after you’ve spotted them once. Will an increase in wages decrease or increase the amount of time people spend working? Well, they’ll earn more money per hour now, so that should lead them to work more. But they can also make the same amount by working less time now, which will induce them to work less. The exact effect is, again, unclear.
And as we get past the point of pure intellectual curiosity, income and substitution effects have massive implications for policy. Would raising the earned income tax credit be a pro-work or anti-work policy? Could taxes actually induce people to work more by making them poorer? Tough luck, you’ll have to dive into a trove of messed up econ papers to find out the answer. Though my general advice would be to assume that substitution effects dominate.
Are Savings Good or Bad for the Economy
When it comes to economic growth, people have two stories in their heads.
One says that savings is bad for the economy. If you save all your money, then you’re not giving it to firms. If you’re not giving it to firms, they can’t pay it out to workers. If they can’t pay it out, they fire workers. Those fired workers then have no money to spend. The cycle continues and you get a Great Depression. That is the strawman version of Keynesianism many appear to believe (it’s akin to characterizing Marxism as “time was invented by clock companies to sell more clocks”).
The second story says that society’s wealth comes from production. Production broadly depends on capital, labor, and technology. Savings affects the level of capital. When you save money, you are in some direct or indirect way investing in capital. Your bank loans out your money to businesses, or you buy stocks directly. Either way, you refusing to spend money on consumption today leads to the creation of more factories, tractors, or whatever form of capital your country makes the best use of. This in turn leads to economic growth as we use these tools to make more goods in the future.
So which story is right? Mostly the second one. Long-run economic growth is indeed related to savings. Savings also explains a large chunk of what we know as the “Asian growth miracles”. They show up in most economic growth models and truly do matter for long-run prosperity.
The first story is a fairly vulgar retelling of Keynesianism. It is true that under some specific circumstances spending drives economic growth. Indeed, if you were to spend as much as possible during the Great Depression, that would spur economic activity. But that is not true under normal conditions (or even the abnormal conditions of the highly uncertain Trump economy).
And though there are many explanations of why the first story is false, I’ll fall back on a monetarist perspective: spending can raise production, but it also raises prices. This will lead to the Fed raising interest rates to stop inflation. Therefore, spending will not result in higher growth, it will instead lead to higher interest rates which in turn induce savings. For spending to matter, the Fed must be trying to get inflation higher and failing. This was true in 2009, it is not true today. It is also why most “stimulus” bills will fail to be stimulative now that we have gone past the era of zero interest rates.