framing

Felix Salmon writes about negative interest rates:

Here’s an example of how the psychology plays out in practice. If you pay a monthly fee on your checking account, you are probably getting negative rates on your money right now. You’re lending your money to the bank, which ends up giving you back less money than you started with. But because the negative interest rates are presented as a fee, rather than an actual negative rate, the banks are able to get away with it.

This strikes me as essentially backward; insofar as you were getting 3% interest on a "free" checking account when the federal funds rate was at 6.5%, you were paying a 350 bp fee; people hate bank fees, but of course the bank has costs, some of which increase with an increase in the number of depositors, so it hides the fees by calling it a negative interest rate (but in an environment in which, aggregated with its cost of funds, it can report a positive overall interest rate.)  That said, the psychology of "0 percent interest" is similar to the psychology of "free" (or "no-fee"); we're seeing the same psychology in different contexts.

Similarly, payday loans are typically accompanied by fees; reporting often divides these by the length of the loan and reports astronomical interest rates.  While it is not necessarily the case that the distinction between fees and interest as stated by the payday loan shops really corresponds to a distinction between transaction costs (or overhead) and a risk-adjusted cost of funds, it is certainly the case that some of the payment really is "fees" in any reasonable disinterested sense; I might just as reasonably accuse McDonald's of charging a huge rate of interest by dividing the gross profits on my sandwich by the time between my paying for it and receiving it.

While I'm on this, I might as well also push back at another common abuse of language,

Does this mean that Nestlé was being paid to borrow money? No. A company has to pay to borrow money if it ends up paying back more money than it borrowed.

He goes on to add up the franc values of the cash flows associated with the Nestlé bond, and notes that the outflows add up to more than the inflows. Even if it had come out the other way, I object to framing this as "being paid to borrow money".

The problem is perhaps clearer in the case of Apple, which does its accounting in dollars.  If Apple had borrowed money at a negative yield in francs, but the franc had appreciated sufficiently over the course of the loan, then Apple would have paid out more dollars than it received.  In the same situation, if Apple did its accounting in francs, it would record the negative interest charges.  One accounting system uses the fiction that dollars in 2020 are equivalent to dollars in 2015; the other uses the fiction that francs are francs.  Of course, we don't currently know whether the franc will appreciate or depreciate against the dollar; the point is that these are not the same thing.

Now, to a good approximation I can convert a dollar now into a dollar in 2020 by holding a physical piece of currency.  There is some nuisance and security cost to doing so; it may be lost or stolen.  These facts notwithstanding, this ability to sit on physical currency is the reason importance is attached to "zero interest rates" and "negative interest rates"; there is, in the lingo, an apparent arbitrage, in which (for example) Apple borrows Swiss francs, holds them in a vault, pays out some smaller number of Swiss francs, and, if it likes, converts the balance to dollars (or cadmium, etc.) at some point in time along the way.  As long as the surplus from this trade is worth the expenses associated with issuing the bond, securing the currency, and making the payments, then it would be worth doing; as a matter of fact, though, those costs are probably at least 10 or 20 basis points; the costs of securing and transacting in physical currency are probably among the reasons bondholders are willing to buy bonds at negative yields in the first place.

The most egregious consequent error that I see smart people make a lot in public policy debate — even smart economists, but ones who think of interest as "the money paid to borrow other money" — is to suggest that, as long as interest rates are 0 or very close to it, we can borrow and spend with impunity.  The simplest retort is that if we borrow the money, we will have to pay back that money. An argument can be made that it is more useful to spend the money now than it will be later, but that argument should be made, not assumed, and the 0 interest rate is not particularly relevant to it.  The tradeoff in value between spending now and spending five years from now may be that $1.10 in spending now is worth $1 of spending in the future, or that it is $1 and $1.10, respectively; there is no reason in economics that precludes the former possibility, in which case any interest rate that is higher than (using rough figures) -2% is higher than should be incurred.

Accounting rules separate interest payments from principal payments, and taxes and regulations sometimes create economic effects, but in and of themselves they are not economically distinct, and the correct way to think about interest rates is that they measure a rate of exchange between (say) dollars at different points in time.  A balance that grows because of "accruing interest" is typically growing, at least in part, because it is being measured in dollars that are worth less and less.