By Simon Johnson

There are three straightforward and rather obvious lessons from the collapse of Lehman Brothers Holdings Inc. five years ago: The financial crisis was devastating; we haven't ended the problems associated with "too big to fail" financial institutions; and we must force the biggest banks to break up — or risk dire consequences.

The crisis had huge costs, in both economic and human terms. Almost every official and politician understates these costs, though fortunately Tyler Atkinson, David Luttrell and Harvey Rosenblum of the Federal Reserve Bank of Dallas have produced a definitive accounting.

I would begin with their measures of lost output. Their crucial point is that this recession recovery wasn't V-shaped, but rather something much more protracted. Lost output — goods and services that we shall never see — was at least 40 percent of 2007 gross domestic product and probably considerably more. Add other costs, including the human damage caused by long-term unemployment, and the shortfall easily reaches at least one year's output — say, $15 trillion in round numbers.

Any claims that the crisis wasn't so bad, or that the government ended up making money on bailouts, or that the banks only received short-term liquidity loans, completely fail to understand the way financial crises wreak havoc. The crisis had a huge cost, even though large financial institutions (after Lehman) were prevented from failing through generous government support.

Hard Questions

The hard question isn't what we would do the next time a Lehman fails or is on the brink of failure. We would still want to save the world from financial collapse, and there are many creative ways for a government or a central bank to provide support to troubled institutions.

We don't want to reach any such point again, so the much tougher question is: How do we prevent any financial institution from becoming so big and so complex that its distress — even without outright failure and losses for creditors — could contribute to a panic that depresses real output and threatens to bring the global financial system to its knees?

The Dodd-Frank Act didn't end "too big to fail." The legislation was intended to reduce — and hopefully eliminate — the implicit and explicit government support that, post-Lehman, could reasonably be presumed to buttress large financial institutions.

Too big to fail will end only when regulators properly implement the financial-reform measure. Title I of the law authorizes the Fed to force financial companies to change their activities — including their scale and scope — when it is determined that the institutions couldn't be resolved in an orderly fashion through the standard bankruptcy process.

Looking at the costs, both direct and indirect, of the Lehman collapse, it's clear that running any of our big leveraged financial companies through bankruptcy would cause major economic turmoil.

Unfortunately, the Fed's Board of Governors has been unwilling to apply the law in a full and fair manner. Banks that submitted inadequate living wills were simply asked to provide new ones, without any discernible consequences. This game probably will go on until the next major crisis.

Title II of Dodd-Frank creates a resolution mechanism known as orderly liquidation that would be administered by the Federal Deposit Insurance Corp. But this takes effect only if determinations under Title I turn out to be inaccurate — that is, if the Fed allows a company to operate and it is later discovered that its distress cannot be handled through bankruptcy, at least not without catastrophic consequences.

Calling Bluffs

I support the efforts of the FDIC to design an approach to resolution that reduces taxpayer risks and limits the disruption to credit markets when a big financial institution gets into trouble. (I'm on the FDIC's Systemic Resolution Advisory Committee, but I wasn't involved in the design of this mechanism.)

The FDIC procedure is designed to impose losses on creditors of bank holding companies while protecting creditors to operating subsidiaries. As a consequence, private lending to a holding company should carry a premium interest rate relative to loans to an operating subsidiary, which involves less risk. This approach will work only if the Fed — and, again, the Board of Governors in particular — requires banks to have enough equity and long-term debt issued by the holding company to absorb potential losses. So far, the signs from the Fed about their thinking on this issue haven't been encouraging.

The upshot is the simplest lesson of all: We must do more. The largest U.S. financial institutions need to become small enough and simple enough to fail without bringing down the global system.

This is the only credible way to threaten the banks — and their executives — with potential failure if they take on and mismanage excessive risk. Everything else is ultimately a bluff that will be called.

Simon Johnson, a professor at the MIT Sloan School of Management as well as a senior fellow at the Peterson Institute for International Economics, is co-author of "White House Burning: The Founding Fathers, Our National Debt, and Why It Matters to You."