Sections

Research

Maintaining Financial Stability in an Open Economy: Sweden in the Global Crisis and Beyond

SUMMARY

Sweden is highly open to the rest of the world, dependent on extensive cross-border transactions
in goods, services, and financial assets and liabilities. Exports are now around half the size of GDP. Cross-border financial assets and liabilities are each 2½ times GDP. The banking system is more than 4 times GDP. Even more than in past decades, Swedish financial institutions and markets are pervasively linked to the rest of the global financial system.

Sweden has been buffeted by financial instability twice in the last twenty years. The dominant sources of the instability in the early-1990s crisis were domestic. In the recent global crisis, however, the underlying causes were predominantly external in origin, stemming from financial shocks emanating from financial markets and institutions outside Sweden.

Financial openness is essential to Sweden’s healthy economic growth. But openness comes with risks as well as benefits. Our report attempts to assess these risks and benefits. We analyze the policy responses of the Swedish authorities to the recent crisis and examine how policies might be adjusted to improve stability in the future. We advocate a continuing review of the desirability of adjustments in policies that would reduce Sweden’s external vulnerability.

When the financial crisis erupted in the fall of 2008 following the collapse of Lehman Brothers, the Swedish authorities responded with alacrity. In addition to adjustments in traditional monetary policy, they took a broad range of other collective-support activities including emergency lending, emergency market support, modification in government guarantees, and facilitating the orderly recapitalization or resolution of institutions coping with possible insolvency.

The immediate problem in the crisis was that assets previously regarded as safe suddenly became suspect. In its initial response, the Riksbank provided substantial liquidity to the banking system both in foreign currency and Swedish kronor at longer than typical maturities and, at different times, at fixed and variable rates. In making these longer-term loans, the Riksbank was doing something akin to the socalled quantitative easing pursued by the Federal Reserve and the Bank of England. To ensure that enough liquidity could be supplied, in addition the Riksbank relaxed collateral requirements and extended the list of eligible counterparties.

Also at the outset of the crisis, the Swedish National Debt Office held extra auctions of Treasury bills. The proceeds were lent out accepting covered mortgage bonds as collateral to help stabilize financial markets. A few days later, the Riksbank sold “certificates” which had interest rates and maturities similar to those of Treasury bills and which were just as safe and as liquid. Issuing Riksbank certificates was another way to enhance liquidity.

A further significant crisis measure was the National Debt Office’s guarantee program. Only newly issued debt was guaranteed. The peak use of the program came to some SEK 325 billion, about 10 percent of GDP, with some two thirds related to borrowing in U.S. dollars and euros. Two notable aspects of the guarantee program were its selective importance for a few institutions and the importance of foreign currency funding.

Still other emergency measures were appropriately adopted in the crisis. These included modifications in the deposit-insurance program, ancillary actions to support liquidity such as the extension of collateral arrangements, and attention to potential or actual insolvencies of a few individual financial institutions.

The complex effects of the various policy measures taken by the Swedish authorities during the financial crisis are difficult to disentangle. Sweden experienced significant declines in output and stock prices. Yet spreads on financial instruments were stabilized, albeit at above pre-crisis levels. Credit to households was relatively stable, and credit to companies rebounded fairly quickly. Importantly, the financial system survived the crisis essentially intact.

Our report here and there raises questions about the details of particular policy measures. And our analysis takes into account some larger questions. Did the authorities move fast enough? Were their actions too timid, or did they intervene too aggressively? Could more have been done in preparing contingency arrangements for managing crisis conditions? When reading our evaluative observations, it should be remembered how much more difficult it is to make decisions in a crisis than when looking backwards with the advantage of hindsight. Crisis actions have to be decided in the heat of the moment with very uncertain foresight.

All things considered, we judge that the Swedish crisis actions were commendably prompt and typically appropriate. The experience in the past crisis, moreover, augurs well for the management of potential future crises.

Our report discusses several aspects of the Riksbank’s conduct of traditional monetary policy during the crisis. The main monetary-policy instrument, the Riksbank’s repo rate, was forcefully reduced, cut from 4.25 percent to 0.25 percent in a nine-month period during 2008-2009. The rate reductions were slower than cuts by the Federal Reserve and slightly slower than those by the Bank of England. It is an open question whether the Riksbank repo rate should perhaps have been lowered more rapidly and whether reductions in the repo rate should have ceased when the rate had fallen to 0.25 percent.

Difficult decisions were necessary when the repo rate had reached the neighborhood of the zero lower bound. As we look in the rear-view mirror at the crisis period, we find the arguments for keeping the repo rate from falling below 0.25 percent not fully convincing. The hypothesis that a literally-zero policy rate would create significant problems has not yet been tested in any country. Nor does the recent experience with low rates appear to strengthen the arguments against a zero or even mildly negative rate. We believe that it would be worthwhile, in Sweden and elsewhere, to devote more resources to studying the issues associated with a zero lower bound for the repo rate, including whether innovative options might mitigate the hesitancy of central banks to cut policy rates all the way to zero.

The common belief is that the Riksbank, unlike the central-bank authorities in the United States and the United Kingdom, did not engage in large-scale asset purchases as part of their crisis response. Yet the Riksbank did make large amounts of fixed-rate longer-maturity kronor loans to the banks, against collateral, in July, September, and October 2009. Most commentators agree that the main purpose of those loans was to enhance the expansionary stance of policy. The collateral arrangements, the credit risk, and the term risk to the Riksbank of the loans were not very different from the credit and term risks that would have been associated with direct purchases of comparable-maturity securities from the banks. The ultimate effects on interest rates paid by households and nonfinancial corporations of the two options, direct lending against collateral versus explicit asset purchases, might not have been all that different either. Similarities between the two options, broadly speaking, were probably even more important than the differences. To put the point more provocatively, we contend that the Riksbank put its toes in the water with a policy having many effects similar to the quantitative easing pursued by the Federal Reserve and the Bank of England.

Our report emphasizes the forward-guidance aspects of the Riksbank’s crisis-period monetary policy. We ask, for example, whether the forward guidance for prospective repo rates might have projected somewhat lower paths and whether greater prominence might have been given to the uncertainty associated with the repo-rate paths and the paths for the inflation and output target variables. We take it as given that the levels of forward-guidance paths and what is said about the degree of uncertainty associated with them are both fundamental aspects of the communication problem. And our predisposition is that – during crisis times of severe financial stress – the uncertainty aspects of forward guidance should be emphasized perhaps even more than forecast levels. In our view, most central banks – including the Riksbank – insufficiently focused on the uncertainty aspects of forward guidance during the crisis period and did not give enough attention to how to incorporate their judgments about forwardlooking uncertainty into their communications with the public. In particular, we are inclined to believe that the Riksbank’s forward guidance in 2008-2009 said too little about the possible consequences of the severe uncertainty for the target variables of monetary policy. And, free from the time constraints facing the Riksbank staff, we believe that it would have been helpful to amend the procedures for presenting the uncertainty bands in the Riksbank’s fan charts.

The recent period 2010-2011, as the severe strains of the crisis were somewhat dissipating, has been characterized by a vigorous debate within the Riksbank about the most appropriate stance for projected repo-rate paths and more broadly about how best to manage forward guidance during exit from the crisis period. The differences of view within the Riksbank have resulted in a persistent division of the Executive Board into majority and minority views. These differences are a first-order issue. Modestly different projected levels for the repo-rate forward-guidance path are associated with likely significant differences in possible outcomes for the economy.

The division in Board members’ views can be explained in large part by differences of judgment about the appropriate analytical approach for making decisions, not least about uncertainty. It is a subtle and unresolved issue whether the analytical treatment of uncertainty in the preparation of forwardguidance paths should help to determine the choice among the paths. The issue is subtle because the existence of uncertainty, great or small, does not by itself constitute persuasive grounds for relying on one or another analytical approach. All approaches, no matter to what extent they are based on explicit models, should try to incorporate sensitivity to uncertainty. Existing models are unable to capture adequately the uncertainty dimensions of financial strains, whether severe or moderate. Hence all model-based analysis must be cautiously amended by judgmental adjustments. The difficult tasks for policymakers are to determine how best to combine model-based and judgmental analysis and how best to explain the process and its associated uncertainties to the public. The ongoing debate within the Riksbank is a prime example of how very difficult these tasks can be.

Analysts and policymakers alike have been forced by the global financial crisis into a much sharper awareness of the deficiencies of existing models used to guide monetary policy. Models of the transmission of monetary policy through the financial system to the real economy have been shown to be more inadequate than was realized before the crisis. One can now discern, fortunately, an intensification of research efforts to improve the modeling of financial behavior, including at the Riksbank. Eventually, modeling of macroprudential instruments and their effects will need to be integrated into the larger, general-equilibrium analytical frameworks underpinning all types of macroeconomic and prudential policy actions.

The turbulence of the last few years has altered the debate about how to conduct financial policies in at least four important ways. First, central banks, market participants, and analysts in general are taking much more seriously the view that traditional monetary policy should give higher priority to financial stability. Second, they are according new urgency to making improvements in prudential policies. Third, they are recognizing that traditional monetary policy and prudential policies have important implications for one another so that they probably should be coordinated if they are to be used to best advantage. Fourth, given these new preoccupations, government authorities and outside observers are focusing anew on the institutional allocation of the responsibilities for the various financial policies – within national governments and among international institutions. The latter sections of the report touch on all these issues.

Before the global financial crisis, most analysts expressed doubts whether the central bank’s policy rate should respond to a financial-stability variable in addition to responding to the usual output and inflation variables. After the crisis, however, the debate has shifted ground. The debate now is broader, more about how to inhibit systemic financial strains and how to support financial stability more generally. Opposition to the general idea of “leaning against the wind,” interpreted loosely as putting greater emphasis on financial stability, has softened somewhat. And even though crisis tensions have partially dissipated, the still vivid memories of the meltdown turmoil have encouraged more sympathy for attempts to reduce the probability of future crises.

Policymakers charged with traditional monetary policy will understandably look to prudential instruments for a major part of the task of ex ante crisis prevention. That proclivity, however, cannot rationalize a complete neglect of the issues of financial stability when making monetary-policy decisions. We share the increasingly widespread agreement that it is unwise to rely solely on prudential instruments for reducing the risks of financial instability.

Prudential policies encompass both micro and macroprudential policies that aim at reducing the risk of financial instability. Although consultations at the BIS play a key role in developing international guidelines under the Basel III accord, domestic authorities implement the policies. In addition to more general increases in capital requirements for banks, the ongoing international discussions are proposing measures to make these requirements countercyclical. Such countercyclical capital requirements (CCRs) can potentially be used to deal with “bubbles” and to moderate credit fluctuations for stabilization purposes. The BIS and others have analyzed how CCRs can be implemented in practice — and in particular which conditioning variables can be used to determine when to build up and when to draw down buffers. Some progress has been made, but more is needed to implement this type of requirement in a systematic way.

Sweden now has in use a loan-to-value ratio cap as a prudential tool. Its effects are being debated; further study is both warranted and promised. Liquidity ratios designed to reduce maturity mismatches currency by currency are another prudential tool potentially important for Sweden (given Swedish banks’ extensive operations in foreign currencies). Many of the proposed prudential instruments are promising for use, in Sweden and elsewhere. But it is still too early to generalize confidently about how effective they will be in reducing vulnerabilities.

Buffers to deal with financial shocks can be built at the national as well as individual bank level. During the crisis in the fall of 2008, the Swedish authorities set up the Financial Stability Fund, which can extend support to troubled financial institutions, with a target size of 2.5 percent of GDP by 2025. In response to the external vulnerabilities in the financial system, the authorities also decided to increase foreign currency reserves in order to be able to support the financial system with liquidity not only in Swedish kronor but also in foreign currency. The size and funding of these buffers have been somewhat contentious.

In Sweden, as elsewhere, the authorities are addressing the issue of the degree to which “the financial system should pay for itself.” Our report considers two situations in which this issue arises. One involves the Riksbank’s foreign-currency reserves. If the Riksbank is to provide liquidity in foreign currencies to financial institutions on short notice without relying on central-bank swap facilities, it has to hold foreign-currency reserves. Funds are obtained through long-term borrowing and used to purchase short-term liquid assets. The cost of holding foreign-currency reserves is the difference between the long-term borrowing rate and the lower return on short-term assets. We believe that financial institutions that want access to emergency foreign-currency borrowing from the Riksbank should pay an “insurance fee” that covers this cost. Such a fee may reduce the amount of foreigncurrency business done by the financial institutions; as things stand now, that business is essentially being subsidized by the Swedish tax payer.

The other situation involves the financing of the Financial Stability Fund. The government started the fund off in 2008 with a contribution equal to .5% of GDP. Over time, as has been argued by the Swedish National Audit Office, private financial institutions themselves should replace the government’s initial contribution by paying that amount into the fund (taking accumulated interest into account). In addition, financial institutions are required to pay fees to build up the Fund until it reaches the announced target of 2.5% of GDP. As recommended by the EU, those fees should be lodged in an account that is invested in a geographically diversified portfolio of liquid assets. Investing in that way actually increases government assets whereas using the funds to buy Swedish government debt does not (because government assets and liabilities are increased by the same amount).

The potential for instability in financial activity cannot be attributed to cross-border finance per se. The causes are deeply rooted in the information asymmetries, the expectational and informational cascades, and the adverse-selection and moral-hazard problems that pervade all aspects of financial behavior, domestic as well as cross-border. Yet the cross-border features unquestionably magnify the potential for instability. How to allocate resolution responsibilities and associated costs among Swedish authorities and foreign authorities for complicated cross-border cases is very much an open question, now under active international consideration.

For Sweden, a small open economy with extensive financial links to the rest of the world, the development of macroprudential tools aimed at external vulnerability of the financial system seems to us a logical priority. Many practical aspects of such efforts remain to be worked out. If macroprudential financial policies have a promising future at all in Sweden, the prospects ought to be bright for those aimed at external-sector vulnerability. In any case, that is where the challenge may be greatest, and perhaps the payoff greatest, for successful measures and procedures.

Traditional monetary policy, to repeat, is relatively better suited for achieving stability of inflation and resource utilization. Macroprudential policies are relatively better suited for achieving financial stability. Yet all the target variables are affected by both types of policies. Thus even though a specialization in the two types of policies might seem appropriate, it would be inefficient – perhaps risky – if the two were conducted independently. Hence the logical question: to what degree, and how, should interactions between the two be managed? Should monetary policy and macroprudential policies be coordinated, even integrated? These issues are now high on the agenda in Sweden (as in most other countries).

Existing theory points in the direction of coordinated decisions. In general, decentralized noncooperative decisionmaking produces outcomes for a society that are inferior to the best attainable outcomes that could result from centralized decisions or the equivalent situation of full cooperation and information sharing among the decentralized decisionmakers. The broad principle is that coordination of decisions has a potential payoff.

Decentralized policymakers should, other things being equal, take account of the effects of the instruments they control on the entire set of target variables (relevant to all policymakers). If, despite the general principle, decentralized decisions without cooperation and information sharing are to be pursued, then it is incumbent on the advocates of that approach to identify benefits from decentralization – such as increased accountability, or improved specialization of function, or the avoidance of an undue concentration of power in a single authority – which offset the potential efficiency losses stemming from the lack of coordination.

Much of Swedish political thinking and political history, we have learned, has struggled with striking an appropriate balance between centralized and decentralized decisions. For Swedish financial policies, a significant degree of decentralization exists. Four separate authorities have important responsibilities. Microprudential policies are the province of Finansinspektionen. The Riksbank is responsible for monetary policies. The allocation of responsibility for macroprudential policies has yet to be clearly determined.

The Riksbank and Finansinspektionen engage in extensive information sharing and coordination at all levels. Integrating the two institutions could increase efficiency by removing the need for many of these activities. All three of the functions – macroprudential, microprudential, and monetary policy – would be under the same roof. Such a change, however, appears unlikely. Whatever the efficiency benefits of a merger of the Riksbank and Finansinspektionen might be, most of those with whom we talked in our interviews were either in favor of, or resigned to, a continuation of something like the current division of responsibilities between the two agencies. Most also did not envisage a major change in the responsibilities of SNDO and the Ministry of Finance. Several interviewees expressed reluctance to have the Riksbank gain more power relative to the other three agencies.

The difficult problem for Sweden, therefore, is how to catalyze coordination among the different authorities’ decisions regarding monetary policy and financial stability while still preserving the perceived advantages of decentralization. The approach under most active consideration entails the creation of a new umbrella institution, a “Financial Stability Council” (FSC). The FSC would have overall responsibilities for financial stability and crisis management. Detailed decisionmaking authority, however, would remain decentralized among the same four agencies who now share the various responsibilities. The FSC would be charged with engendering the desired amount of information sharing, analysis, and coordination of decisions. The presumption seems to be that the FSC would primarily act as a vehicle for joint consultation and peer pressure.

Problematic challenges lie ahead in working out precise responsibilities and detailed procedures for the FSC. Difficulties will arise, for example, when macroprudential-policy and monetary-policy considerations call for different actions. When the Riksbank participates in a shared-responsibility approach to macroprudential policymaking, will it be possible for Sweden to retain all the gains that have been attributed to political independence for monetary-policy decisions? The task of the FSC may be made more difficult by institutional features of the four Swedish financial-policy agencies that appear unlikely to change. It would be unfortunate if beneficial coordination were to be undermined by an understandable albeit regrettable tendency of decentralized institutions to insist on agency prerogatives predating the establishment of the FSC. We conjecture that the inter-agency problems can be resolved successfully provided that all parties are fully committed to the new institution.

Two procedural guidelines, as proposed by some advocates, would make it more likely that FSC joint recommendations to an individual agency would receive serious consideration. The first feature would be, following a FSC recommendation to an agency, a “comply-or-explain” obligation. The agency’s response might be subsequently published. The second feature would be a commitment to publish the minutes of FSC meetings, perhaps with some lag. We believe both procedural guidelines would be supportive steps encouraging constructive cooperation.

Our analysis in this report focuses on coordination issues within Sweden. But we are mindful of the broader European and world context in which Swedish decisions are made. Intra-European and international considerations are powerful constraints on Swedish policymakers. The complications arise for all prudential policies, microprudential and macroprudential. And they arise powerfully for traditional monetary policy. Swedish policy must take into account, and try to contribute to, the evolution of European Union financial policies.

Perhaps the greatest uncertainty facing Swedish policymakers – about financial policies but also about every aspect of Sweden’s economic policies – stems from doubts about the future of the Eurozone within the European Union. In late 2011 as this report was written, no one could clearly foresee whether a 17-member Eurozone struggling with sovereign debt issues would stay intact. Key aspects of the mandate of the European Central Bank were being debated. Although the issues were less explicitly discussed, it was also quite unclear how the European Union in the future would handle within its single-market framework the tensions between Eurozone countries and non-Eurozone countries. Those tensions are likely to become increasingly important for Sweden, as all the non-Eurozone countries — especially the United Kingdom and Denmark as well as Sweden – try to work out arrangements for themselves that are satisfactory and politically feasible.

The Eurozone member nations will be under continuing pressure to move faster toward measures of “fiscal union” (unless the Eurozone itself fractures). The European Central Bank will probably be pushed to play a stronger role as a lender of last resort for the Eurozone. Amid such pressures, it is unclear whether the issues of financial policies will evolve as a Eurozone responsibility rather than as a European Union responsibility. Perhaps even more in the next than in the last decade, the future of Europe – and Sweden within Europe – will continue to dominate financial, economic, and political discourse.