CHAPTER 1

If Stable and Efficient Banks Are Such a Good Idea,Why Are They So Rare?

The majority of economists ... tend to assume that financial institutionswill grow more or less spontaneously as the need for their servicesarises—a case of demand creating its own supply.... Such an attitudedisposes of a complex matter far too summarily.

Rondo Cameron and Hugh Patrick,Banking in the Early Stages of Industrialization (1967)

Everyone knows that life isn't fair, that "politics matters." We say itwhen our favorite movie loses out at the Academy Awards. We sayit when the dolt in the cubicle down the hall, who plays golf with theboss, gets the promotion we deserved. We say it when bridges to nowhereare built because a powerful senator brings federal infrastructure dollarsto his home state. And we say it when well-connected entrepreneurs obtainbillions in government subsidies to build factories that never stand achance of becoming competitive enterprises.

We recognize that politics is everywhere, but somehow we believe thatbanking crises are apolitical, the result of unforeseen and extraordinarycircumstances, like earthquakes and hailstorms. We believe this because itis the version of events told time and again by central bankers and treasuryofficials, which is then repeated by business journalists and television talkingheads. In that story, well-intentioned and highly skilled people do thebest they can to create effective financial institutions, allocate credit efficiently,and manage problems as they arise—but they are not omnipotent.Unable to foresee every possible contingency, they are sometimes subjectedto strings of bad luck. "Economic shocks," which presumably could notpossibly have been anticipated, destabilize an otherwise smoothly runningsystem. Banking crises, according to this version of events, are much likeTolstoy's unhappy families: they are all unhappy in their own ways.

This book takes exception with that view and suggests instead that thepolitics that we see operating everywhere else around us also determineswhether societies suffer repeated banking crises (as in Argentina and theUnited States), or never suffer banking crises (as in Canada). By politicswe do not mean temporary, idiosyncratic alliances among individuals ofthe type that get the dumbest guy in the company promoted to vice presidentfor corporate strategy. We mean, instead, the way that the fundamentalpolitical institutions of a society structure the incentives of politicians,bankers, bank shareholders, depositors, debtors, and taxpayers to formcoalitions in order to shape laws, policies, and regulations in their favor—oftenat the expense of everyone else. In this view, a country does not"choose" its banking system: rather it gets a banking system that is consistentwith the institutions that govern its distribution of political power.

The Nonrandom Distribution of Banking Crises

Systemic bank insolvency crises like the U.S. subprime debacle of 2007–09—a series of bank failures so catastrophic that the continued existence ofthe banking system itself is in doubt—do not happen without warning,like earthquakes or mountain lion attacks. Rather, they occur when bankingsystems are made vulnerable by construction, as the result of politicalchoices. Banking systems are susceptible to collapse only when banksboth expose themselves to high risk in making loans and other investmentsand have inadequate capital on their balance sheets to absorb thelosses associated with those risky loans and investments. If a bank makesonly solid loans to solid borrowers, there is little chance that its loan portfoliowill suddenly become nonperforming. If a bank makes riskier loansto less solid borrowers but sets aside capital to cover the possibility thatthose loans will not be repaid, its shareholders will suffer a loss, but it willnot become insolvent. These basic facts about banking crises are knownto bankers or government regulators; they are as old as black thread.

By contrast, consider what occurs when bank capital is insufficient relativeto bank risk. Bank losses can become so large that the negative networth of banks totals a significant fraction of a country's gross domesticproduct (GDP). In this scenario, credit contracts, GDP falls, and the countrysustains a recession driven by a banking crisis. Governments can preventthis outcome by propping up the banking system. They can makeloans to the banks, purchase their nonperforming assets, buy their sharesin order to provide them with adequate capital, or take them over entirely.

If such catastrophes were random events, all countries would sufferthem with equal frequency. The fact is, however, that some countries havehad many, whereas others have few or none. The United States, for example,is highly crisis prone. It had major banking crises in 1837, 1839,1857, 1861, 1873, 1884, 1890, 1893, 1896, 1907, the 1920s, 1930–33,the 1980s, and 2007–09. That is to say, the United States has had 14banking crises over the past 180 years! Canada, which shares not only a2,000-mile border with the United States but also a common culture andlanguage, had only two brief and mild bank illiquidity crises during thesame period, in 1837 and 1839, neither of which involved significantbank failures. Since that time, some Canadian banks have failed, but thecountry has experienced no systemic banking crises. The Canadian bankingsystem has been extraordinarily stable—so stable, in fact, that therehas been little need for government intervention in support of the bankssince Canada became an independent country in 1867.

The nonrandom pattern of banking crises is also apparent in their distributionaround the world since 1970. Some countries appear immune tothe disease, while others are unusually susceptible. Consider the patternthat emerges when we look at data on the frequency of banking crises inthe 117 nations of the world that have populations in excess of 250,000,are not current or former communist countries, and have banking systemslarge enough to report data on private credit from commercial banksfor at least 14 years between 1990 and 2010 in the World Bank's FinancialStructure Database. Only 34 of those 117 countries (29 percent)were crisis free from 1970 to 2010. Sixty-two countries had one crisis.Nineteen countries experienced two crises. One country underwent threecrises, and another weathered no less than four. That is to say, countriesthat underwent banking crises outnumbered countries with stable bankingsystems by more than two to one, and 18 percent of the countries inthe world appear to have been preternaturally crisis prone.

The country that experienced the most crises was Argentina, a nationso badly governed for so long that its political history is practically a synonymfor mismanagement. The close runner-up (with three crises since1970) was the Democratic Republic of the Congo, the nation whosebrutal colonial experience served as the inspiration for Joseph Conrad'sHeart of Darkness, which was governed after independence by one of thethird world's longest-lived and most avaricious despots (Mobutu SeseSeko, who ruled from 1965 to 1997), and whose subsequent history is atemplate for tragedy.

The 19 countries that had two banking crises are also far from a randomdraw. The list includes Chad, the Central African Republic, Cameroon,Kenya, Nigeria, the Philippines, Thailand, Turkey, Bolivia, Ecuador,Brazil, Mexico, Colombia, Costa Rica, Chile, Uruguay, Spain, Sweden,and ... the United States. One of the striking features of this list is thepaucity of high-income, well-governed countries on it. Of the 117 countriesin our data set, roughly one-third are categorized by the World Bankas high-income nations. But only three of the 21 crisis-prone countries,14 percent, are in this group. This suggests that, for the most part, beingcrisis prone is connected to other undesirable traits and outcomes. But thatraises another troubling question. Why is the United States on this list?

The Nonrandom Distribution of Under-Banked Economies

There is, of course, more to having a good banking system than simplyavoiding crises. Equally problematic are banking systems that provide toolittle credit relative to the size of the economy—a phenomenon known asunder-banking. This outcome, too, appears not to be randomly distributed.Consider the striking contrast between Canada and Mexico, theUnited States' partners in the North American Free Trade Agreement(NAFTA). From 1990 to 2010, private bank lending to firms and householdsaveraged 95 percent of GDP in Canada, but in Mexico the ratiowas only 19 percent. The dramatic difference in those ratios means thatMexican families have a much more difficult time financing the purchaseof homes, automobiles, and consumer goods, and Mexican business enterpriseshave much more difficulty in obtaining working capital, thantheir Canadian counterparts. The result is slower economic growth. Littlewonder, then, that over 500,000 Mexicans—roughly half of all newentrants to the Mexican labor market—illegally cross the border to theUnited States each year.

As figure 1.1 shows, the stark difference between Canada and Mexicois part of a recurring pattern. In the world's poorest countries (those onthe far left-hand side of the figure), including for example, the DemocraticRepublic of the Congo, the ratio of bank credit to GDP averagesonly 11 percent. In the richest countries (shown on the far right-hand sideof the figure), the ratio of bank credit to GDP averages 87 percent.

Crucially, there is also substantial variance across countries within eachof the four income groups, which suggests that the amount of credit extendedwithin countries is not solely a function of demand for credit butalso reflects constraints on the supply of credit. In other words, the factthat some countries in each income group extend much more credit thanothers in the same income group (or even the next-highest income group)suggests that many countries in all categories are under-banked. For example,Mexico appears to be under-banked relative to other countries inthe same income group, and even has a lower ratio of credit to GDP thanmany countries in the next-lowest income group (e.g., the Philippines).

Being under-banked has huge social costs. A large and growing academicliterature has shown that under-banked countries suffer lower economicgrowth than other countries. Economic historians have shown thatHolland, Great Britain, and the United States experienced revolutions infinancial intermediation and financial institutions before their rise to globaleconomic hegemony in the eighteenth, nineteenth, and twentieth centuries,respectively. They also found that Russia, Germany, and Japan underwentsimilar financial revolutions before they narrowed the gap with theworld's economic leaders in the late nineteenth and early twentieth centuries.Financial economists using statistical methods to analyze the growthof contemporary economies have reached similar conclusions. Whetherthey look at variance in outcomes across countries, across regions withincountries, within countries over time, or across industries, their studies alldemonstrate that higher levels of financial development produce fasterrates of physical capital accumulation, faster economic growth, morerapid technological progress, faster job creation, and increased opportunitiesfor social mobility. Given the relationship between economic growthand the ability to project power internationally, under-banked countriesare also at a disadvantage in defending their sovereignty and influencingevents abroad. In short, being under-banked is a far more serious state ofaffairs than lacking capacity in the real sectors of the economy, such astextiles, sugar refining, or automobile manufacturing: finance facilitatesthe efficient operation of all other economic activities, including industrialsectors crucial to the defense of the state.

How Many Efficient and Stable Banking Systems Are There?

If very few countries have been free of banking crises since the 1970s, andif much of the world is under-banked, in how many countries is bankcredit plentiful and the banking system stable? Answering this questionrequires us to draw a line between those economies where bank credit isabundant and those economies that are under-banked. If we define a countrywith abundant credit as one that has an average ratio of bank credit toGDP one standard deviation above the mean for the 117 countries in ourdata set (83 percent), which corresponds to the ratio in Australia, anddefine a stable banking system as one that has been free of systemic crisessince 1970, we arrive at a shocking answer: only six out of 117 countries—5percent—meet those criteria.

The Puzzling Pervasiveness of Dysfunctional Banking

The puzzle of why societies tolerate unstable and scarce bank credit deepenswhen one considers the costs imposed by unstable and under-bankedsystems on those societies. In addition to the slower long-term growthproduced by under-banking, unstable banking systems entail other costs.Banking crises magnify recessions, resulting in greater job losses, and taxpayersare forced to pay the price of rescuing bankers from the consequencesof their own mistakes. Why do citizens tolerate this? Worldwide,that tab has been huge. Over the period 1970–2011, the median directfiscal cost of banking crisis resolution was 6.8 percent of GDP, and themedian increase in country indebtedness during a crisis was 12.1 percentof GDP. The cost of banking crises in terms of lost GDP (due to the effectsof credit contractions, heightened sovereign-debt risk, and currency collapseon economic activity) also tends to be enormous: from 1970 through2009, the median lost output during a banking crisis amounted to 23 percentof GDP.

In thinking about this puzzle, one shouldn't assume that taxpayershave always been willing to pay for bank bailouts. Taxpayer-funded bailoutsof banks are a recent phenomenon. Until the mid-twentieth century,the costs of failure tended to be borne by the bankers themselves, alongwith bank shareholders and depositors. Since then, however, the costshave been progressively shifted to taxpayers. How did bankers, regulators,and politicians come to impose these costs on taxpayers, and why dotaxpayers put up with bearing those costs?

This shifting of losses onto taxpayers is especially puzzling because ittends to produce much larger losses and deeper recessions than a systemin which shareholders and depositors bear the cost. A broad literature infinancial economics has demonstrated that a system in which shareholdersand depositors have money at risk imposes discipline on the behaviorof bankers: at the first sign of trouble, stockholders start sellingtheir shares, and depositors start moving their funds to more solventbanks. As a result, some banks fail, some of the holders of bank liabilities(shareholders and depositors) are wiped out, credit contracts as bankersrush to reduce their exposure to risky classes of loans, and economicgrowth slows. The result is painful, but not tragic. Most important, bankersknow the consequences of imprudent behavior and thus tend to maintainlarge buffers of capital and large portfolios of low-risk assets. As aconsequence, systemic banking crises are rare. Contrast that outcome withthe system that has come to be the norm since the mid-twentieth century.When losses are borne by taxpayers, the incentives of stockholders anddepositors to discipline bankers are much weaker. Bankers are willing totake bigger risks, thereby increasing the probability of failure. As a result,after 1945 banks in the world's most developed economies became morehighly leveraged and maintained smaller amounts of low-risk assets.