By his own admission, Alan Greenspan was wrong for six and a third years of his 21-year tenure as chairman of the Federal Reserve. Greenspan is just not saying which of those 76 months were erroneous.

The acknowledgment of the formerly great man's part-time fallibility came during Greenspan's testimony earlier this month before the Financial Crisis Inquiry Commission. As the Associated Press reported: "Regarding his own missteps over his two decades as Fed chair, Greenspan said, 'I was wrong 30 percent of the time, and there were an awful lot of mistakes in 21 years.' "

Now he tells us.

The thing is, if we really want to get to the root of the financial meltdown, we're going to need more specificity. For example, was Greenspan only wrong during certain seasons -- say autumn months? Or months that end in the letter "y"? Or was he incorrect at random times throughout the year? June 1988. February 1991. Then maybe three consecutive Aprils?

Or, and this is where it really gets tricky, was he mistaken only on certain days? Were Wednesdays particularly fraught with blunders? Clearly, Greenspan's miscalculations mainly occurred during the week. After all, if his stumbles had been confined to weekends, the worst results of his tenure as chief monetary policymaker might have been an overabundance of burnt toast and spilled milk over at Chez Greenspan. Instead, we got an economic calamity of galactic proportions.

In fact, the severity of the crisis suggests that we witnessed the worst-case scenario: that Greenspan was wrong for 30% of his daily office hours. If so, he might have committed two or three bloopers before lunch and one howler just before clocking out ... every blessed day for 21 years.

Is it any wonder the world's financial system came crashing down?

In the absence of more detail about Greenspan's periodic ineptitude, we are left to ponder two related questions. First, is it fair to insist that monetary officials spend less than 30% of their time screwing up? And second, if a 30% blundering rate is too high for a central banker, just how much inexactitude will the system tolerate?

On the first point, it should be clear by now that public officials charged with making critical decisions and setting policy for the world's largest economy should be held to a nearly error-free standard. Granted, no one is perfect. But if you can cause trillions of dollars to disappear overnight, you can't be wrong 30% of the time. Unless your mistakes are limited to the repeated typing of "Untied States of America" on the cover sheet of your Humphrey-Hawkins reports.

So how much wrong-headedness is permissible before it's time to make a change at the top? It depends. If a Fed chairman's miscues are no worse than typos or misspellings, then 30% is probably harmless. But if the misjudgments occur in the context of the Fed's response to dangerously bubbly economic events, then 2% is the upper limit.

When you're trying to prevent the collapse of a critical segment of the economy, even dialing a wrong number can be catastrophic: "... And so that's why we need a concerted international effort to raise rates immediately. If we don't act now, New York, London and Frankfurt will be reduced to rubble by nightfall. So do we have a deal? Wait, what? Isn't this Mervyn at the Bank of England? 7601 4444? No? 4445. Damn. Third time that's happened this week."

At that point, we have reached the threshold and it is time to move on. Assuming there is something to move on to.

Once we've set the boundaries for misapprehension, monitoring becomes the issue. How will Congress and the White House know when a Fed chairman has reached his or her quota of cluelessness? An outside auditor is probably too intrusive -- and unnecessary.

Apparently, Alan Greenspan knew he was stumbling along 30% of the time. And the Fed chairman is already required to appear regularly before Congress. Therefore, a self-reported account of brain hiccups should be made a mandatory section of those congressional reports. Furthermore, given what we now know about the attitude toward accuracy over at the Fed, the law should require the central banker to run the reports through a spell checker before submitting them.

See the complete archives of Deal Sense