If the Yglesias plan were to have negative nominal interest rates on bank deposits always, then perhaps Avent would have a point. However, the nominal interest rate on bank deposits can (and should) be varied according to supply and demand.
If issuing money is very costly, perhaps because making loans is very risky, then charging a competitive fee to people who want to hold money clears the market. It simultaneously coordinates the quantity of money with the demand to hold it and saving with investment. (Saving is that part of income not spent on consumer goods and services and investment is spending on capital goods.)
If, on the other hand, issuing money is very profitable, because the loans and other financial assets that banks can fund with it are lucrative, then banks can and should pay positive interest on money. This will still coordinate the quantity of money with the demand to hold it as well as saving and investment.
What is Avent's alternative? Is it to raise the inflation rate whenever the market clearing interest rate needs to be negative? Should the inflation rate then fall again when this is no longer an issue? Or do we have high inflation at all times so that real interest rates can go as negative as necessary with nominal interest remaining above zero? If there were many financial instruments that require negative real interest rates frequently, or even always, this might make some sense. Negative real interest rates would be practically normal and so maybe centering the entire macroeconomic order around allowing for them would make sense. As for the fluctuating inflation approach, varying inflation in final goods prices and wages might be worth it if just about all interest rates needed to be negative sometimes.
But in reality, it is a few, very short and safe financial assets, that only rarely need to have negative real interest rates to clear markets. Painting a picture where all nominal interest rates are temporarily negative, much less permanently negative, is unrealistic. I am sure that neither Yglesias nor Avent have this in mind. Unfortunately, care must be taken not to create that impression.
Generally, interest rates on short and safe assets are lower than on longer and riskier assets. If some interest rates are negative, including on checking account balances, people who want a saving vehicle for retirement or making a down payment on a house would need to save using something that has a longer term to maturity. Those who want more yield, would have to take more risk. Maybe they should consider stocks.
And, of course, if people chose to save by accumulating gold or silver, then there is no problem. The price of gold or silver simply rises enough to clear those markets. Those who were already holding the gold or silver earn capital gains, which allows them to purchase consumer goods or else buy capital goods. And, of course, those accumulating gold or silver bear risk, just as they would if they purchased stocks, risky bonds, or capital goods.
In my view, occasional fluctuations in inflation rates to create occasional negative real yields is needlessly disruptive. Of course, if all prices were perfectly flexible and everyone had a perfect understanding of what is happening, it would hardly matter. But that isn't the real world. I think that slow, steady growth of spending on output is the least bad macroeconomic environment for microeconomic coordination. Spending should grow with the productive capacity of the economy, leaving final goods prices stable on average. Nominal incomes should generally grow with real incomes.
I don't favor keeping the price level absolutely fixed. Decreases the the supplies of particular goods and even slowdowns in productivity should result in a higher price level, and so transitory inflation. And increases in the supplies of particular goods or rapid growth in productivity should have the opposite effect. But creating persistent inflation or fluctuating inflation so that a few asset markets can sometimes have negative real yields is a mistake.
Why is it that real interest rates sometimes need to be negative? We live in an uncertain world where production takes time. Real investment projects involve risk and take time. If someone wants to save but bear no risk and be able to spend their money at any time, then they are proposing to shift risk to someone else. Sometimes, they may have to pay for someone else to bear that risk. The least bad way to handle the situation is for nominal interest rates to be negative on those particular financial assets that require negative real interest rates for quantity supplied to match quantity demanded.
Does this make hand-to-hand currency impossible? Not at all. The problem isn't the existence of hand-to-hand currency. The problem is basing the entire monetary order on government guaranteed hand-to-hand currency. It is perfectly short and perfectly safe and has a zero nominal yield. No other nominal yield can be any lower than the cost of storing currency.
If hand-to-hand currency were privately issued, and there was no effort to make it especially safe and secure, then it could continue to be used for small, face-to-face transactions. It could continue to serve as a medium of exchange. But it would not be a particularly secure store of wealth, and so would not prevent the nominal yields on T-bills or FDIC insured bank deposits from being negative for just as long as necessary to adjust the quantity demanded to the quantity supplied.
Are there alternatives to privatizing currency other than persistent or fluctuating inflation? The alternative is for the government, as issuer of money, to purchase long and risky assets and bear the risk for those who want to hold short and safe assets. Fortunately, this should be no more common than the need for negative nominal interest rates. Interestingly, a monetary order that is committed to heroic open market operations when necessary, much like one that would allow for negative nominal interest rates when needed, might be less likely to have the event occur.