Memories are short. But those in finance are even shorter. Before the credit crunch began in 2007, policymakers in advanced economies were flirting with the idea that we should just accept that financial crises occur every seven years or so and plan accordingly, as seeking to avert or limit them would suffocate the financial system. At the time, greater financialization of the economy, which is when the financial sector accounts for an ever rising portion of gross domestic product, was seen as an unambiguous measure of progress. The most financially liberalized economies—the United States and the United Kingdom—were held up as exemplars for others to follow, and the rallying cry was "set finance free."
The credit crunch ended such talk. It reminded the financial crisis-deniers just how traumatic crises can be and how slow and hesitant the recoveries are. It also reminded us of the Faustian pacts that policymakers are forced to make at the height of a severe financial crisis—for example, trying to revive an economy after a debt-driven bust through more debt, or employing the very same individuals who caused or contributed to the crisis to try to mend things, simply because only they understand the instruments that need to be disentangled.
During the 2008 US Democratic presidential primaries, one of Senator Hillary Clinton's mantras was that she was more able than rival candidates to take that 3:00 a.m. phone call from the generals about some overseas calamity. During a financial crisis, the 3:00 a.m. phone call invariably comes not from a general but from a banker. And invariably he will tell you that unless you bail out his institution, the whole financial system will collapse.
It's all very well to sound brave in the abstract and say the banks should not be bailed out, but when the authorities tried just that in September 2008 and allowed Lehman Brothers to fail, financial meltdown followed. In the wake of that collapse, no major financial institution was able to fund itself without state support. Calling the bankers' bluff is much harder than it seems to the wider public. Where possible, it is best to avoid having to make that call.
In the shadow of the crisis we have returned to a more nuanced consensus, along lines similar to that which existed in a previous age: Financial firms can play an important role in financing growth, but only as part of a financial system that does not accentuate boom and bust. Let us hope that we do not forget this lesson too quickly.
In the 10 years prior to the crisis, financial policy was driven by three main objectives: transparency, standardization (of value and risk measures), and the removal of restraints on financial trade-like transaction taxes or capital requirements for the trading books of banks. Bankers persuaded everyone else that achieving this holy trinity would deliver an effective financial system and therefore greater prosperity for all—though it was also considered foolhardy for regulators to second guess what an effective financial system should look like.
While there is undoubted merit in transparency and standardization, and in the removal of trading restrictions, the manner with which these goals were pursued caused financial systems to become larger, yet more fragile. It also led to financial systems with high degrees of trader liquidity, but which lacked systemic resilience.
Regulators must dare to consider what a resilient financial system would look like. I would venture that it is one where a shock in one part of the system can be absorbed by another part and not spread and amplified across all the others. For this to happen, we need a financial system in which the different parts assess, value, hedge, and trade the same assets or activities differently, not because they have different information, different forecasts of the world economy, or different risk appetites, but because they have different objectives, or, more precisely, different liabilities.
Systemic liquidity does not come from the amount of turnover or size of markets but from the degree of heterogeneity. When a shock of some kind leads to a sudden jump in the precautionary demand for cash, and all banks have to sell assets to raise cash, the financial system will be better able to absorb these stresses if life insurance funds or pension funds valued these same assets on the basis of their ability to meet a future pension or insurance liability.
On this long-term basis, they may decide that the assets are now cheap and should be bought from the bankers. This would not only make the financial system more resilient, it would also do so in the economic interests of the customers of these different institutions and without the requirement for onerous amounts of unproductive capital.
At the heart of this approach is the notion that it is economically sensible for different institutions to value the same assets differently if they have different liabilities. But this notion often runs counter to the practices that are put in place to support transparency and common standards—such as the spread of mark-to-market accounting or the use of third-party credit ratings and bureaus. And it could reduce trader liquidity between crises.
If all financial institutions are required to value the assets in the same way, through their audit rules, their capital adequacy calculations, or their solvency rules, then when one firm sells an asset, it induces other firms to sell more, causing a vicious spiral driven by herd-like panic, aggravating a collapse in market prices. In these circumstances almost no amount of capital would be enough to prevent a bank from suffering a run.
Homogeneous behavior, not risky assets, is the main avenue of systemic risk. Homogeneity, not size, determines whether a financial system is shallow, fragile, and prone to falling over. A commonality of standards and rules can support trading activity in quiet times, giving the illusion of liquidity. But trading liquidity is a false god that vanishes at the first sign of trouble.
We must not confuse modes of operation with goals. The holy trinity of transparency, common standards, and the removal of trading restrictions is a good mode of operation but not the ultimate goal, which is to create a financial system that is resilient and serves the multiple needs of the consumers of finance.
I have earlier proposed a "mark-to-funding" accounting system to address this issue. It does so by allowing institutions to value their assets based on the time they have available in which to sell them. A bank with short-term money-market funding has to value assets based on their price if they were to be sold tomorrow, as with mark-to-market accounting. In fact, we currently allow banks to define many assets as "hold-to-maturity," even if they lack the maturity of funding to keep them that long with any certainty.
Life insurance funds with long-term liabilities rarely need to sell significant assets tomorrow and could therefore be allowed to value them based on long-term valuation, such as a model that discounts to the present value of all the future income the asset produces. Instead, the proposed Solvency II regime for long-term savings emphasizes a mark-to-market approach that will make the financial system more fragile. In some ways, Solvency II is even worse for financial stability than the original Basel II accord on bank supervision, which had to be quickly revised after the crisis erupted.
Segmenting the financial system according to the maturity of liabilities would bring other benefits. Contrary to popular belief, financial crises are not caused by people knowingly taking bad risks. The human desire for retribution would like this to be the case, and there is never any shortage of poor, deceitful behavior prior to crashes. But that is there all the time. We need to root it out and ban many people from financial activities, but doing so will not prevent financial crashes.
Financial crashes always follow financial booms. Financial booms take place because, collectively, people do things that they believe to be virtually risk-free—so safe, indeed, that it makes perfect sense for them to double up. It is the doubling and tripling up by almost everyone, not malfeasance from a few crooks that drives the booms that lead to the busts. In booms, traditional bank lending and leverage surge. Indeed, it is the widespread and unbridled optimism of the times that will make it hard to secure criminal convictions against individuals. Much of the blame is a collective one.
Mark-to-market valuations and price-sensitive risk systems accentuate this process. As asset prices steadily rise, the value of collateral goes up and volatility goes down, suggesting that it is safe to lend or borrow more when, with the wonderful vision of hindsight, we can see that lending and borrowing should be scaled down in the face of increasingly unsafe valuations. The opposite process occurs in a bust. The collapse in valuations makes it safer to lend but banks don't, as the current value of collateral and volatility suggest to the banks' risk systems that there is no room to lend safely. Banks forget that the credit mistakes are made in the boom, not the bust.
We need a financial system that is less circular and less self-referential if we are to wean ourselves off boom-and-bust. Another way of putting this is that we need a method of risk-managing the financial system that is less statistical and more structural: less dependent on us doing what we repeatedly fail to do—to correctly measure fluctuations in risk and value through time.
Structural risk management is about recognizing that there is no one thing called risk and that it is not possible to aggregate all risks into a single number, but rather that there are a handful of different risks—mainly market risk, credit risk, and liquidity risk. These are fundamentally different risks. We know this not because we give them different names, but because they have to be hedged differently.
Credit risk, the risk of a default, is hedged by diversification across a number of different credit risks. Liquidity risk, the risk that you cannot sell an asset immediately, is not hedged by diversification across a number of different but illiquid assets; it is hedged through having time before assets need to be sold. Market risk, the risk of price changes, is hedged through a combination of diversification across assets and across time.
By virtue of having different liabilities—including the need to repay a depositor, or to provide a pension or a life insurance payout—different institutions and people have different capacities for these different risks. A bank with short-term depositors has a capacity for taking credit risk but not liquidity risk. A young pension or life insurance fund has a deep capacity for taking liquidity and market risk, but not credit risk.
Institutions should be required to put up capital against any mismatch between their risk capacity and their risk-taking, thereby encouraging them to stick to taking risks that they have a natural capacity to absorb (making them naturally hedged if their estimates of value and risk prove wrong). This is fundamentally different from the ring-fencing that has been proposed in some quarters. It will encourage risk transfers between the different parts of the financial system—the right kind of risk transfer, not the kind that went on before. Under our proposal, banks and life insurance funds might continue to serve their customers as they do, but banks would later strip out and sell their liquidity risks to pension funds and buy credit risks from them.
We are a long way from this approach, but the attempt to make banks more liquid through the application of short- and long-term liquidity ratios in the revised accord on the supervision of international banks (revised Basel II) is an important start, albeit one that is being strongly resisted by the banks.
Systemic resilience is not about the size of a market or the degree of trading activity and liquidity, but about its heterogeneity. The natural diversity of any economy—small or large, rich or poor—must be nurtured, not artificially snuffed out by sacrificing long-term investment and systemic resilience on the altar of short-term trading liquidity and undue reverence to today's price. Risk capacity, not some statistical risk sensitivity, should be the watchword of regulation.
"Market-to-market" accounting—in which financial institutions value their assets based on the price they could be sold for the next day—encourages herd-like selling in downturns and herd-like buying in upturns, exaggerating financial market manias and panics. It is thoroughly procyclical. It suits banks, given their dependence on short-term money-market funding but is less appropriate for life insurance and pension funds, given the longer-term nature of their liabilities and different risk capacity. Were insurers permitted to value their assets based on the time they have in which to sell them, it would bring much greater equilibrium to financial markets. Under "mark-to-funding" accounting, asset valuations would be based upon an assessment of the cash flows generated over the period for which funding is guaranteed. An approach to accounting and asset valuations that better reflected the real diversity of liabilities and risk capacity in the financial system would break up of the financial herd—reducing the risk and severity of financial crashes.