In trying to avoid financial crises, as in wars, generals tend to prepare for the last war. In Greg Ip's retrospective "A Decade After Bear’s Collapse, the Seeds of Instability Are Germinating Again," Greg amplifies a 2014 warning by Hyun Song Shin to that effect by pointing out the wide variety of ultimately falsified ideas that have driven financial excesses and their ensuing crises: 

Crises surprise because they usually start with an assumption so sensible that everyone acts on it, planting the seeds of its own undoing: in 1982 that countries like Mexico don’t default; in 1997 that Asia’s fixed exchange rates wouldn’t break; in 2007 that housing prices never declined nationwide; and in 2011 that euro members wouldn’t default. James Bianco, who runs his own financial research firm in Chicago, speculates that the equivalent today might be, “We will never see higher inflation or higher growth.” If either in fact occurs, the low interest rates that have raised household stock and property wealth to an all-time high relative to disposable income won’t be sustainable.

Rather than hoping to predict the shock that triggers the next financial crisis, we need to make the system more shock resistant. Bear's collapse provides a clue to how. Here is Greg Ip's brief explanation of that collapse:

Bear ... arranged mortgages that financed the housing bubble while borrowing heavily with short-term IOUs. When those mortgages went bad, Bear’s creditors yanked their funds—a de facto run on the bank.

In short, any financial institution is rickety if the people who gave you the money can pull their money out on short notice. If most of the funds behind a financial institution were given in return for long-term bonds—or even better, in return for stock—that financial institution is much more resistant to failure.

In the case of ordinary mainstreet consumer banks, we solved the danger from people being able to pull out their funds on a moment's notice with a government guarantee of deposits. We could have a system with explicit government guarantees of private bets in many places in the financial system, or a continued commitment to bailouts in the future that come as a surprise to the electorate. But if we don't want explicit government guarantees or an implicit government bailout policy, what we need to do is to put a cap on the amount of borrowing financial institutions can do—and especially the amount of short-term borrowing. This is another way of saying we need an upper limit on the amount of leverage that financial institutions can have and a lower limit on the amount of capital—equity—that a bank is financed by.

Though they wouldn't put it in these terms, banks and other financial companies don't like leverage limits because leverage limits limit their ability to make big profits by relying on an implicit too-big-to-fail or a too-many-to-fail expected bailout subsidy from the government. On this, see "Martin Wolf: Why Bankers are Intellectually Naked." 

In Larry Summers's keynote speech at the October 22-23, 2015 University of Michigan Financial Stability Conference, he added two important angles on effectively regulating the financial system. At the 22:25 minute mark in this video of his speech, he points out a big problem financial regulation shares with many other forms of regulation: how would someone come to know enough about finance to be an effective regulator unless they either believed finance was a force for good in the world or mainly cared about earning a lot of money? In either case, they are likely to be "cognitively coopted" by the financial industry. It is possible for someone to believe that finance is a force for good in the world but that particular financial practices are very dangerous to the economy (indeed I put myself in that category), but an enthusiasm for finance often veers over into thinking that existing practices are good.

Larry Summers offers as a partial solution locating important chunks of financial regulatory power in the Federal Reserve. The Fed is a good choice for financial regulator because it has an outsized share of the few people who understand finance without being cognitively coopted by the financial industry. Larry doesn't say that explicitly; I do. Explicitly, Larry links his recommendation of the Fed as financial regulator to the idea that the Fed is long-lasting, prestigious institution—which makes it easier for the Fed to stand up to the financial industry. Of course, the Federal Reserve already has substantial financial regulatory powers. Maybe it should have more. 

At the 27 minute mark, Larry points out that capital equity requirements and leverage limits are often stated in terms of book values on the balance sheet. The trouble with book values is that they our far out of date by the time a crisis occurs. The market values of a firm's stocks and bonds provide a much better early-warning system than those book values! Firms need to be required to have enough stockholder equity in the good times so that even if their stock prices tank, they still have enough of a cushion of stockholder equity that they won't go bankrupt. At a minimum, when a firm's stock price goes down enough that it doesn't have enough of a capital cushion by market prices, it should be forbidden to dissipate its capital cushion by paying dividends or buying back its own stock. 

We have seen enough financial crises in recent years that those in charge can no longer plead ignorance that anything could go wrong. A decision to have low effective capital requirements, or to base those capital requirements only on book value, is a decision to play bailout roulette. Those who act to lower capital requirements or who resist raising them should be held to account now and in the future. The next big financial crisis and any associated bailout will be their fault. On the other side, if there is no "next big financial crisis" it will be either because capital requirements have been dramatically raised or because of an extraordinary streak of luck. 

Let me note in closing that the depth and length of the Great Recession and its aftermath required two big mistakes. First, the mistake of having capital requirement too low. This led to the initial fall of the dominoes. Second, an inadequate monetary policy response, as I discuss in "America's Big Monetary Policy Mistake: How Negative Interest Rates Could Have Stopped the Great Recession in Its Tracks." I lay out the theory behind this view in "On the Great Recession" and have been working hard on the details of how to do effective negative interest rate policy, as you can see in "How and Why to Eliminate the Zero Lower Bound: A Reader’s Guide."