The "it doesn't matter" perspective

Why lots of stuff matters less than you might expect

Clearly lots of things matter, lots of things are important. But economics provides a different way of looking at the world, which often leads to the conclusion that things don’t matter in the way that common sense might suggest.

Readers may already know much of the following as individual concepts, but perhaps you don’t yet see the big picture. Indeed, is there a big picture? Is there a unified way of thinking about why so many things don’t matter? I’m not sure. But it might be fun to see if we can find one. I’ll group 12 examples into 5 general categories:

Zero long run competitive rate of profit:

Students often get confused when principles textbooks describe a zero long run rate of profit in competitive industries. In those examples, economists are referring to something known as “economic profit”, which includes the opportunity cost of capital as a part of the cost of production.

It might be easier to visualize zero economic profits as a “normal rate of accounting profit.” Thus in highly competitive industries with no barriers to entry, the normal rate of accounting profit might be 12%. If profits rise above that level for an extended period, new firms enter and that drives profits back down. The reverse happens if profits fall below the normal level. This assumption has profound implications for a wide range of issues, including price gouging, ATM fees, and mandatory employee benefits.

#1: Consider a simple six period example for price gouging. Assume that every 6th day there is a shortage of a good such as bottled water. Sellers have to decide whether to price gouge when there is a shortage, or to continue charging the normal price. We can think of two pricing options:

Price gouging: $4, $4, $4, $4, $4, $10

No price gouging:  $5, $5, $5, $5, $5, $5

The exact figures I’ve chosen are arbitrary, but the key point is that due to the assumption of competition and free entry into retailing, we assume zero economic profits in the long run. This means that firms experience normal accounting profits in the long run. In my example, water sellers need to receive revenue equal to $30 for the total sales over 6 days, for each 6 bottles sold. If they are not allowed to price gouge on days when there are shortages, then they need to charge more during normal times.

People often find these sorts of examples to be implausible, but I suspect that’s because they misunderstand the purpose of economic models. The point of an economic model is not to explain how real people in the business world make decisions—I’m quite certain that almost no real world retailer thinks of things in precisely this way. Rather the point is to explain how underlying economic principles such as competition lead (through a sort of “invisible hand”) to certain outcomes. If the retailer that avoids price gouging doesn’t charge $5 every day, he or she eventually goes broke.

Most people don’t think deeply enough about the implications of our models. Thus even some opponents of price gouging will concede too much to the other side. They might say, “Yes, the poor might benefit from a cap on prices when price gouging occurs, but it’s not worth the efficiency cost of misallocation of goods and shortages.” In fact, the poor will pay the exact same $30 for water over a 6-day period, regardless of whether price gouging is or is not allowed.

Mistaken reasoning on these sorts of problems often occurs because people forget that we are dealing with a repeat game. Yes, if this only happened once, then a price ceiling might have the distributional effects its proponents hoped for. But if the past 250 years of the US economy is not a repeat game, then nothing is. Do you really believe that entrepreneurs in highly statist economies like Argentina don’t understand the regulatory environment they’ve been dealing with since Juan Peron took power in the 1940s? Do you really believe that business people don’t price their goods in a way that reflects the hassle of dealing with all of those statist regulations?

#2: ATM fees are another good example. Would bank depositors benefit from a government regulation that capped ATM fees at one dollar? I suspect if you asked them, many would say yes. In fact, banking is a relatively competitive industry, with several thousand banks serving depositors in the US. One way or another, bank profits would be maintained in the long run, as banks increased other fees (or reduced interest on deposits) enough to maintain the same long run rate of economic profit.

Someone could quibble that this isn’t exactly right; bank profits might slightly decline in the long run. I agree, but only because regulation is not a zero sum game. A cap on ATM fees would cause banks to offer fewer ATMs, reducing real GDP. Banks would be a bit poorer, as would the non-bank public.

#3: Governments occasionally mandate certain employee benefits. Yet economic theory suggests that total employee compensation is determined by the supply and demand for labor. Thus if you mandate something like health insurance, employers will respond by offering a lower salary. This doesn’t necessarily mean that mandated benefits are a bad idea, rather it suggests they are a good idea if employees benefit more from something like $10,000 in mandated health insurance than they would from an extra $10,000 in cash income.

At first glance, all employee benefit mandates look like a bad idea. If workers would have preferred the benefit to the extra cash income, why weren’t the employers already offering that benefit? Again, this perspective doesn’t necessarily mean mandated benefits are bad, rather you need to find some sort of deeper justification. Thus the government might mandate benefits for distributional reasons, or because they fear non-insured employers would show up at public hospitals and demand treatment that they could not afford. These are complex issues, beyond the scope of this post.

International Trade:

International trade is another area rich in cognitive illusions. (My second choice for a title for my new blog was “The Many Illusions”.)

#4: You might think that international trade causes more unemployment, as workers are displaced by imports. In fact, jobs lost due to imports are offset by jobs gained in other sectors, such as exports, investment and the production of consumer goods and services. Even a big trade deficit doesn’t result in higher unemployment, which is one reason why Europe has far more unemployment than the US, despite a massive trade surplus.

#5: I’ve seen other pundits make all sorts of arguments against tariffs, claiming that they will indirectly hurt our firms by driving up the cost of inputs such as auto parts, or that other countries will retaliate. Those claims may be true, but there’s a much simpler explanation for why tariffs don’t have the impact their proponents expect; our current account balance reflects the gap between domestic saving and domestic investment (which is a negative number in the US.)

A tariff that raised a great deal of revenue could conceivably reduce the deficit by raising national saving. But the sort of populist politicians that favor tariffs also tend to run massive budget deficits, so it’s unlikely that any plausible US tariff policy would actually reduce the trade deficit. For example, the deficit did not improve as a result of the Trump/Biden tariffs. Tariffs tend to appreciate the domestic currency, offsetting any competitive advantage to domestic firms from higher taxes on imports.

Macroeconomics:

In a previous post, I argued that the famous GDP = C + I + G + (X - M) equation caused all sorts of mischief. It led some people to wrongly assume that imports reduce GDP, and it led others to wrongly conclude that government spending raises GDP.

#6: As with tariffs, there are very limited cases where there might be a grain of truth in the claim that fiscal stimulus raises output. Thus if the government spends a lot of money on wasteful projects like building pyramids in the desert, this can make the country poorer. Workers respond by working harder, boosting measured GDP (but not actual living standards.) Most real world fiscal stimulus, however, is merely tax cuts and transfers, and the effects on aggregate demand are largely offset by monetary policymakers that target inflation at 2%.

The only sense in which this sort of fiscal policy matters at all is the degree to which it makes our long run fiscal regime more or less efficient. The ideal path of budget deficits (for any given level of spending) is produced by a path of tax revenue that minimizes the long run deadweight loss from taxes. Thus you want to elect leaders that are wise and far-sighted, who care more about future generations than about their current political popularity. (OK, you can stop laughing now.)

#7: Money illusion is another area where things matter much less than one might imagine. A one-time, exogenous 10% increase in the money supply will tend to raise all wages, prices and asset values by 10%. Because some wages and prices are sticky, however, this sort of monetary shock really will matter in the short run. But because it matters in the short run, many pundits forget that it doesn’t necessarily matter in the long run, as all nominal variables adjust by an equal amount to the money supply increase.

I’ve seen pundits call for an expansionary monetary policy to generate higher wages, which makes no sense if by “wages” you mean real hourly wages. When I have called for easier money (say in 2009), it’s been precisely because I believed it would lead to lower real hourly wages, more employment, and more total real income in the short run. In a deeply depressed economy, monetary stimulus will initially cause prices to rise a bit faster than wages, but the number of total hours worked will increase as well, resulting in more total real income.

#8: Do supply shocks cause higher inflation? Much less than most people believe. There are two arguments worth considering here, a simple argument that is popular with conservatives and a more subtle long run argument. Both have some value.

The simple (right wing) argument is that supply shocks such as an oil embargo cause a change in the relative price of a single good, not the overall (absolute) price level. Many Keynesian economists push back against this claim, pointing to the fact that monetary policymakers given a “dual mandate” will often partially accommodate the supply shock, allowing a temporary period of above normal inflation.

While the Keynesian critique has some merit, at a practical level its importance is wildly overstated. To take a recent example, over the past 5 years the US has been hit by both positive and negative supply shocks. The positive supply shocks have been slightly larger than the negative supply shocks, and as a result real GDP growth has been a bit above trend. So it is approximately true that roughly zero percent of the total cumulative inflation since late 2019 was caused by supply shocks. Supply shocks boosted inflation in 2021-22, and depressed it even more in 2023-24.

That conclusion is pretty clear to anyone looking at recent growth rates of prices, real GDP and nominal GDP. As in 2008-09, “The real problem was nominal.” And yet you’d be hard pressed to find a single Keynesian economist that agrees with me on this point. In most cases, they are fooled by the fact that supply shocks are more newsworthy, and thus they don’t notice all the positive supply-side developments that held inflation down. In a recent Econlog post, I pointed out that real oil prices have changed very little from 50 years earlier, and hence energy explains roughly 0% of the total inflation we’ve experienced since 1974.

Compensating differentials:

#9: Is it better to have a safe job than a dangerous job? Not necessarily. According to the theory of compensating differentials, risky jobs must pay higher salaries to attract workers. Thus a worker who welds I-beams on the 60th floor of a skyscraper makes more than a worker who builds one-story ranch houses. You many not think that West Virginia coal miners earn much money, but they earn more than blue-collar workers in less dangerous jobs in West Virginia.

#10: Compensating differentials also impact asset prices. A house located next to a noisy airport costs less that a similar house located in a quiet neighborhood. This sort of reasoning underlies the Efficient Markets Hypothesis, which suggests that asset prices incorporate all publicly available information about the value of an asset. Thus the observation that Nvidia has a brighter future than US Steel does not in any way suggest that Nvidia is a better investment than US Steel from this point forward. The “dartboard theory” of stock picking is a classic example of the “it doesn’t matter” mindset.

Other examples:

#11: There are two ways of getting a Harvard education. You can pay $60,000/year in tuition, or you can get it for free by auditing lots of classes and sitting in the back of the room. Lots of students want the expensive version, whereas almost no one wants the free version. This insight underlies many of the “signaling theories” of education. (Bryan Caplan has a book entitled “The Case Against Education.”)

I recently read a book about China entitled Other Rivers, written by Peter Hessler. I was struck by the fact that China’s extremely rapid modernization of 1978-2012 was accomplished by workers with a pretty low level of education. And even those who had gone to college often ended up working in areas totally unrelated to their studies. (To be clear, China’s population has recently become much more highly educated; I’m speaking of an earlier period in their history. I’ll do a post reviewing the Hessler book)

To be clear, some level of formal education is clearly useful. A country benefits when its general population can read, write and do basic arithmetic. And you need a few specialists with more advanced knowledge. But at the margin? I’m just not sure it matters all that much. Chinese workers seem to have been quickly trained by their employers. Some of the most useful skills described by Hessler were self taught or learned on the job.

Robin Hanson has made a similar argument about health care. Obviously you’d want to have a system that can treat acute injuries, do an appendectomy and provide antibiotics and vaccines. But Hanson argues that most health care is complete waste, and indeed some is positively harmful.

#12: The replication crisis suggests another example of where things matter less than you might assume. Because academic journals are biased in favor of papers that reject the null hypothesis, researchers keep torturing the data until they can find a “statistically significant” result. In the end, research in many fields becomes little more than an empty game, and many published papers fail to be replicated. When you read a news article about study suggesting that X affects Y, you are probably better off assuming the finding is spurious, especially if it is from a field like social science or nutrition.

Summary: Is there a common theme to all of these examples? Probably not. But in many of these examples, certain concepts keep reappearing. For instance, people move to places with the best opportunity. Thus even though California has a minimum wage of $20 for fast food workers, while Texas has a $7.25 minimum wage for hamburger flippers, you would not expect California fast food workers to be better off than those in Texas. If they were, workers would move from Texas to California.

In many examples above, either firms adjust or people adjust or prices adjust to reach some sort of “equilibrium”. If you try to move to a new equilibrium with regulatory changes, your policy might well be frustrated by the response of the public. Or if you subsidize things that the public would otherwise not purchase, the subsidy might have surprisingly little impact on people’s welfare.

But the “nothing matters” perspective goes well beyond cases of markets reacting to compensate for some sort of policy shift. I also suspect that for aggregate outcomes, all sorts of individual policies matter less than one might think. For instance, one viewpoint on which many Democrats and Republican seem to agree is that elections are very important for the future direction of the economy. But are they? The entire stock market response to the recent election could be explained by a single issue—lower corporate tax rates under a new GOP administration. Who is president might matter for the broader economy, but how could we ever know for sure?

Does having a big welfare state matter? Maybe, but it is surprisingly hard to know for certain. One could compare Sweden, where government spending is 47.3% of GDP, to the US, which spends 36.3%, but is that the right comparison? Sweden has about 10 million people, lots of immigration, and was neutral during the two world wars. By those criteria, the most similar country is Switzerland, where the government spends only 31.5% of GDP. If you compare those two affluent European countries, then I see little evidence that social spending has much impact on things like poverty rates. It might, it’s just hard to know for sure.

My baseline assumption is that most economic policy issues are 99% about efficiency and 1% about distribution. That’s because I see the world as mostly being about issues such as the price-gouging example discussed at the top of the post. You can interfere in the free market as much as you wish, just don’t expect it to matter very much for income distribution.

PS. Other possible examples:

1. Sanctions often don’t seem to have much effect.

2. Affordable housing mandates don’t make housing more affordable.

3. After more than a century of fighting the drug war, the drug problem appears to be worse than in 1895 (when drugs were legal.)

4. Recent abortion bans don’t seem to have reduced abortion.

5. Canada bans the payment of money for blood, but gives Canadian patients blood provided by paid donors in the US.

6. Taxes on sellers or buyers have the same impact. Ditto for taxes on imports and exports.

7. Low wages don’t give a country a competitive advantage, as they reflect lower productivity.

8. Except for Covid, real shocks have little impact on the US business cycle.

9. Changes in interest rates often have little effect, especially if they mirror changes in the natural rate of interest (as in 2022.)

10. And when it comes to “nothing matters”, who can forget JM Keynes’ famous remark:

In the long run we are all dead

In this print by Yoshitoshi, the warrior plays a flute, in preparation for committing suicide: