The Broken Window Fallacy and Stimulus from Destruction, War, and Natural Disasters

Whenever you hear of some destruction from natural disasters or war, you’ll always find some people (like Paul Krugman of NY Times) praising the events for stimulating economic growth. Their argument is that, because of the destruction of all these homes and buildings, people have to rebuild them, and this causes people to spend and circulate money and it creates jobs for lots of people. They tend to think that such destruction is actually good for the “economy”. This is actually a somewhat common myth, and it has been debunked already long ago by economist Frederick Bastiat in 1850. Before I summarize his argument, pause for a moment to reflect on the utter stupidity of the implications. If such destruction was a good thing, then all we would have to do to gain economic prosperity is break all of our stuff and burn down all of our houses. If we all did that, do you really think we would be better off economically? Besides this intuitive hint of stupidity, let’s get more technical.

Bastiat illustrates the concept with the story of the broken window. Consider a shop keeper. And, consider a boy that comes up and carelessly throws a rock threw the window and breaks it. Now, the upset shop keeper has to spend $100 to replace the window. This money goes to the window maker, who now gains $100. So, the window maker is better off. The destruction seems to have helped sustain a job. After all, without the breaking of windows, the window maker would have less income. Furthermore, the window maker will then spend the $100 somewhere else, again “stimulating the economy”. Fair enough. This is the part that is easily seen. But, we should also consider the unseen. If the window did NOT break, then the shop keeper would simply have spent that same $100 somewhere else, maybe to get a suit from the tailor. In this case, the tailor would be helped, and he would then spend the $100 somewhere else, again “stimulating the economy”. There really isn’t much of a difference if you’re just counting money flows. $100 flow either way. Breaking the window simply benefited the window maker at the expense of the tailor.

But, that’s not all. The net effect is not zero sum. Breaking the window didn’t just divert money flows to a different direction. The destruction of the window actually made people poorer on a net basis. In the first scenario, with the destruction, the end effect is that we still have a window (because it was replaced), but no suit. However, in the second scenario, without destruction, we would still have the original window, but we would have a suit in addition. The choice is between having only a window or having both a window and a suit. The net effect of breaking the window is to have one less suit. The “point” of the “economy” is to produce goods and services. That is wealth. When you destroy goods, you decrease wealth and the quality of life for people. Destruction is not good for the economy.

The illusion of destruction-caused stimulus is a short term positive effect with long term negative effects. Mass destruction forces us to spend the money from savings (or borrowing and going in debt) all at once instead of spread out spending over time. Any short term boost in spending now to replace destroyed items is offset by less spending later on new goods and services. The net effect is significant harm and less goods and services.

4 thoughts on “The Broken Window Fallacy and Stimulus from Destruction, War, and Natural Disasters

  1. Gross savings is around 18% in the US, so it’s not fair to say the shop keeper would simply spend his $100 elsewhere. On average he would spend $82 elsewhere. Saving that money may stimulate the economy so other way… perhaps the shopkeeper would hire help, perhaps not. But you can’t argue that it all evens out.

    Also, you’re only considering goods (windows and suits) and have failed to consider services, which make up 77% of our GDP.

    • Hi Stranger,

      I see that your argument is basically that I may not be correct when I say that the money flow is merely diverted to different channels. You’re saying that the total amount of money flow (regardless of where it’s flowing) is likely to be lower under the nondestructive case because the shopkeeper may save some money instead of immediately spend it. I will offer several distinct explanations, and on different levels, on why your argument is not valid.

      If you recall, the first part of my post focused just on money flows as the sole measure of how to judge the case. I made the point that the only difference between the two cases was which channels the money flowed through, and that the same flow occurs in both cases, and therefore the same amount of “stimulus” occurs. However, the focus on money flows is not how you judge the benefits of economic situations. Focusing on money alone will certainly lead to false nonsensical conclusions. The only reason I focused on money flows was to temporarily place myself on the same playing field that most other people (who do not study economics) play on. I’ll explain this more later. But, for now, let’s pretend that we should judge things based on money flows. If so, I still think your point may be lacking. After all, even if someone is saving money, the whole point of saving is to spend it later. So, the money will still be spent, just later. The “stimulus” is just shifted to later in time. The money isn’t disappearing. That’s one thing to consider, but this will not be my main argument. None of that even matters. Forget money flows. Even if you have the perspective that the nondestructive case has less money flows, this does not matter AT ALL. I’ll now back up and explain how to properly judge economic matters, and the nature of wealth.

      I remember you said (in other conversations) that “Economies are stimulated by spending”, and I believe that you were implying that this is how to gain prosperity for an economy. Not really. Firstly, “wealth” (in the economic sense, not spiritual) is NOT money, for money is useless by itself. We only want money to buy the things that we need and want. Money is only the intermediate medium of exchange, exchange of the real things we want. The real things that we want are products and services to satisfy needs and desires. Therefore, the definition of wealth is actually well-being, satisfaction of needs and desires, pleasures, comfort, happiness, etc. “Following the money” sounds like a wise practice, but it’s actually unwise because it only tells a small portion of the story. Money is NOTHING except the temporary, intermediate, medium of exchange that we use to effectively trade actual wealth, actual products and services. If money flows are what make us prosperous, then this would mean that we could gain prosperity simply by printing mountains of money and giving $1 billion to each person. This will increase prices through inflation and cause money flows to increase, but it obviously won’t work to make peoples’ lives really better. So, if you want to judge economic cases, the proper way to do it is to determine which case has the most wealth, i.e. the most products and services that are maximally satisfying peoples’ needs and desires.

      With this in mind, let’s revisit our thought experiment. Remember what I said: “Breaking the window didn’t just divert money flows to a different direction. The destruction of the window actually made people poorer on a net basis. In the first scenario, with the destruction, the end effect is that we still have a window (because it was replaced), but no suit. However, in the second scenario, without destruction, we would still have the original window, but we would have a suit in addition. The choice is between having only a window or having both a window and a suit. The net effect of breaking the window is to have one less suit. The “point” of the “economy” is to produce goods and services. That is wealth. When you destroy goods, you decrease wealth and the quality of life for people.” So, as I’ve already said, by the proper standard of measuring real wealth (regardless of irrelevant money flows), the nondestructive case has more real wealth for people. Therefore, it is better for human wellbeing and prosperity.

      I will elaborate further by also accounting for your complaint about savings. Let’s say I accept your case that the shop keeper does not spend all the money, but keeps a portion. Sure. Instead of buying a $100 suit, he buys a pair of $82 shoes. He saves the rest so he can spend that at some later time (or give to his kids, which will spend it eventually). Even if I focus only on the $82 shoes, and even if I pretend imagine that the $18 disappears forever as if it were burned with fire and it never ever gets spent, let’s again compare the two cases. In the first scenario, with the destruction, the end effect is that we still have a window (because it was replaced), but no shoes. However, in the second scenario, without destruction, we would still have the original window, but we would have a pair of shoes in addition. The choice is between having only a window or having both a window and shoes. The net effect of breaking the window is that society has one less pair of shoes. When you destroy goods, you decrease wealth and the quality of life for people. So, regardless of savings, my argument still holds.

      Also, I ask you to realize the ridiculous implications of your beliefs. If destruction was a good thing, then all we would have to do to gain economic prosperity is break all of our stuff and burn down all of our houses. If we all did that, do you really think we would be better off economically? Really, do you think that? Because that’s what your belief demands.

      Your belief of stimulus spending is misapplied. I guess you’re thinking of how govt spending stimulates the economy during a crisis, and you are trying to apply it here. It may be the case (although, it’s debatable) that stimulus spending may help during emergency economic crises when there is some temporary mass scare in the markets. But, this does not carry over to normal nonemergency cases. Even Keynes himself (the guy who brought us that stimulus theory) sees such spending only as a temporary primer pump to jolt the scared and stagnant market (during a depression or recession) back to normal. Even Keynes would refute you.

      Lastly, and continuing on the lesson that real wealth is satisfaction of needs and desires, you must realize that saving money is also a way to create wealth. After all, saving money is useful, prudent, wise, and satisfies needs, namely the need to buy a very expensive product or service in the future (which they are saving money for), or the need for security and peace of mind in case of a “rainy day”. But, even if it’s saved for a rainy day and no rainy day comes, the old man will eventually spend it on products and services, or his kids will. Now, let’s look at the two cases again. In the first scenario, with the destruction, the end effect is that we still have a window (because it was replaced), but no shoes or peace of mind wealth from savings. However, in the second scenario, without destruction, we would still have the original window, but we would have a pair of shoes and peace of mind in addition. The choice is between having only a window or having all of a window, shoes, and peace of mind. The net effect of breaking the window is that society has one less pair of shoes and also less peace of mind. Or replace peace of mind with “ability to buy something very expensive in the future”. So, once I add this part of wealth addition from savings to the analysis, it makes my point even stronger than before, even though technically my argument can stand even if I don’t consider this last part.

      Lastly, lastly, you say that I forgot to account for services. I don’t understand your point about that. Sure, I used products to demonstrate my point because products are easier to think about. But, I technically could set up this same thought experiment using services as the example instead of products, although it would be a little more awkward to set up. Explain to me what I’m missing. It’s as silly as saying, “Well you only accounted for windows and suits, but you forgot to account for houses and cars.” My example used products, but the general principle is applicable to all wealth and trade.

  2. I brought up services to challenge your argument about increased wealth. I understand that you were simplifying the ideas for the sake of argument, but I think it’s worth exploring. You made the comparison between having only a window in one case, and a window and suit in the other. Easy to see how one is greater wealth.

    Let’s say that instead of purchasing a suit, I paid a lawn care service to mow my lawn. In a few weeks time, after the lawn has been cut a second time, I have a non-broken window and a mowed lawn in either scenario. Has wealth increased? Perhaps I’m more fulfilled in my desire to have a well maintained yard, but I doubt any economists are arguing that natural disasters increase happiness.

    That doesn’t make a case for disasters stimulating the economy, but it challenges the idea that the sum of wealth after destruction is an appropriate measure here.

    Measured wealth isn’t what we’re after. It is more telling to know, in a month or year with a natural disaster, is GDP increased from what it would have been without it. Of course saving money and building wealth is good, but we’re not interested in the economy when my kid goes to college, or the economy when I die in 50 years and I pass my savings along to my kid, and he buys a house or a suit. We’re interested in whether or not a broken window or some natural disaster marks an increase in GDP, right now. Generally speaking, money spent now is more stimulating to the economy than money not spent.

    My belief does not demand the implications you’ve outlined. Pruning bushes promotes growth, chopping it down at the trunk does not. Reductio ad absurdum. But this underscores an interesting point…

    It would seem that you find it hard to consider that a healthy economy might not be good for wealth. As you know, I have socialist leanings, and would argue that a healthy free market does *not* create wealth in the sense of happiness and well-being. I’ll quote Thomas Piketty, “a market economy based on private property, if left to itself, […] contains powerful forces of divergence, which are potentially threatening to democratic societies and to the values of social justice on which they are based.” Free markets concentrate wealth and lead to wealth inequality.

    Disasters can be good for the economy and bad for wealth, for people.

    • As I said before, changing the analogy to include services doesn’t change the underlying point. Sure, I’ll go with your modification and consider lawn mowing for the two cases. Let’s assume that each lawn mowing is roughly $33 and it’s mowed every two weeks, so his $100 can pay for 6 weeks of lawn mowing. In the first scenario, with the destruction, the end effect is that we still have a window (because it was replaced), but really tall and ugly grass that causes him (and his neighbors) unhappiness for 6 weeks. However, in the second scenario, without destruction, we would still have the original window, but we would have, in addition, a beautiful lawn that gives the shopkeeper happiness for 6 weeks. The choice is between having only a window or having both a window and a beautiful lawn that provides happiness for 6 weeks. The net effect of breaking the window is to have less happiness, and therefore less wealth. My argument is just as fine as before. If you want, you can take 18% and save it, but this also doesn’t affect my argument, as I already demonstrated with the shoe-instead-of-suit modification in my previous response.

      You say, “[He has] a non-broken window and a mowed lawn in either scenario.” This is false. I don’t know why you said that. In the destruction case, he has a nonmowed tall grass lawn because his money went to replacing the window instead of paying for lawn services. However, as I’m trying to anticipate your reaction to the above, maybe you will say that, “Tom, but he can likely still afford lawn services even if he had to pay for replacing a window. He’s probably not at the very edge of not affording basic things. So, even if the window broke and he had to pay to replace it, he still pays for lawn service as usual and has a cut lawn, and therefore a cut lawn in both cases.” Is this what you were thinking? If so, you’re missing a glaring problem in that reasoning. Even if he replaced the window and still has other money to pay for lawn services, this only pushes the problem back one step to something else. It doesn’t get rid of the problem, it only pushes it back to somewhere else. Instead of having no money for the lawn service (let’s say he still pays for lawn service), he may now have no money for going to the theater. Since it is certain that he now has less money in the destructive case (since he had to spend it to replace the window), this means he must forego purchasing some other product or service, somewhere, whatever it may be.

      You say, “I doubt any economists are arguing that natural disasters increase happiness. […] That doesn’t make a case for disasters stimulating the economy,…” OK. Now I’m really confused as to what your conclusion is. It sounds like you’re agreeing with me, maybe. Are you saying that you agree with me that destruction does not bring economic prosperity? Or, do you disagree with me and do you believe that destruction brings economic prosperity, and so you welcome it as good?

      You say, “Measured wealth isn’t what we’re after.” False. This IS what we are after. How could you possibly say this? Given the definition of wealth that we’re using, it’s like you’re saying, “We shouldn’t judge the goodness of economic outcomes by how much it helps people, allows them to live happier lives, and satisfies people’s needs and desires.” On the contrary, yes, we should. The only reason why you favor stimulus spending and think GDP is good (which I’m not necessarily disagreeing with totally) is not for their own sake in themselves, but because you think these things will better people’s lives and lead to the satisfaction of their needs and desires. So, yes, the maximization of wealth for people is exactly how to judge the goodness of economic outcomes (if we are using utilitarian standards, which is what matters to you).

      You say, “It is more telling to know, in a month or year with a natural disaster, is GDP increased from what it would have been without it. […]We’re interested in whether or not a broken window or some natural disaster marks an increase in GDP, right now. Generally speaking, money spent now is more stimulating to the economy than money not spent.” This is false. As I just explained, wealth is what matters to people. You can’t eat GDP or money, but you can eat food, and so food is wealth. I already explained that money flows themselves are not important in and of themselves, and that money itself is not wealth. Money is just the temporary intermediate medium of exchange to trade real wealth. I understanding what you’re thinking. You think GDP is a measure of how much wealth is being generated for people to better their lives, but this shows that what you are really after is wealth, since that’s what you’re ultimately trying to measure. It’s true that GDP is indeed roughly correlated with wealth generation, under *normal* market conditions, but that’s all it does. If you go and read a standard (not super libertarian) college economics textbook, you’ll find that GDP is simply a rough indicator that may be appropriate under assumed normal market conditions, and that it can sometimes be misleading. The only reason why it’s used is because it’s too difficult (maybe impossible) to measure true wealth on a society level, so we try to use some other surrogate “indicator” of wealth generation. Also, GDP may be more necessary if we’re talking macroeconomic measurements because it’s easier to determine from data and it’s too difficult to measure true wellbeing and wealth at such large scales of millions of people. But, if you’re thinking of a small scenario of a few people, you no longer have to rely on GDP because small scenarios afford us the easier opportunity to look at the true wealth creations up close. So, for our discussion, we can look directly at wealth because we are close enough to see it. I’m not saying GDP is completely invalid, it’s pretty good, but we have to see what’s behind the numbers. You only use GDP when you can’t measure the real thing called wealth. In this small thought experiment of the broken window, why use GDP as an indirect surrogate indicator when you have the real thing right in front of you?

      The above discussion of GDP should be sufficient and stands on its own. However, I’ll try to come up with two examples to help illustrate how GDP has limitations and can give false impressions (besides the huge inflation example I already gave you).

      Example 1. Imagine that two people named Bastiat and Hayek want to do a little exercise just to demonstrate the limitations of GDP. Bastiat has a car worth $10,000. Hayek wants to buy. They exchange. But, they decide that they will sell the car back to each other every single day. So, the next day Bastiat buys the car back, and the next day Hayek buys the car back. They do this until an entire year passes and Hayek does the final buying and gets the car. The above case will be called Case 1. In Case 2, imagine that they didn’t do this weird thing and they did the normal thing of exchanging just once so that Hayek buys the car only once, but then Bastiat puts his earned $10,000 under his mattress for the rest of the year so that he would only spend it next year, once this year’s national statistics are already completed. In both cases, the end effect is exactly the same in terms of wealth and how satisfied people’s needs are: Hayek gets the car and Bastiat gets $10,000 that he can spend on other goods starting the next year. Ignore taxes for the sake of simplicity. We should judge that both cases produce the same economic wellbeing and practical benefits of helping people. But, the two cases have completely different money flows and GDPs. In Case 1, every day $10,000 flowed and changed hands and therefore the national GDP is $3,640,000 ($10,000x(365-1days)) more so than the GDP produced from Case 2. If GDP accurately measured economic well-being under all conditions, the GDP should be the same in each case, but it’s not. If you were just looking at national GDP data, you’d probably think that Case 2 is very much better than Case 1. But, it’s not. They are the same on account of what matters.

      Example 2: Imagine there’s a company called Sand-O-Lot. They sell sand. Then, the govt passes a law that grants them a monopoly on sand, the govt also fixes the price of sand at $5,000 per bucket, and the govt also requires every person in the US to spend 75% of all their income and 100% of all their savings on buckets of sand from Sand-O-Lot. People are allowed to spend the other 25% on whatever they want. If you have savings, you have to take all your savings out and buy sand. In return, Sand-O-Lot is required to spend all its revenue to repurchase those same buckets from all the customers. What’s more, this must happen every single week. Huge cash flows and many many buckets of sand simply trade hands back and forth between the general populace and Sand-O-Lot. They go back and forth every single week. That’s a lot of money flowing and trading hands. If this happened, the GDP would skyrocket to extremely high levels. Money flows would be very high. Stimulus would be tremendously high, especially since savings were being depleted in order to go into circulation. Incomes would also be very high because Sand-O-Lot has to pay the populace to buy back the sand buckets and that counts as revenue income to the populous. According to your theory, all this stimulus and GDP and much higher incomes would be good for people. But, people would be miserable with all this worthless sand and the fact that they can’t spend most of their money on things they need and want. They would not be wealthy or happy. All that money flow does nothing to satisfy needs.

      You say, “Of course saving money and building wealth is good, but we’re not interested in the economy when my kid goes to college, or the economy when I die in 50 years and I pass my savings along to my kid, and he buys a house or a suit. We’re interested in whether or not a broken window or some natural disaster marks an increase in GDP, right now.” Technically, I already made my point and don’t have to respond to this, but I will. This is arbitrary and very prejudice of you. You’re saying, “Something is good if it benefits a specific group of people, even if it harms a different group of people.” One of the most common economics fallacies is to only look at the effects of something on a certain group of people, instead of on all people, and to only look at the effects of something right now, and ignore long term consequences. You are committing these fallacies. You’re saying that we should only think of “now” and ignore long term effects. You’re saying that the group of people that matters is the present group, and not the future group. This is prejudice against the future group. Should we applaud a thing if it helps us now, yet hurts us an equal or greater amount in the future, or hurts our children? Even if I place myself on the “stimulus and GDP itself are the goal” playing field (which is an illogical playing field), your comment still doesn’t make sense. If the destructive case causes an extra $18 more of money flow today (because the $18 is no longer saved for the future, but spent now), then a future society (likely including yourself) will have $18 less of stimulus and money flow. So, no net benefit over the long run.

      You say, “My belief does not demand the implications you’ve outlined. Pruning bushes promotes growth, chopping it down at the trunk does not. Reductio ad absurdum.” I disagree. I still think your belief demands them. The specific nature of the bush allows pruning to be beneficial. It’s unique nature causes the degree of cutting to have a different type of effect depending on degree. But, why should I believe that destruction to property has an analogous behavior? Maybe the destruction-economy situation is more like a bucket of 20 apples, to where you lose welfare if even one apple rots, and so you want all apples to be healthy and able to be eaten. The relationship between economic wellbeing and degree of destruction may be monotonic (in math) instead of non-monotonic. Also, even if I accept your bush analogy, it still leads to absurdity. Since pruning destroys only a small part of the bush to make it better in the long run, your analogy would imply that we should have a lottery where we randomly select 3% of all homes, cars, and other property, and intentionally burn down and destroy this 3% every year, and that this would be a pathway to prosperity. Still seems absurd. Also, why knock Reductio ad Absurdums? RADs don’t always work, and one must use them wisely, but they are useful and important when used well. RADs are particularly good for monotonic relationships, and so I think they may apply here.

      You said some broad stuff about free markets being not that great. We won’t talk about this at the moment because it takes much more time and argumentation and study of history and economics to get to that broad point, and right now we’re still just on one of the early steps toward understanding economics. I’m aware of Piketty’s silly ideas, too.

      Lastly, why are you fighting this post’s conclusion so hard? You won’t lose anything sacred by agreeing with me on this particular post. By agreeing on this post, it does not mean that libertarians and capitalism win. You can still be even a socialist and agree with this stuff. Don’t worry so much. This wasn’t meant to be a controversial post, it was meant to be a simple post that everyone can agree on.

Leave a Reply

Your email address will not be published.

*