By Dr. A. Michael Spence

Since 2008, PIMCO has had the privilege of a special consulting arrangement with Dr. Michael Spence, a Nobel laureate and Professor Emeritus of Management in the Graduate School of Business at Stanford University. He has dedicated his intellect to deepening our understanding of economics and repeatedly received recognition for his contributions to the field. In his capacity as special consultant he has brought his knowledge and experience into the core of our investment process as a participant at the PIMCO Secular and Cyclical Forums. In the following article, he helps us to understand the myriad forces at work in the current crisis. He also shares his thoughts on how to emerge from the malaise, highlighting:

Downward Dynamics
The economic downturn is still accelerating. There are increasingly persistent questions about the role and the effectiveness of government intervention. There also appears to be some confusion and disagreement about what government should and can do to ameliorate the situation.

Terms like ‘bailout’ are unhelpful as they suggest the government’s primary role is to help out particularly adversely affected people, households or institutions, whether they deserve it or not, with no clear statement of the purpose. While that may sometimes be one legitimate redistributive role for government, generally and in this context, it does not capture its principal potential function and contribution.

This note is about the role of governments (including central banks) in the midst of the crisis as distinct from whatever their roles are in normal times, or may have been in contributing to an unstable economic and financial configuration that led to the crisis. The latter is important in the context of redesigning the regulation and oversight of the system, once the crisis is over. But it is not the central issue now.

Given the history, a downturn was inevitable. Consumption, savings, leverage and risk spreads were at unsustainable levels and are being reset at different levels. A return to the pre-crisis path is neither possible nor desirable.  

Government’s main role in the context of the crisis was and is not to prevent this but rather to attempt to limit the length and depth of the downward journey, or put differently to try to prevent a downward overshoot – in asset values, employment and consumption. There are good reasons to think that such an overshoot with the attendant asset destruction and prolonged period of high unemployment is likely absent government intervention. 

In truth government has several roles to play in this kind of crisis. One is to protect the most adversely affected. A second is to act as an emergency repair service and fix things that break – like the credit system – when leaving them unattended would cause widespread and destructive collateral damage well in excess of the costs of the temporary fixes. These roles are understood, and in large part have been acted on in the current crisis. The Fed’s expanding balance sheet, capital injections into damaged financial institutions and direct provision of credit would be examples of what might be called the auto mechanic role. Keep the car running even if only on a couple of cylinders.

But there is a third important role, which has to do with introducing a circuit breaker or at least brakes into the powerful, reinforcing negative dynamics. It is this role that is the most complex and controversial.

Nash Equilibria and Tipping Points
This global economic downturn has several interacting adverse feedback loops involving asset prices, credit availability and real economy variables – consumption, investment, employment, savings, and profits. In past writing I have characterised this (at the risk of oversimplification) as a double downward spiral involving the interaction of balance sheets and income statements (including those of households). The powerful downward dynamics consist of self-reinforcing feedback loops within and between the parts. Asset price declines reduce consumption and with a short lag investment, employment and profits, producing further deterioration of balance sheets, prospects and asset prices. 

The important pieces of the system are all ENDOGENOUS. Where asset prices will and should level out is determined by the depth and length of the real economy decline and vice versa. This presents a fundamental conceptual problem with the concept of finding the bottom. Value investors are supposed to reenter and stabilise asset prices when asset prices overshoot in the downward direction. But overshooting is defined in terms of reasonable expectations about the present value of earnings, the evolution of balance sheets and of credit quality. And those are all endogenous and determined by the unknown future performance of the real economy.

Expectations are a key element of this dynamic. And they are also endogenous because they are influenced by data from markets and economies, and they prominently influence investment and consumption behaviour and hence outcomes in the real economy and the capital markets. People have a natural tendency to think of beliefs and expectations as accurate or inaccurate, realistic or not. But because they are endogenous and determinative of outcomes, it is better to think of them as self-confirming or not. 

At this point, negative expectations are producing highly risk-averse behaviour by investors and consumers, which leads to results that are in line with their expectations. 

Take consumers for example. Asset prices have plunged and their future course is highly uncertain. Households tend to consume 3-4% of a rise in net assets once they believe it is permanent, according to Alan Greenspan. This is consistent with what we observed in the run-up to the crisis. The same occurs in reverse. Asset declines with the prospect of further reductions produce contracting consumption patterns that will not begin to reverse until households believe that asset prices have stabilised and have the prospect of rising.

Similar incentives apply to investors; they are sitting on the sidelines waiting for asset prices to stabilise before (and with a lag) reentering the market as value investors. 

The current dynamic can be thought of as the playing out of a Nash equilibrium in a large and complex game, an equilibrium that no one wants.¹ The incentive structure of this game is somewhat akin to what is called the prisoner’s dilemma in game theory. Roughly speaking it is a situation in which all players have dominant strategies (that is, a strategy that is preferred regardless of what the other players do), and when they all choose the dominant strategy the result is inferior to another outcome that is achievable only if the players act in their collective interest and not independently in their individual interest.

If various players in the game could find a way to agree to simultaneously consume and invest a bit more, to shift their asset allocation a bit, the dynamics and the embedded expectations would shift.

But that is not practically possible. It is a coordination and contractual problem of such complexity that it cannot be solved in a decentralised fashion. Government’s most challenging and probably most important role is to act so as to adjust the payoffs and hence the behaviour in such a way that the new Nash equilibrium involves less income and employment loss and wealth destruction.

Now let me modify this analysis slightly. Investors and consumers, the players in this game, in reality have conservative and risk-averse strategies that are almost dominant in the following sense. Given that the vast majority is holding back for rational reasons, the preferred strategy is to defer consumption, rebuild assets and wait for asset market prices to stabilise.  But if an unknown but large fraction of consumers and investors switched to a less conservative strategy, then unemployment and asset prices would decelerate in the downward direction and turn. Then the rational individual strategy, once convinced of the new dynamic, would be to adopt the less conservative strategy.

This means that there are two qualitatively different sets of stable equilibria and a delicate (and unstable) tipping point between them. At the moment, we are on one side (the bad side) of that tipping point, the side where the extreme caution is individually rational and the high speed downward dynamics are in play. The government, by providing credit, and spending and investing for us (as if we had agreed to do it in a cooperative manner) is attempting to shift the actions and expectations together to the other side of the tipping point. If they are successful, then over time, expectations and behaviour will adjust and the non-cooperative Nash equilibrium dynamics will take us in the reverse direction. In effect, the government is in part simulating (albeit imperfectly) the cooperative behaviour that would shift the dynamics to those associated with the more beneficial side of the tipping point.

Even with large scale government intervention, this shift will not happen instantly because expectations do not shift overnight. But it can happen quite quickly, particularly in the asset markets. Now that government actions in the real economy and the financial sector are about to gain traction, it is a critical period particularly for investors.

One implication of the above framework is that a little intervention has a high probability of being ineffective, while a much more substantial intervention has a higher probability of success. The reason is that the more modest intervention will not cross over the tipping point and shift the dynamics in a favourable direction. 

Why, in principle, should citizens vote for taking the costly actions that might shift the expectations and hence the dynamic evolution of asset prices, credit spreads, consumption, employment and investment? In the short run the actions are costly in terms of present and future taxes. The government is spending our money for us. In addition some of the actions cause or risk collateral damage. The short and correct answer is that citizens will support the interventions in principle if the expected benefits in terms of increased income and wealth (that is, reduced income loss and wealth destruction) exceed the costs. Otherwise the actions simply amount to a set of redistributive activities with an important intertemporal or intergenerational component. Thus the potential to meet this benefit-cost test is dependent on the ability to solve the coordination problem and hence reduce the magnitude and length of the downward dynamics. 

Alternative Views
I believe that there is a reasonable chance of favourably altering the dynamics and that the effort is worth the risk. But it cannot be timid or it won’t work.

Others disagree and believe that the downward resetting of asset values, debt levels and savings rates are basically not changeable in terms of magnitude or speed (except for the aforementioned emergency repair work on the credit side).² Those who consciously or implicitly take the view that beyond the emergency repair service, government is largely powerless, therefore believe that the government’s activities should be much more circumscribed. Essentially we just have to ride this deleveraging and asset price resetting process out. Some would argue further that the collateral damage from intervention – unpredictably abrogated property rights, potential fiscal instability – is significant and comes with little provable benefit.

There is no definitive body of theory, model or body of evidence with respect to whether government intervention can shift the dynamics in a positive way. Judgment is required. The argument for the large scale intervention model as outlined above is that investors and consumers have huge amounts of discretion especially with respect to timing, and when they exercise that discretion rationally in the direction of extreme caution, the negative dynamics become more virulent. We would probably collectively rationally choose not to do that if there were an effective way to commit together to an alternative course.

Implementation
Let me turn to the interventions that might partially solve the coordination problem and begin a deceleration. To be maximally effective the interventions need to operate on each of the interacting parts of the system. These include the real economy, the financial system and credit provision, and asset prices (balance sheet effects); and this must be done at a global level or, at least, in countries that are systemically important (which goes beyond just the US or even the advanced countries). The stimulus package is in place and is intended to operate on the real economy, specifically on consumption, disposable income, investment and employment in the public procurement portions with hoped-for multiplier effects.

The stabilisation of the financial system is a work in progress and is less far along, though significant progress was made at the end of March in the form of a trillion dollar commitment by the Fed followed by a fairly detailed public private partnership program to acquire toxic assets. Credit is still tight and spreads on a variety of debt instruments like high grade corporate bonds are very high. Major financial institution balance sheets are damaged and remain in a fog created by the uncertainty, quantity and value of the toxic assets and the underlying collateral. 

I am going to skip over the complex debates currently under way as to how best and with the least collateral damage to stabilise the financial system. It would be too big a diversion. Suffice it to say that there is broad agreement that the impact of the stimulus package will be very much diminished if the financial system remains in an unstable and damaged state. The reason is that the multipliers from a stimulus package are substantially diminished when credit is unavailable or tight and when asset prices are expected to diminish net wealth further. The reverse argument is also true. It is hard for asset prices to stabilise when the real economy is sinking.

Asset prices are declining rapidly and as far as I can see, there is no plan other than the indirect effects of the items above to alter their course. In particular, a program to stabilise housing prices is needed. 

It is a commonly expressed view that a combination of the stimulus package (I will come to the international dimensions of this a little later) and an aggressive program to stabilise the financial system and restore normal credit channels will have a major positive effect. I am inclined to agree but I think it is a bit optimistic. The downward trajectory of asset prices has its own momentum. And while the restoration of cancelled dividends, an uptick in profits or sales and an improvement in credit quality would surely help, there remains the issue of who is going to jump in first in the bargain hunting phase. Right now, and given the history, the benefits of waiting appear to outweigh the benefits of jumping in. This could change very quickly. But we need a trigger. It is the classic coordination problem in almost pure form.

If this were Canada, I would recommend the creation and capitalisation of a crown corporation with a mandate to start buying bargains in the asset markets. It would be a transient entity with a clear mandate to sell at a profit when the markets turn. Ideologically this is sufficiently remote from our conceptions of government’s role that it is unlikely to be a component of a coordinate intervention strategy in the US.

As noted above, expectations are a very important endogenous part of the equation. This means that apart from the substance of the government interventions, speed, clarity and communication are crucial ingredients. Until recently these have been in short supply. As Alan Greenspan has said, extreme caution comes from fear, and fear in turn comes not from risk, but from uncertainty. Government is a potential source of risk and uncertainty, absent clarity and communication. It is a little bit like being locked in a room with an elephant; quite risky unless you know in advance where the elephant is going to step.

Constraints on Government Intervention
There are many. In our system there are multiple voices and multiple decision-makers. It is hard to act quickly, decisively and with clarity and effective communication. 

There is philosophical and political resistance to government participating in the economy in certain ways. These constrain the options.

There are potential conflicts of interest.  Government as major shareholder of a bank or other financial institution has different (social) objectives than the private shareholders. 

Then there is the debt. We are financing most of the short run cost of intervention with debt, held domestically and by foreign investors, private, sovereign and mixed. That amounts to deferring the cost to those who may not wish to pay the cost. Since whatever happens, we are not going to know what the counterfactual would have been (though lots of words will be spoken about it), it will be hard even after the fact to “prove“ that the benefits outweigh the costs for the future taxpayers. 

I think the Administration is approaching this in the only sensible and practical way, and that is to combine the intervention plans and strategy with credible plans for reducing deficits and debt levels in the future once the crisis is behind us. That of course entails normalisation forecasts and those in turn depend on the interventions succeeding. It is a hard problem.

The bottom line is that the government is trying to solve a problem of effective collective action that practically cannot be solved by individuals and institutions by themselves. This is sometimes described as the government having the only balance sheet that is big enough to intervene in a meaningful way. But that potential balance sheet comes from having the authority to issue debt and tax within broad limits, that is, to redistribute income and assets across individuals and markets and intertemporally. Its appropriate use is to act as a temporary substitute for risk-averse consumers and investors and damaged financial institutions in order to apply a brake to the downward dynamics.

International Dimensions of the Crisis and Intervention Strategies
The economic and financial crisis has become global with the transmission channels reasonably well understood. A long period of increasingly open markets in capital and goods and services has increased interdependence. The rapid fall in aggregate demand in the advanced economies reduced effective demand in the rest of the world. Exports and production are falling rapidly everywhere.

On the financial side, while much of the world did not hold toxic assets, capital moving to shore up balance sheets in the US and Europe caused a near instantaneous credit tightening everywhere. The evidence of rapidly reversing capital flows is the near universal depreciation of currencies in the developing world relative to the euro, dollar and yen. Even China, which was appreciating in a managed way up until the summer of 2008, levelled off with respect to the dollar. For those like Eastern Europe who borrowed in euros or dollars, the increase in the debt burden caused by the depreciation is destabilising for them and for their financing banks in Europe.

If intervening in a coordinated way to alter the downward dynamics is economically and politically difficult in the US, the complexity increases exponentially in the international setting given the highly interconnected global economy. 

This is a very large and complex subject, too large for this note. I will confine myself to some observations.

On the real economy side, the hoped-for “first best” solution is a set of coordinated stimulus packages combined with a commitment to avoid protectionism. Because of real economy linkages in the global economy, stimulus packages leak. Hence the perceived need for coordinated and simultaneous commitment. This has the same incentive structure as the coordination problem earlier described. The two self-interested strategies are (1) no stimulus and free ride or (2) stimulus package with provisions that caused the stimulus to benefit mostly domestic individuals and institutions, i.e., protectionist measures. Further there are asymmetries in the leakage that result from different amounts and kinds of trade exposure. So depending on the country (its size and trade exposure) either of the above two strategies may be the optimal one.

My personal view is that this problem is too hard to solve and that it will likely be impossible to achieve the first best outcome. The second best, which people won’t like, may be stimulus packages with some protectionist measures and hopefully a commitment to get rid of them when the crisis has largely passed. If we insist on no protectionist measures the alternative may be an extreme reluctance for citizens to support stimulus packages. We have to be realistic about what the alternatives are. The choice may be between stimulus packages with protectionism or no stimulus packages, rather than between stimulus packages with and without protectionism.

Stabilising the global financial system is also challenging in part because it raises the issue of who is responsible for what. Is Europe largely to be responsible for instability in Eastern Europe and its collateral damage and why? Generally the European Union (EU) has a harder problem in this area (and in the stimulus area) because the connectivity and leakages are very large. 

One thing has become very clear. A trusted, well-capitalised (in the sense of access to very large reserves – on the order to two trillion dollars) and well-governed/legitimate International Monetary Fund (IMF) is a crucial asset that needs to be created to carry into the future. Had that been in place, the rapid provision of credit to the developing world would have acted as an important circuit breaker in the exportation of the credit tightening. Further, relying on the IMF side-steps many of the issues of who is mandated or responsible for what. A crisis is no time to have an extended debate about this.

Concluding Thoughts
Our government and a number of others are embarked on a course of using resources to try to shift the dynamic market trajectory away from a depression-like scenario toward just a deep and longish global recession. To do this, they need support and a rationale. The rationale is that there is a credible coordinated set of interventions that are likely to produce a different Nash equilibrium dynamic and that the benefits in terms of reduced income loss and wealth destruction are worth the cost and risk. 

Finally, there will be collateral damage.  There already is. The onset of the crisis has led to widespread distrust (and anger) with the financial system and its regulators. Property rights will be adjusted and risk aversion generally partially embodied in future regulation will raise the cost of capital and slow growth. And it will take time to overcome the risk aversion and mistrust. How much will likely depend on the success of government in its collective action role in the crisis.

¹ A Nash equilibrium, sometimes called a non-cooperative equilibrium, is a configuration in which each participant is acting rationally given the behaviour of the other players.
² The vast majority of participants and analysts would agree that keeping the car running is a crucial if potentially expensive function – and for that set of actions the costs of inaction are so large and widespread that the benefit cost test is met. Or to put it differently, a complete credit lockup is the depression scenario.