Critics of Keynesian economics often use the so-called broken window fallacy, advanced in the 19th century by the French economist Frederick Bastiat, to reject the role of government spending in stabilizing the economy. According to this fallacy, if a hooligan breaks the window of a bakery, the subsequent repair expenditures by the baker will have no net benefits for the economy. This is supposedly because if the baker had not spent his money to fix his broken window, he would have spent it on something else, say, buying a new jacket. Thus, while it would be true that the fixing of the window would have a positive multiplier effect on the economy, this positive effect would be exactly offset by the negative multiplier effect of foregoing the purchase of the jacket.
The above argument is often used by opponents of Keynesian economics to reject any role for government fiscal policy in macroeconomic stabilization. Specifically, and just like the above bakery example, critics assert that any money raised by government through taxation to finance its spending would have no net effect on the economy, because such taxation leaves less money for the private sector to spend. In other words, any government spending would simply “crowd out” an equivalent amount of private spending, leaving total aggregate spending in the economy unchanged. If correct, the above fallacy could spell the end of Keynesian economics as we know it. Indeed, such an argument underlies much of the anti-Keynesian attacks on government stimulation efforts during the recent economic crisis. The fallacy, however, is not correct.
To see why, we need to note that most Keynesian critics fail to recognize an important assumption underlying the window fallacy, namely, that the economy is fully employed. Under full employment, aggregate spending in the economy is constant and is equal to the full employment level of aggregate output. Furthermore, for full employment to hold, all earnings must be duly spent. Otherwise, there will be a recession. Thus, going back to the baker in our example, if he had not used his money to fix his window, he MUST have used it to buy a new jacket. As we all know, under recessionary conditions, this is not always the case. If the baker had expected harder economic times ahead, he would have most likely hang on to his money, instead of spending it. Under these conditions, it would not be true that if the window had not been fixed, the baker would have instead used his money to buy a jacket. This means that breaking and fixing the broken window would have actually helped the economy by increasing the aggregate level of spending.
To sum up, by assuming full employment, the broken window fallacy simply assumes away the problem we need to solve, namely, pulling the economy out of recession. That is why Keynesian economics asserts that during the periods of high unemployment, it would be better to pay workers to dig out holes and fill them up again, or even to break and fix windows, than doing nothing at all. Of course, it will be much more desirable to pay workers for socially desirable projects, but you get the drift of the argument.
Hassan Shirvani, Ph.D.
Professor Cullen Foundation Chair in Economics