Supposing we decide a formal monetary constitution is a good idea, what properties should it have? What essential features of money must a monetary constitution safeguard? Again referencing James Buchanan, we know that money has properties that render it similar to weights and measures. Money is a yardstick; if the definition or value of money is constantly in flux, this severely inhibits its ability to serve as a medium of exchange. It seems, then, that predictability, in the form of a stable purchasing power of money, should be the primary goal of a monetary constitution. However, we immediately encounter the following problem: sometimes the purchasing power of money should change! The constellation of relative prices in an economy is supposed to help producers and consumers ascertain real resource scarcities. If there is a negative supply shock—say the price of oil suddenly and unexpectedly rises—then production in general will become more costly. This is exhibited by the prices of particular goods and services rising across the economy, which will reduce a given amount of money’s ability to purchase goods and services. This is what we mean when we say that money’s purchasing power has fallen. The rise in prices following a negative supply shock, while unfortunate in that it is accompanied by increased unemployment, is the least bad outcome for economic efficiency. Since final goods and services have become scarcer relative to money, the ability of money to command goods and services must fall if the price system is to continue to facilitate efficient resource allocation. (Note that the most important aspect of this process is the change in relative prices resulting from the negative supply shock. More money will be required to purchase goods and services in general, but the particular changes in prices will be different for each good and service. The fall in money’s purchasing power is simply the consequence of these many relative price changes.) The above suggests we embrace a broader definition of predictability. We should be able to know, not the particular purchasing power of money at any point in time, but that whatever the purchasing power of money, it will be permitted to adjust to allow the price system to serve best it essential role as a communication device. Mandating a constant purchasing power of money fails to do this, since adherence to the mandate would mean impeding the efficacy of the price mechanism.