Tim Harford has a gift for explaining complex economic concepts in plain language. It’s therefore regrettable to see him suggest in his latest column in the Financial Times that regulators encourage the adoption of a new financial instrument that is likely to cause more harm than good.

Harford rightly notes that banks now benefit from subsidies that distort the economy. The reason is that the failure of a large bank is so destructive that governments usually prefer bailouts. That in turn comes from the fact that most of the debt issued by financial firms is used as money. (Think of bank deposits, commercial paper owned by money-market funds and securities lent out for cash in the repurchase markets.) Markets know this. Regulators can try to micromanage the activities of bankers and other financiers in an attempt to offset these effects but their track record is poor.

The better solution is to require banks to fund themselves in ways that make them less vulnerable to sudden failure. I have argued that this would involve separating money-creation from lending. That idea has a distinguished intellectual pedigree -- including Irving Fisher, Milton Friedman, and James Tobin -- although some find it too extreme.

Harford is one of those people. According to him, society will pay a price if lenders are forced to fund their activities by issuing equity rather than debt, although he doesn't explain why. His preferred solution comes from a new paper by Stanford’s Jeremy Bulow and Oxford’s Paul Klemperer. Among other things, they argue that banks could be made immune to sudden collapse if “all (non-deposit) existing unsecured debt” were replaced with something they call an equity recourse note.

These new instruments, which don't currently exist, would pay regular coupons as long as the issuing bank's share price was above a certain level. If the bank's shares fell below this threshold, the coupons would no longer be paid in cash but in new shares. The number of shares issued and the threshold share price would be determined by the share price at the time the ERNs were issued. The main advantage of this system is that a bank cannot literally go bankrupt because it is always capable of using (increasingly worthless) shares to meet its debt obligations.

I see two basic problems with this.

First, these securities are bonds with a complex set of equity options buried inside. There are certainly people who would know how to price and hedge the embedded options, but most of them are found in the very banks that are supposed to be issuing them. That makes ERNs unsuitable for many of the holders of bank debt, such as bond and money-market mutual funds. Those people want to own plain-vanilla bonds, not complex securities that they don't know how to value.

Bulow and Klemperer think that ERNs could be transformed into more familiar instruments by bundling them and then creating a mix of tranches according to risk and maturity. As they note, this would be similar to what was done to create collateralized debt obligations, although more complex. That this level of financial engineering is required to make ERNs palatable to ordinary people doesn’t exactly inspire confidence.

The second major problem with ERNs is that they may not meet their promise of moderating banking's procyclicality. Right now, excessive leverage means that even small losses can drive firms into bankruptcy. Bankers and shareholders therefore have large incentives to take excessive risks in order to maximize returns during the good years. ERNs won't change this.

A quick look at the data shows that bank share prices rose gradually in the years before the crisis only to plunge. In practice, this means that almost all of the ERNs issued during those years would suddenly start “paying” shares instead of cash at the same time. Those shares would lose value as the bank continued to issue more of them, further depressing prices and ensuring that even the ERNs issued when the share price was lower would eventually convert. It would be a quick downward spiral to zero.

Anyone owning these instruments would be wary of this possibility. The smart ones would dump their holdings at the first sign of trouble, or perhaps try to offset their losses by short selling the bank’s stock. The bank wouldn't go bankrupt, since it would always be capable of paying its obligations with new shares. But the resulting dilution might produce losses just as large for shareholders. Rather than encouraging moderate management, the ease with which a small downturn could cascade into near-total annihilation might encourage even riskier banking than we had before the crisis.

Some financial innovations are helpful. Equity recourse notes, however, probably wouldn't be.

(Matthew C. Klein is a writer for Bloomberg View. Follow him on Twitter.)