David Andolfatto has an interesting post suggesting that a Taylor rule would could generate a "bad" low inflation equilibrium where the nominal interest rate is zero and real output remains below potential. (HT to Mark Thoma.) Targeting the quantity of money avoids this problem. (Or my understanding is that it is only when the quantity of money is allowed to adjust with the demand to hold it, and only interest rates are targeted, that this bad equilibrium was possible.)
Of course, Andolfatto's discussion is in terms of a pretty arcane model, where everyone saves all their income when young and the only choice is whether to buy government bonds or capital goods to fund consumption when old. His definition of potential output is based upon some efficient capital stock and the natural interest rate is the marginal product of that capital. Output is below potential when people buy too many bonds and not enough capital and so less output is produced. To me, output is always equal to potential in this model and potential output is just lower when the capital stock is lower. Further, I don't think the natural interest rate is the marginal physical product of some optimal amount of capital.
Suppose that people can invest in gold or else capital goods. If gold is expected to appreciate more quickly, then people shift to gold and out of capital. The capital stock is lower and presumably the marginal product of "capital" is higher. The least productive uses of capital goods are sacrificed, and those that are remaining are somewhat more productive. Also, because there is a smaller capital stock, productive capacity is a bit lower. What happens to the resources that are not used to produce the capital goods? They are used to produce consumer goods enjoyed by those who earned capital gains on their existing gold holdings.
Now, suppose that there is a gold standard. A lower inflation rate is gold depreciating less rapidly, and when it shifts over to deflation, it is gold appreciating more rapidly. If people choose to hold more gold and less capital goods because there is a higher deflation rate, then potential output is a bit lower than otherwise and the marginal product of capital a bit higher.
I think this is the effect that Andolfatto is capturing in his model. If this is an issue, one answer is an inside money where real money balances are invested in capital goods. Of course, even in that scenario, people can still save by accumulating stocks of commodities. In my view, trying to ban that to increase investment in production processes is a bit draconian.
Andolfatto's model is in terms of bonds, though he does mention that bonds can be interpreted as interest bearing money. (Good, keep up with that. Most money does bear interest. Don't fall into the trap of identifying money with noninterest bearing hand-to-hand currency.) Can government debt play the role of gold?
Everyone wants to hold government bonds. The nominal interest rate is driven to zero. Inflation slows and perhaps turns to deflation. This makes holding government bonds even more attractive. More wealth is held in the form of government bonds, and less in capital goods. Potential income falls and the natural interest rate rises.
One solution to this problem would be for the government to sell additional bonds to fund the purchase of capital goods. If there are some capital goods (road and bridge repair) that it is sensible for government to purchase, that would seem desirable. Otherwise, it appears that lower real interest rates on government bonds would result in more investment, capital, and potential output. This could be generated by a higher inflation rate or else a negative nominal interest rate on government bonds. With government issuing zero-interest rate currency, a negative nominal interest rate on government bonds is impossible.
In my view, an important part of our current problem is a failure of markets for short and safe (mostly government guaranteed) debt to clear properly. Negative nominal yields are necessary on some financial assets. In my view, the first best answer isn't a higher inflation rate, but rather getting government out of the business of issuing of hand-to-hand currency. If the nominal and real yields on "government bonds" can always adjust enough to clear, then households and firms will switch to funding capital goods.
Still, for the U.S. economy, it seems to me that these problems will show up as an increase in current consumption and a decrease in investment. Those already holding government bonds would earn real capital gains until they felt wealthy enough to expand their real consumption. However, investment and capital accumulation would be less, so productive capacity would shift to a lower growth path.
I don't think that the U.S. economy is in this situation. Real consumption has not recovered to the growth path of the Great Moderation. Still, I am pleased that Andolfatto is seeing some problems with this fixation on the "Friedman rule" of deflation equal to the negative of the real interest rate. Sure, it is great for those holding money to have a risk-free, perfectly liquid asset that pays the real equilibrium interest rate. But does it make sense to pay them any interest if they are not providing resources to fund real investment and expand the production of consumer goods in the future? Pay positive nominal interest rates on money, reflecting the actual risks (both in terms of failure of investment products and duration) for actual funding investment in capital goods.
Currency? Currency! Quit making the tail wag the dog. Yes, it is useful for a limited set of transactions. But don't make it the center of the economic universe and require everything else to accomodate it.