Regulators have yet to solve too-big-to-fail. Eight domestic systemically important financial institutions — Bank of America, Citigroup, JPMorgan Chase, Wells Fargo, Goldman Sachs, Morgan Stanley, Bank of New York Mellon and State Street — are, as a group, larger than they were before the financial crisis. They are just as dangerous to the financial system and more likely to be saved with taxpayer dollars if they falter. The problem is that regulators have attempted to change bankers' behavior while ignoring the way that bankers' risk appetites impair the effectiveness of new rules.

Bankers at the big eight institutions have an incentive to maximize the value of their implicit government guarantee in order to achieve their return-on-equity targets and bonuses. Hence, they have high risk appetites. When regulators pass new rules intended to increase the stability of the financial system, bankers find ways to offset the impact of those rules so as to return to their desired risk appetite. It is as if there is a Newton's third law of banking: for every regulation, there is an equal and opposite reaction by bankers.

Capital regulations illustrate the dilemma. Basel I placed a flat capital requirement independent of a bank's portfolio risk. Banks responded by loading up on higher-risk assets to maintain their return-on-equity. Basel II sought to correct the issue by weighting assets according to their risk. Banks with higher-risk assets were required to hold more capital than those with lower-risk assets. But banks responded by gaming the weights to render them useless. Basel III now reinstates a modest leverage ratio backstop to control the gaming.

Bankers at the big eight have also compensated for the reduced financial risk under higher capital requirements by taking on more credit risk. The objective is to take risk in a form unrecognized by regulators, typically in one of two ways. The first is by investing in instruments with embedded leverage like derivatives, real estate and leveraged transactions. The second is to take on risk with assets like collateralized debt obligations, which have frequent small gains in normal economic times but suffer from rare, steep losses during market downturns.

Regulators have played catch-up as bankers find ways to work around new rules, becoming more aware of the dangers posed by structured products and leveraged transactions. They try to plug the leaking dike as new holes emerge. But eventually regulators run out of fingers and a new form of risk-taking slips through. That is why the next crisis is unlikely to resemble the last.

Permanent structural changes, such as breaking up the big eight banks into entities small enough to fail or raising their capital levels to a level so high they cannot fail, are unlikely in the current political climate — at least until the next crisis. We can, however, address the root problem of bankers' risk appetites. It is difficult to compel bankers to take less risk than they want to take. Therefore, the key is to increase bankers' vested interests in the risks they create. They need more of their skin in the game.

Bankers at the big eight currently benefit from a "heads I win, tails the taxpayers lose" arrangement. Consequently, their incentive is literally to bet the bank. Three changes could persuade them to act with more caution.

First, board directors must hold bank management accountable for problems in the same way non-financial firms do. Just look the difference between Target Chief Executive Gregg Steinhafel, who stepped down in the wake of the retailer's massive data breach, and bankers who suffer only bonus reductions after losing billions. Bankers who know their jobs are on the line are likely to be more careful.

Second, a significant portion of bankers' total net worth must be at risk. Executives should not be permitted to walk away from bailouts or bank failures with hundreds of millions of dollars in hand. Clawback provisions — in which employees later discovered to have engaged in inappropriate activities are required to pay back the compensation they received — are a meek attempt to reinstate a limited form of personal liability. They are not enough. Policies similar to the old partnership model of investment banks, in which senior employees were responsible for all liabilities as well as profits, should be considered.

Lastly, criminal liability for the banker — not just the bank — must be expanded. The Romans used to require engineers to sleep under the bridges they designed. Some of those bridges still stand today. Perhaps this kind of accountability with senior management at the big eight could achieve similar results.

The regulatory focus must shift from prohibiting undesirable risk behavior to managing the incentives that cause that behavior. Otherwise, risk-taking simply morphs to fit bankers' risk appetite, leaving the overall threat unchanged.

J.V. Rizzi is a banking industry consultant and investor. He is also an instructor at DePaul University Chicago.