In the March 1, 2013 Wall Street Journal, John Cochrane had an eye=-opening review of The Banker’s New Clothes, summarizing the argument of authors Anat Admati and Martin Hellwig:

Running on Empty: Banks should raise more capital, carry less debt–and never need a bailout again.

It motivated me to buy the book and download it to my Kindle.

In Admati and Martin’s argument, as distilled by John, the first thing to understand is that capital requirements do not by themselves reduce the amount of funds available for lending; they are on the liability side, not the asset side: 

“Capital” is not “reserves,” and requiring more capital does not reduce funds available for lending. Capital is a source of money, not a use of money.

Moreover,

Capital is not an inherently more expensive source of funds than debt. Banks have to promise stockholders high returns only because bank stock is risky. If banks issued much more stock, the authors patiently explain, banks’ stock would be much less risky and their cost of capital lower. “Stocks” with bond-like risk need pay only bond-like returns. Investors who desire higher risk and returns can do their own leveraging—without government guarantees, thank you very much—to buy such stocks.

And yet, banks choose very high debt/equity ratios than other firms. What is going on? 

Nothing inherent in banking requires banks to borrow money rather than issue equity. Banks could also raise capital by retaining earnings and forgoing dividends,…

Why do banks and protective regulators howl so loudly at these simple suggestions? As Ms. Admati and Mr. Hellwig detail in their chapter “Sweet Subsidies,” it’s because bank debt is highly subsidized, and leverage increases the value of the subsidies to management and shareholders.

Equity is expensive to banks only because it dilutes the subsidies they get from the government. That’s exactly why increasing bank equity would be cheap for taxpayers and the economy, to say nothing of removing the costs of occasional crises.

Would high capital requirements inevitably gum up the works of banking? No:  

… it was not always thus. In the 19th century, banks funded themselves with 40% to 50% capital. 

How high should capital requirements be? Here is John Cochrane’s answer, which I second: 

How much capital should banks issue? Enough so that it doesn’t matter! Enough so that we never, ever hear again the cry that “banks need to be recapitalized” (at taxpayer expense)!

To be specific, I would say 50%. After all, equal amounts of equity and debt would not be an unusual debt/equity ratio for a non-banking firm that didn’t face a massive implicit subsidy from the likelihood of a bailout. Indeed, even the debt/equity ratios seen for non-banking firms are likely to tilt toward a higher debt/equity ratio than would be socially optimal as a result of the favorable tax treatment of debt relative to equity. (See Simon Johnson’s Congressional testimony on that point here.) And equal amounts of equity and debt did not constitute an unusual debt/equity ratio for banks in the era when the implicit bailout subsidy was not yet in place.