David Longworth noted the work of Borio and Lowe (2002a, 2002b, 2003, 2004) as well as the more recent work by Borio and Drehmann (2009), which looks at the growth in credit, housing prices, and stock prices as predictors of financial crises. He wondered whether these relationships are causal and, if so, whether central banks need tools to lean against the growth in these variables and what should be done in this area. Longworth also wondered whether central bankers may have exaggerated what could go wrong if monetary policy leaned at least a little bit against the cycle. This issue is often put inappropriately in terms of "pricking bubbles," as opposed to "leaning in a mild way." Furthermore, central bankers may have underestimated the effects on the real economy of bubbles. For example, during the run-up of the high-tech bubble, huge amounts of physical cable were being laid that were never going to be used. As such, there were real consequences in the future when investment dried up. Leaning against the bubble also involves leaning against what is happening on the real side.
Nick Le Pan responded that there are rates of credit expansion that do build up bubbles that no regulatory system can stand against when they burst. Hence, it is not realistic to expect a solution to financial stability problems that can ensure a close-to-zero failure in such economic and financial conditions. The interesting issue is what contribution various instruments can make to dealing with such problems. Le Pan argued against automatic credit-cycle capital requirements because of its target instrument problem. In his opinion, using such automatic capital requirements is akin to asking a blunt instrument like monetary policy to prick a particular bubble. Furthermore, counter-cyclical capital policy has a cost that is politically undesirable and unjustifiable. U.S. regulators were unable to stand against the decline in underwriting standards driven by Fannie Mae and Freddie Mac. This slackness in standards from these government-sponsored entities was then propagated across the system with huge implications. Le Pan thought that automatic counter-cyclical capital requirements would be unhelpful in dealing with such situations. Instead, he proposed joint collective action on behalf of the authorities who are concerned about financial stability and an effort on identifying credible, realistic tools instead of putting things on automatic pilot.
Charles Freedman argued that, insofar as there is information contained in credit growth and asset price bubbles that is pertinent to forming inflation outlooks, central bankers should and do respond to them indirectly, since inflation targeting monetary policy works by responding to inflation forecasts. Freedman noted that the recent bubble had much stronger effects on the economy than other bubbles because it caused an implosion of the financial sector. From a policy perspective it is acceptable if excessive risk taking on the part of society's wealthy people results in their private losses. However, during this crisis financial institutions were so interconnected and behaved in such a way that the risky behaviour of some financial market participants resulted in major consequences for many others. In sum, Freedman thought that central bankers should respond to the growth of credit and other similar variables since at times they can contain important information about the economy's path. However, responses should not be made too automatic, since at times they may contain false information.
John Murray mentioned that Lars Svensson is a great proponent of flexible inflation targeting, where all useful information is taken into account in making judgments and inflation forecasts. In this crisis, there was a sense that some potentially useful information was not taken on board. This, however, should be seen as an implementation error, not a flaw in the framework. The question is whether it would be sensible to go further than flexible inflation targeting if it were apparent that a real problem was at hand. While instruments other than monetary policy are needed for this, it is not clear as yet which tools would be useful, and it may be appropriate to use monetary policy to some extent in such circumstances if other instruments are not available.
Charles Goodhart summarized the current state of macro modeling by recalling that the workhorse model is actually a real business cycle model with price and wage stickiness. In the workhorse model, there is by assumption no default-everyone pays off their debts with probability one. Under these circumstances, money and banks are unnecessary and there is no risk premium. Thus the current macro model assumes away everything that should be of interest to central banks. The crisis has shown that we now have a need for models that incorporate default, banks, money, and risk premiums in a serious way. Goodhart expressed the hope that this crisis will have a large number of people working away at this problem. These insufficient models have been in use for so long because they worked fairly well. In normal times there are very few failures, risk premiums are low and stable, and money and banking is indeed a passive veil. Given that most times are normal, Goodhart wonders whether it is better to have a model that works only in normal times and a separate one for when a crisis hits, or to have a much more complicated model that is not needed most of the time.
Freedman noted that one reason why it has been so difficult to include credit channels in macro models is that their importance is episodic. He prefers a suite of models since the complexity of complex models with credit makes it much harder for senior management at central banks to deal with them in normal times. By way of comparison, a simple four- or five-equation model is much easier to explain than a DSGE model, even if there is another, more complicated model behind it. But we do have to do work on the more complex models that do incorporate credit.
Goodhart asked Le Pan why he seemed to rule out restraining bank dividend payments as a mechanism for keeping cash in the banking system. That is, when the banking sector falls into difficulty and confidence erodes, people start taking their money off the table. In these circumstances, instruments like contingent convertible bonds bring in very little cash to the banks compared to that being taken out through dividend payments. Goodhart noted that the amount of dividends paid out between August 2007 and September 2008 was much higher than the new capital that banks were able to raise through issues of bonds and equity.
Le Pan clarified his remarks by noting that the possibility of restraining dividends has always been a tool of regulators. His objection was to proposals that advocate automatic, system-wide dividend restraint in response to only a certain number of systemic banks being under stress. This might lead to a signaling problem that could suggest a broader lack of confidence. During a crisis, markets tend to assume that problems emerging in certain areas might also be present elsewhere. He recalled the common perception that in the intermediate ranges of severity, stopping dividends will negatively impact a bank's image and hinder its ability to raise capital when it needs it the most. In addition, differential tax treatment may make capital sourced from dividends more costly than that raised in the marketplace. During the crisis virtually all Canadian institutions stopped increasing dividends while some introduced automatic dividend reinvestment programs as a source of additional common equity capital.
Le Pan did agree with Goodhart that contingent capital was not the most effective way to raise a significant amount of new capital. However, it may be a way to reduce the cost to others and make a bank resolution more likely by making debt holders become equity holders. But, unless the official sector explains clearly the mechanisms of contingent capital instruments, as well as the expectations that underlie the authorities' responses, there is a risk of these instruments not working. During the crisis, a number of institutes had innovative instruments that were deliberately callable, but this feature was under-assessed by rating agencies. As a result there was a huge market surprise when the instruments were called by the institutions. Lack of credibility and clarity in authorities' strategies regarding contingent capital may also cause surprise, which in turn risks authorities recoiling from taking action.
Nick Rowe recalled Scott Sumner's theory claiming tight monetary policy in late 2008 might have caused the crisis and the recession. This theory notes that the path many economic variables have taken during the crisis matches the consequences of tight monetary policy-declines in asset prices, aggregate demand, and subsequently, declines in prices and nominal interest rates. The theory therefore suggests that the crisis was caused by tight monetary policy. Freedman dismissed this explanation because he could not see what actions taken by the Federal Reserve at the time could be defined as tightening monetary policy. Economic developments at the time can much more easily be explained by such matters as changing market views as to solvency of the financial institutions.
Huntley Schaller supported Goodhart's observation that many of the issues that played a prominent role in the recent crisis are absent in standard macro models. Many people have suggested that normal and crisis times are different in nature and therefore require different models. However, another perspective is that the issues central to crises may still be important in normal times but we have neglected them. John Cochrane (1994) claimed that technology and monetary policy shocks have difficulty accounting for a large proportion of the variation in output that we observe. It may be the case that some shocks arising from financial intermediation are important in normal times in accounting for fluctuations we see more generally, but they have not been prominent in the literature because we have not been looking for them, and because our models have not been oriented to thinking about these issues.
John Murray noted that there has always been interest and motivation to incorporate these issues in enriching the understanding of transmission mechanisms at the Bank of Canada, even in normal times, but technological and modeling constraints have restricted these efforts. Technological improvements and computing capacity is now making some interesting work more feasible, and recent events have brought focus to these efforts.
Freedman concluded the session and the conference by thanking the conference sponsors-SSHRC, TD Financial, Carleton's Vice President, Research and International, the Dean of Faculty of Public Affairs, and the Department of Economics-paper givers, discussants, session chairs, and audience members as well as the people working behind the scenes for contributing to the making of a most interesting and stimulating event.