This is a talk I gave on 21 June at the European Political Strategy Center, which is the in-house think tank of the European Commission. The audience was very generous in listening to my presentation. The point I tried to make is that the capital markets union is an important project, but we should be careful to ensure that policymakers supplement the efforts to make capital markets more efficient with efforts to make them more resilient. This is an argument that I have made before and yet it is probably worth repeating. Given the dynamics behind Europe’s economic and financial crisis, there is simply too much at stake.
Thank you for giving me the opportunity to share some thoughts about the relevance of U.S. experience for Europe’s capital markets union. My argument is that U.S. experience is relevant both in terms of its successes and in terms of its mistakes. The most important lesson I draw from the United States is about the importance of managing or channeling the flight to quality when financial markets come under duress. In jargon, my specific concern is when a sudden increase in liquidity preference translates into a spontaneous return of home bias. In plainer language, what interests me is how we handle situations where investors decide to place priority on protecting the value of their assets.
Let me start with four qualifications.
First, I support the European capital markets union project both in whole and in its many parts. I think financial market disintegration is the engine behind many of the worst effects of the recent crisis and I believe that the Juncker Commission has identified a clutch of policy measure through which it can improve what we all acknowledge to be a sub-optimal situation.
Second, I am not here to tell you to copy U.S. institutional arrangements. We Americans have made a number of mistakes in integrating our capital markets that I would not encourage you to repeat. Moreover, the institutional legacies of those mistakes make it hard for us to put things right. It is probably more effort than it would be worth as well. Many times it is easier to make the best of a bad situation than to build something new from scratch. That is no reason for you to try and mimic the idiosyncrasies that we find in the U.S. context.
Third, I don’t think that there is some obvious best-of-all-possible-worlds in terms of institutional design – particularly when you are building on the legacies of so many different national experiences. Hence the best I can offer is a way to think about things based on my understanding of the U.S. case and my twenty-five years studying European integration. Along the way, I can help identify a range of institutions that you will want to develop in some way, shape, or form, but I understand that whether and how you do that is a decision that European politicians and policymakers will have to take.
Fourth, although I am presenting this work, the conclusions are the result of a long collaboration with Geoffrey Underhill of the University of Amsterdam and our research has been supported by the SWIFT Institute in London.
My presentation has four parts. First, I want to talk about the United States as a financial system in terms of three different sub-systems – for distributing capital, for redistributing liquidity, and for managing periods of duress. Second, I want to introduce the challenge of integrating those systems across different regulatory jurisdictions. Third, I want to focus attention on the dynamics that emerge during periods of distress both in the United States and in the European Union. Finally, I want to use this analysis to put the Commission’s capital markets union project into a wider context.
Let’s start with an overview of the U.S. financial system. The temptation is to regard the United States as a model fully formed – with the banks, the exchanges, the futures markets, the infrastructure, and the regulators all in place. That said, it is probably more useful to think of the U.S. financial system as a work in progress. That work is built on goals and relationships, it manifests in terms of stocks and flows, and it operates over time as a function of feedback mechanisms and information exchanges. In other words, when I describe this as a ‘system’ I mean that in a methodological sense. And my goal is to try and understand both the boundaries of the system and its underlying dynamics.
That in mind, I find it useful to break the financial system into three different sub-systems. Let’s start with the simplest capitalist arrangement. You have savings and you have investment. The goal of the system is to collect savings and channel it into investment. Along the way, you run into all different kinds of risks associated with temporal mismatches and unforeseen events. We can build a parallel system for risk management or we can try to build that risk management into the institutions for converting savings into investment. In the United States, we have had a host of different arrangements to achieve those linked objectives – redistributing capital while managing providential risk. They include everything from savings and loan institutions to commercial banks and investment banks. Most of these institutions were created in specific places to serve (or create) specific local markets. As it emerged, the U.S. financial system was very fragmented and inefficient from a global perspective and yet it each step in its evolution offered some improvement for the local context.
A second sub-system focuses on the distribution of liquidity. Here I use the term ‘liquidity’ as a placeholder for the things you need to convert capital into highly-mobile and widely-accepted fungible assets. These assets are important more for their functional characteristics than as investment opportunities. They provide a medium of exchange, store of value, and unit of account, but they don’t necessarily or even usually offer a high rate of return. Hence when I talk about this liquidity redistribution, I am thinking more in terms of payment and collateral management, but you might also think about the whole host of post-trade activities insofar as the more efficient these are, the more easily you can convert illiquid assets into liquid ones. Although there was extensive use of commodities like gold or silver as money in the United States, much of the ‘liquidity’ in the U.S. system was created locally as bank paper by the same institutions that collect savings and channel it into investment. [Indeed, Americans still rely heavily on paper checks today.] Where the markets were large enough, we also saw the development of specialized institutions like clearing houses and exchanges. And, at a certain point in its history, for reasons of its own and in many ways unrelated to the matching of savings and investment, the U.S. government decided to issue a unified national paper currency.
Historically, the U.S. system was not very good at managing duress. Whenever there were periods of financial turmoil and savers sought to protect the principal of their capital holdings, they almost inevitably tried to liquidate their investments all at once. As a result, they overwhelmed the local institutions for liquidity redistribution by depleting their stocks of highly-mobile and widely-accepted fungible assets. Worse, since many of these assets were created by financial institutions, liquidity vanished as the credibility of the institution collapsed.
Ironically, the inefficiencies of the US system were its saving grace. Local financial crises created only local disasters. In ‘It’s a Wonderful Life’, George Bailey never worries that the failure of his family’s thrift would bring down the world economy. The exceptions were the larger markets. Fortunately, those larger markets also had specialized institutions for liquidity management that could be leveraged to strengthen the crisis response mechanism. That is what happened in 1907 when the big banks used the clearing houses to stop the collapse.
So we can imagine the early U.S. financial system as a collection of local institutions operating in markets of varying dimensions where local crises were locally contained. It was inefficient but resilient – at least until it faced a powerful symmetrical shock. The Great Depression of the late 19th Century was one such shock; the 1907 market crash was another; and the Great Depression of the 1930s was a third.
The policy response in this progressively worsening context was to strengthen the system for liquidity management in two ways – one is through the creation of the Federal Reserve System as a mechanism for providing emergency liquidity and hence also for shoring up market confidence; the other was through the introduction of inter-district liquidity transfers between the different Federal Reserve Banks. Much of the private-sector activity within the U.S. financial system remained local, and crises were still locally contained, but the system was more resilient because of the creation of institutions that improved the management of liquidity in times of duress.
During the postwar period, the United States engaged in a process a gradual process of domestic financial market integration. Important parts of that process started already during the Great Depression with the creation of institutions to facilitate home ownership. It gained strength through the gradual liberalization of exchanges, the expansion of investment banking activities, the liberalization of commercial banking, and, eventually, the emancipation of the savings and loans institutions. At each stage, the goal of the liberalization process was to increase the efficiency of the financial system at distributing capital by, for example, taking surplus savings in New York and making it available for home building in California.
In order to achieve this objective, however, the United States also needed to deepen the pool of available liquidity by creating new assets that were both highly mobile and widely acceptable. This is where banks started experimenting in the securitization of mortgages and the creation of tradable certificates of deposit. For these assets to be liquid, however, they needed some kind of external support, because the underlying assets or institutional commitments existed far away from the investment community and this additional support was necessary to overcome that community’s traditional home bias. You will recognize that this is where the federal mortgage institutions like Fannie Mae and Freddie Mac become important for our story. Federal deposit insurance also played a significant role.
Alas the integration of U.S. domestic capital markets had unintended consequences for financial institutions that operated exclusively at the state and local level. The problem started as these institutions found themselves priced out of their traditional investment markets and worsened as they faced increased competition for funds. As a result, the savings and loans moved into ever riskier asset classes. Even there, however, the prices of the assets and so the yields on the investments were distorted by competition from outside the local community. Should these cross-jurisdictional investors suddenly pull out of the local market, they would overwhelm the local institutions for liquidity management. Meanwhile, financial institutions with access to national liquidity supports would be at a competitive advantage. Funds would flow from savings and loans to federally insured commercial banks as quickly as they would flow from the local community back to New York or whatever other part of the country they originated. This is what happened during the savings and loans crisis in the 1980s. The point to note, however, is that while this crisis had local epicenters, it could not be locally contained. The federal government not only found itself having to backstop state-funded deposit insurance schemes but also to fund the resolution of a very large number of state-chartered financial institutions.
In the language of international economics, the savings and loans crisis triggered a ‘sudden stop’ on the economies of states like California, Colorado, Florida, Maryland and Texas. Now I am sure that many of you are going to argue that the United States could afford this because of its integrated federal fiscal system. That is the obvious assertion. But that is only partly true. The federal government was able to create the backstop for deposits and it was also able to resolve the failing banks. But economic performance suffered nonetheless. If you were to compare per capita income in Texas and Massachusetts, for example, you would see a sudden gap emerge between the two states and relative to the US average. That gap is even wider in terms of per capita federal transfers – Texas is poorer, on average, and also receives considerably less in federal transfers on average than Massachusetts. Moreover, the automatic stabilization of income taxes does not make up a substantial difference. Roughly 85% of federal income taxes – in terms of volume – are paid by the top 25% of the adjusted gross income distribution in terms of tax filers. The federal income tax system is remarkably progressive in that respect. And the implication is that it does very little for income stabilization, particularly in remote rural areas (like much of Texas) where high income tax payers are few and far between.
All of the Texas banks disappeared in the savings and loans crisis and it took the Texas economy decades to recover. The situation, in that sense, is not unlike what happened recently in parts of Europe. There is, however, another important difference. Much of the liquidity used in the Texas economy was the same as the liquidity used elsewhere in the United States. Here I am talking not only about the currency but also about the highly mobile and widely accepted assets that are used for payment and collateral. This did not stop the bank-sovereign nexus entirely. State governments and deposit insurance mechanisms could still go bankrupt trying to save their savings and loans institutions – which is why the federal government had to step in. The federal government may have a policy about not bailing out states after they go bankrupt but that does not stop the federal government from preventing banks from going bankrupt in the first place. Federal income taxes cannot cushion a fall in per capita income for people who pay little or no federal income tax in the first place.
Of course not all forms of liquidity were equal during the savings and loan crisis. Some assets that were both highly mobile and widely accepted before the crisis were no longer ‘liquid’ once the crisis struck. The list of suddenly illiquid assets is long and includes many of the same instruments that facilitated the integration of financial systems across the United States in the first place. In this way, the institutions that connected markets to make them more efficient also connected them in ways that made them more vulnerable and in ways that ensured a crisis that started in any one part of the country would have a greater impact on all the rest.
Here I want to draw a comparison with recent European experience. In doing so, I want to highlight similar dynamics. In the late 1980s and early 1990s, Europe’s heads of state and government sought to liberalize national capital markets. That was the first of three attempts to create a capital markets union. As part of that effort, they made it easier for savers in any one part of Europe to purchase assets in any other. They did not eliminate home bias and neither did they make investors suddenly more eager to take on risk. But they did make it possible to move savings from where it is in surplus to markets with attractive opportunities for investment. The Italian sovereign debt market offered one such investment opportunity. Italian sovereign debt instruments are liquid in the sense that they are highly mobile and widely accepted; those same instruments are also widely used in Italy as collateral and even for payment. In 1989, roughly 4 percent of Italian sovereign debt instruments were held by foreigners; by the end of 1998 that figure was just over 27 percent; and by 2007 the share of Italian sovereign debt instruments held by foreigners was more than 40 percent.
This increasing foreign presence in Italian sovereign debt markets is a good indicator of European financial market integration. That integration did not transform the Italian banking system, which – apart from Unicredit and Intesa San Paolo – remained very parochial and locally-oriented. Italian households did not increase their debt holdings dramatically, although many did get access to credit cards and long-term mortgages for the first time. And Italy did not experience a housing bubble or a significant current account deficit. In other words, Italy participated in Europe’s capital market integration and garnered relatively modest (albeit still important) gains in the efficiency of matching savings to investment. Through the introduction of the euro and concurrent reduction in nominal interest rates, Italy also benefited from improved access to liquidity.
And yet nothing in Europe’s capital market integration or the Lamfalussey process that succeeded and strengthened it prepared Italy for the sudden sale of Italian sovereign debt instruments by foreigners that took place starting in late June and early July of 2011. Before the end of the summer, foreigners sold more than EUR 100bn in Italian sovereign debt and took their capital back out of the country; by the end of 2011, they had sold close to EUR 200bn. Europe’s saving grace in this context was the real-time gross payments mechanism that connects corresponding institutions within the European system of central banks – also known as Target2. If that mechanism had not been in place to redistribute liquidity across the euro area, Italy would have required an unprecedented swap agreement to meet its liquidity requirements; alternatively, it would have experienced a combined currency and balance of payments crisis.
Just as in the US case, what looked like liquidity in normal times ceased to be liquid in times of duress; and just as in the US case, this created a sudden stop for the Italian economy. The difference is that the Italian banks also relied on Italian sovereign debt instruments for their treasury operations. Hence a threat to state solvency became a threat to the banks. Moreover, financial market participants were well-aware of this dynamic. Hence as soon as they started selling Italian government debt they also started shorting Italian bank shares. The Italian economy has never recovered from the sudden withdrawal of foreign capital. Instead, economic performance has deteriorated and so chipped away at the balance sheets of Italy’s still very parochial and conservative banks. And now we face the prospect that the bank sovereign nexus will work in the opposite direction. The worsening of the banks is threatening the government. Ironically, the new European banking resolution and recovery directive is making matters worse and not better.
So what does this tell us about Europe’s capital market union? To begin with, I think we can all agree on the goal of making the distribution of capital more efficient. That said, I think U.S. experience shows that any increase in efficiency is likely also to increase the requirements for redistributing liquidity. This is true particularly when you are integrating financial markets across formerly self-sufficient and locally-oriented institutions. In other words, it is important that efforts to liberalize capital market flows and to strengthen market infrastructures move apace. I think the Juncker Commission appreciates that challenge and the current raft of proposals offers significant promise in terms of both goals of the European financial system.
What has me worried is how resilient this system will be in handling financial distress. A more efficient and more liquid financial market is also more prone to large and sudden movements. The recent experience of the U.S. financial crisis provides stark evidence of that dynamic. So the question is whether Europe has the institutions in place to channel the flight to quality and to minimize the collateral damage. The European banking union is a step in the right direction but it remains incomplete. The single supervisory mechanism can ensure banks have adequate buffers and it can shore up market confidence, but that is not enough to avoid a major financial crisis. Meanwhile, the lack of a common backstop for deposit insurance and banking resolution is a source of concern. It is easy to see why pooling financial resources is controversial; but U.S. experience suggests it is also necessary – of only to prevent firms from being ‘too big to bail’ given the resources available at the national or local level. This is not a plea for fiscal union. It is certainly not to suggest that Europe adopt federal fiscal institutions like those found in the United States. Rather it is to argue that without some kind of pooled backstop then the credibility of deposit insurance and resolution financing will continue to vary across jurisdictions in ways that make it attractive to move liquidity across jurisdictions when markets come under duress.
Another concern is that Europe does not have a large enough pool of common liquidity into which it can channel the flight to quality without interfering with collateral needs or, ultimately, undermining cross-border payment. Remember, by liquidity I mean highly mobile and widely accept assets. This issue lies at the heart of the bank-sovereign nexus. Simply restricting bank exposure to home-country sovereign debt instruments will not solve the problem. Neither will commitments by the ECB to engage in unlimited purchases of distressed country assets or to expand its balance sheet until the crisis passes. Europe needs a more robust set of institutions to channel the flight to quality. It also needs a common pool of risk free assets that is wide enough and deep enough to meet the liquidity requirements of Europe’s integrated financial markets in times of distress. The capital markets union project is an important step in the right direction. What U.S. experience tells us, however, it is that – as important as it is – the capital markets union is only a step.Follow @Erik_Jones_SAIS