Donald Kohn delivered the following remarks during a panel presentation at the Federal Reserve Bank of Dallas conference “The Federal Reserve’s Role in the Global Economy” on September 18, 2014.
For my contribution to this panel, I will reflect on the role of the Federal Reserve in the last crisis—focusing on the 2007-09 period when the crisis was most intense in the United States, was tending to spread to the rest of the world, and I was inside the institution helping to formulate policy. The Fed played a central role in dealing with the international aspects of the crisis through this period. We provided liquidity to many foreign banks–to their U.S. subsidiaries and affiliates at our discount window here at home, but importantly and innovatively also to their home offices through a network of liquidity swap arrangements with other central banks, which I will discuss at length. But we also were central to the monetary policy response around the world and I will touch on this aspect of our involvement briefly at the end of my remarks.
The centrality of the Fed’s role reflects a number of factors. U.S. financial markets and institutions are themselves at the center of an increasingly integrated global financial system. Deep, liquid markets for dollar assets are a natural investment outlet for investors in the rest of the world, including many official entities. The dollar remains the most important reserve currency, many international transactions are denominated in dollars, and dollar foreign exchange and swap markets are critical venues for managing risk. The largest U.S. financial institutions are global in reach and any impairment of their normal functioning—their provision of intermediary, risk management, liquidity, and payment services– affects economies around the world.
Moreover, the buildup of risks in the period leading up to the crisis and the transmission of their realization around the globe had a lot to do with domestic U.S. markets. The U.S. was “ground zero” in the 2007-09 crisis years—especially the subprime mortgage market. U.S. subprime assets were widely held around the globe so problems with those assets were quickly transmitted overseas, and the U.S. was exposed to the decisions of foreign financial institutions and governments about how to respond. Those global interconnections were key to understanding the Fed’s response in the international sphere.
Barry Eichengreen and others have noted a “modern Triffin dilemma”. The original Triffin dilemma in the Bretton Woods exchange rate system pointed out the tension for the U.S. and the global financial system between providing for the world’s growing needs for dollar liquidity, which required that the U.S. run a current account deficit, and assuring that the growing amount of foreign dollar holdings was exchangeable into gold from what was a relatively stable stock. The modern Triffin dilemma highlights the tension between the growing demands for dollar assets—primarily the debt of U.S. entities—and financial stability. Stability could be threatened when the increasing demand for dollar assets induces a buildup in U.S. debt to levels that might be difficult to service, especially when it is based on inflated collateral values. In the run up to the crisis, foreign demands for U.S. debt were rising rapidly for several reasons.
One was the U.S. current account deficit, which reached 6 percent of GDP. The U.S. was spending far more than it was producing, importing the difference and financing those net imports by borrowing from abroad, resulting in a faster increase in net indebtedness than in income. This deficit was shaped to some extent by exogenous spending/saving decisions in the U.S.—i.e. the federal fiscal deficit stoked by tax cuts. But it also was the consequence of the choices of foreign governments—in particular the decisions of the Chinese and other governments to pursue export-led growth. This entailed artificially holding down the value of their currencies on foreign exchange markets and in the process accumulating huge dollar reserves. These were invested mostly in safe, liquid assets, and the demand for these assets depressed longer-term interest rates in the U.S. and globally.
The buildup of vulnerabilities was exacerbated by the investment choices made by foreign private parties—and here I will emphasize gross as well as net flows from overseas into dollar markets. Foreign banks looking for low-risk higher-yield investments latched onto various types of U.S. securities, most especially super senior tranches of pools of subprime mortgages. Their demand, along with demand from U.S. sources, fed the inventiveness of the U.S. financial sector in constructing a supply of increasingly opaque, but nominally very safe, instruments and contributed to the loosening of credit standards for subprime mortgages. Foreign financial institutions financed their holdings of these fundamentally long-term assets in large part with wholesale short-term liabilities. These included asset-backed commercial paper (ABCP) issued by SIVs conduits sponsored by the banks and deposits from money funds. They were also funded by domestic deposits converted into dollars in short-term swap markets. European banks were prominent in these trades.
On the eve of the crisis, then, foreign institutions were looking at a currency mismatch and a maturity mismatch, and they were relying on the illusion of liquidity, especially in the MBS market, to manage risks. I’m not blaming the crisis on these foreign governments or institutions. Obviously many of the same weaknesses (not the currency mismatch) and more were shared by U.S. domestic lenders and it was U.S. regulators that had responsibility for overseeing U.S. markets. But the actions of foreign banks and governments contributed, and the role of the Fed in the international aspects of the crisis very much reflected the character of the risks that had built up across borders.
After house prices started to decline and the adequacy of the collateral backing for many of those subprime loans was called into question, uncertainty about where the losses would fall disrupted interbank lending markets—both here and abroad. The functioning of these markets was impaired and funding tenors became even shorter. The runoff of ABCP increased the direct exposure to subprime and other mortgage loans of foreign and domestic banks that had provided liquidity backstops to these conduits and SIVs, further increasing uncertainty about solvency. The pull back by global banks as funding became more expensive and its availability uncertain, along with the onset of recessions in some industrial economies, transmitted problems to economies where banks hadn’t taken risks—including many emerging markets.
Swaps with foreign central banks for the purposes of allowing those central banks to provide dollar liquidity to their commercial banks was a major aspect of the Fed’s response and one that was new to this crisis. We saw it as a logical extension in interconnected global financial markets of a basic central bank function–supplying liquidity when uncertainty causes the usual funders of banks and other financial institutions to back away. In these circumstances, central bank liquidity becomes necessary to break or at least to damp the adverse feedback loop between funding difficulties and credit supplies to the economy.
From the onset of the crisis in August of 2007, the financial markets were characterized by disruptions to bank and later nonbank funding as uncertainty about the solvency and viability of counterparties mounted. The lack of transparency in structured investments and in the balance sheets of some complex institutions and the sense that events were moving fast with unknown outcomes meant that market participants could not and did not discriminate well between good and bad counterparties. Their actions came to be dominated by fear and a run to safety and liquidity resulting in a sharp cutback in lending to businesses and households as well as rapid sales of assets at declining market prices. Reduced lending and lower asset prices contributed in turn to a weakening economy, greater disruption of funding markets, and adverse effects on market liquidity. We were in a doom loop, whose intensity varied from late 2007 on, but didn’t fully abate in the U.S. until after the banks were recapitalized and the stress tests of 2009 brought transparency to their condition. This type of market response could not be countered by open market operations because the reserves and liquidity would not be distributed through the financial system to where they were needed.
As a consequence, from the beginning central banks utilized their lending facilities to respond to the emerging crisis. This was seen as classic Bagehot central banking: when funding is generally disrupted lend freely at a penalty rate against good collateral (valued as in normal times) to solvent institutions. This is the way to stem the panic, avoid fire sales, limit the reduction in credit availability, shore up confidence and enhance market functioning.
The first liquidity swaps to deal with the cross-border aspects of disruptions to funding markets were announced in December 2007 with the European Central Bank and the Swiss National Bank. Increasingly intense foreign bank bidding for short-term dollar funding was putting upward pressure on the federal funds and other short-term dollar interest rates. Sales by foreign banks were adding to downward pressure on the prices of mortgage assets and having adverse effects on the liquidity in U.S. financial markets.
By doing swaps with other central banks in addition to domestic discount window lending, the Federal Reserve was able to help relieve pressure in U.S. funding markets without itself needing to make judgments about the solvency of foreign institutions and without taking risks of lending to these institutions. And they allowed the foreign central bank to make the moral hazard judgment that necessarily accompanies any provision of liquidity insurance. This seemed appropriate since home country authorities were overseeing these large globally active banks, and it would be home country taxpayers that could bear the consequences if liquidity failure turned into capital failure and if the failure of an institution impaired the functioning of home country financial markets. Moreover, global banks are often managed on a consolidated basis; the swaps enabled the banks to borrow dollars where collateral was located and then to redistribute that liquidity around their systems. The initial press release emphasized that the swaps were intended to help the functioning of U.S. and global markets.
The December 2007 announcement was coordinated across several foreign central banks in addition to the ECB and SNB in context of a broad array of measures to relieve pressures and enhance the functioning of funding markets. For example, in the U.S. the same press released announced the Term Auction Facility for auctioning discount window credit here. The swaps were one element in a broad effort to make liquidity available so as to bolster confidence and reduce the adverse effects of market disruptions.
As the crisis deepened after the failure of Bear Stearns and then Lehman Bros., the dollar swaps network grew in scope and size. It came to include 14 countries; the ECB, Bank of Japan, SNB, and Bank of England were running TAF-like auctions of dollar loans, with no upper bound on what they swap with us and lend to their banks. At the end of 2008 $554 billion was outstanding to 9 different authorities; Mexico borrowed later so at least 10 central bank counterparties took advantage of these facilities. At the same time, of course, the scope and size of discount window lending in the U.S. was greatly expanding. The volume of lending through swaps came down rapidly as panic abated; this indicates that they were priced right in accordance with Bagehot principals—at a penalty to rates that would prevail if markets were functioning normally.
Did they work? Here’s the bottom line of an extensive study published in 2010 by William Allen and Richhild Moessner in a BIS working paper: “We conclude that the swap lines provided by the Federal Reserve were very effective in relieving U.S. dollar liquidity stresses and stresses in foreign exchange markets, so that the Fed’s objectives were substantially met. It seems plausible that had the Fed not acted as it did, global financial instability would have been much more serious and that the recession consequently would have been deeper. The effectiveness of the Fed’s actions was most likely due to the fact that funds were provided quickly, limits were raised flexibly as the financial crisis intensified, especially after the failure of Lehman Brothers, and that large amounts were provided via the swap lines. “
Still, the swap lines raised some difficult issues as the FOMC debated their expansion in the fall of 2008. The first issue was the boundary problem: who to include and exclude. This came to the fore in particular when the swaps were extended to emerging market economies. We included Brazil, Korea, Mexico, and Singapore, which were characterized as “four large and systemically important economies” in the press release that announced the expansion, but undoubtedly other countries saw themselves as fitting into that category as well.
In the discussion at the October 2008 FOMC meeting, Nathan Sheets, the director of the division of international finance, put forth three criteria. First, that each country has significant economic and financial mass so problems there can spill over into the U.S. (Singapore was a systemically important financial center.) Second, that each had been well managed with prudent policies in place so that the stresses they were experiencing were a consequence of problems in the U.S. and other advanced economies. And third, that the swaps would help—that their banks had experienced or were subject to stresses related to dollar funding. 
The FOMC members generally agreed with these criteria and found that the four countries in question met them. Nonetheless, it was uncomfortable for the FOMC to be the arbiter of the soundness of other countries’ policies, the liquidity requirements of their banks, and their systemic importance. The FOMC is always assessing the likely course of events in the rest of the world as they might affect the U.S. and progress toward the FOMC’s objectives, but this issue raised the required knowledge and judgment to a very much higher and more detailed level, and the results of who might be in or out could have major effects on the countries involved. And how would the FOMC monitor whether the funds were being used for the intended purposes and not to avoid needed adjustment.
The second difficult issue related to the availability of alternative sources of liquidity for the authorities—to what extent should the Fed insist on other sources being utilized before it became the lender or swapper of last resort. Several of the 14 had a large volume of dollar reserves that might be used to lend to their banks. In most cases, however, those reserves were being held for purposes of currency intervention, not liquidity provision to banks. Forcing them to run down their reserves before swaps were activated might send an adverse signal in exchange markets, where some were already under pressure. Moreover, in most cases they didn’t have enough dollar reserves to meet the potential liquidity needs of their banks and those limits would undermine confidence enhancing effects of the dollar loans. Effective lender of last resort requires the possibility of unlimited resources and the Fed was the only institution that met this criterion.
Another alternative might have been borrowing from the IMF. We were in near constant contact with the IMF as the swaps for Brazil, Korea, Mexico, and Singapore were being put together. The IMF’s resources are quite limited, however, and loans form the IMF were perceived as carrying a substantial stigma, as in the past those loans had signaled crisis conditions in the borrowing country and were accompanied by many conditions for major reforms and often austerity. As such they carried considerable political as well as economic risk for the leaders of the borrowing country. The IMF initiated a new facility that relied on pre-qualification and was for liquidity, not solvency, purposes and didn’t have the conditionality of other IMF facilities. But this facility was just getting started in fall 2008, and it was unclear whether it would work with enough resources and with largely unconditional access for sound economies and institutions.
How best to handle cross-border lender of last resort for a wide array of nations is still an open question—one that Steve Cechetti will address in his comments. It’s important that the global policy community address this now. Uncertainty about whether a lender will be available in a crisis will only contribute to reserve accumulation as countries self-insure, putting contractionary pressure on global growth and output.
As I noted at the beginning of this presentation, swaps and liquidity provision weren’t the only areas in which the Fed was deeply involved in international aspects of the crisis. Another was monetary policy. On October 8 2008 six major central banks, including the Federal Reserve, announced simultaneous adjustments of their policy stances, with a view toward “effecting some easing of global monetary conditions.” This coordinated action was unprecedented. I don’t know who made the first phone call that started the banks down this path, but the participation and leadership of the Federal Reserve were essential. Importantly the coordinated cut provided a mechanism to help the ECB turn away from its focus on inflation, which had precipitated an increase in rates in the Euro area in August. It was intended to boost confidence; the central banks were on the job coordinating actions on the thought that working together quite visibly would be more effective than acting separately. Bank capital injections, borrowing guarantees, and more expansionary fiscal policy would be required, but these would all take more time, and meanwhile central banks could act quickly and together.
More generally, the Federal Reserve, under the leadership of Ben Bernanke, led in innovating ways to ease financial conditions even after short-term rates had effectively hit zero—innovations that have been followed by other central banks. We cut rates aggressively at the early stages of the crisis and then after it deepened on the failure of Lehman Bros. Once at the zero lower bound for nominal rates, the Fed used combinations of asset purchases and guidance about future interest rate targets to effect as further easing of financial conditions in order to stimulate growth and limit disinflation. Other central banks may have implemented parts of this program earlier, but the Fed put it together, innovated as more became needed, and explained why it was necessary for global economic recovery, even if some other countries were uncomfortable with the resulting capital flows. Over time other advanced economy central banks have adopted many of the elements of the Federal Reserve’s program.
 Central bank cooperation and international liquidity in the financial crisis of 2008-09, BIS working paper 310, May 2010, page 75.