WITH the words ''financial collapse'' among the most common phrases in newspaper headlines lately -- whether with reference to Barings Bank, to Orange County, or to the peso -- it's hard to see that a rush to repeal the Glass-Steagall Act is in the public interest.
Glass-Steagall is the Depression-era law that keeps commercial banks out of the insurance and securities businesses. The moves to abolish the statute are not directly connected to the disasters mentioned above. But what they all have in common is that they are part of a global financial world going through immense changes, changes under way for some time but not yet complete or fully absorbed by the people who make up the marketplace: new kinds of partnerships and networks, new kinds of financial instruments. In this digital age, the effects of good or bad decisions can be felt at the speed of light.
And with so much legislation, especially deregulatory legislation, being force-marched through Congress, it's hard to feel confident that the various repeal proposals -- one from the Clinton administration and two from Congress -- are going to get the attention they deserve. How confident can the public be, for instance, of the ''firewall'' that would protect insured deposits from a bank's other funds?
Indeed, the very fact that so many different parties are pushing to repeal Glass-Steagall should give us pause. The decision that everyone agrees to right away, the open-and-shut case, can often be the one that causes problems later.
We recall that there was an earlier surge of enthusiasm for the repeal of Glass-Steagall a few years ago, until the savings-and-loan debacle got people thinking twice.
Glass-Steagall was passed out of concern that stock-market speculation had been responsible for bank failures during the Great Depression. Recent research has shown that, on the contrary, banks failed because of inadequate reserves and defaults on loans, and that diversification would have helped banks.
It may well be time to punch through the Glass-Steagall ceiling. But it needn't be done in haste.