Monetary Policy Under UncertaintyCapítulo de Martín Redrado para el libroThe Oxford Handbook of Latin American Political Economyeditado por Javier Santiso y Jeff Dayton Johnson 1. Introduction
The intrinsic complexity of economic relations, i.e. the interconnection among the relevant variables, the changing behavior of the economic structure and the interpretation of the different phenomena by the economic agents, forces monetary policy to be developed in a highly uncertain world.
The analogy of the car driver is appropriate to describe the monetary policy process. In this analogy, the economy is represented by the car, the monetary authority is the driver and policy actions are taps on either the brake or the accelerator. Accordingly, if the economy is running too slowly, then policy makers cut the interest rate (pressure on the accelerator), thereby stimulating aggregate demand. On the contrary, if the objective is to reduce the level of output, then the Central Bank switches to the brake by raising the interest rate.
However, monetary policy-making is far from simple, which renders the previous analogy misleading. First, policy-makers deal with informational constraints that are far more severe than those faced by real-world drivers. The second problem with this analogy arises from the central role of private-sector expectations in determining the impact of monetary policy actions. Therefore, if making monetary policy is like driving a car, then the car is one that has an unreliable speedometer, a foggy windshield, and a tendency to respond unpredictably and with a delay to the accelerator or the brake (Bernanke, 2004).
We live and, therefore, have to pursue monetary policy, in a highly uncertain world. Alan Greenspan defined this phenomenon in his own words: “uncertainty is not just an important feature of the monetary policy landscape; it is the defining characteristic” This statement may seem evident, almost redundant, especially to those of us who have had to weather crises, either as economic policy makers or in the private sector under changing financial conditions.
Financial crisis demonstrates that being a central banker is not an easy task. And, being a governor of an emerging market economy with a fragile institutional framework could be even more complex. Trust me: a year at the Central Bank of Argentina could be like a whole tenure at the Fed!
The purpose of this paper is to describe how the presence of uncertainty limits the effectiveness of monetary policy and, as a result, its implementation.
Monetary and financial policy in developed economies during the so called “great moderation period” tended to focus on price stability, with little regard to developments in asset and credit markets. Financial stability was left in the background, perhaps in the belief that it would follow from price stability and that risks could be properly limited by banks’ self assessment mechanisms. In contrast, developing countries have long (and painfully) learnt the lesson of the importance of financial stability on macroeconomic performance –as evidence by recurring crises in the 1980s, 90s and 2000s, in different circumstances, but sharing the substantial role of some sort of financial imbalance. Thus, if only a short time ago monetary policy could be adequately described by a simple “Taylor rule”, nowadays a more complex central bank “reaction function” is called for. This entails a rebalancing of monetary policy objectives, putting the proper functioning of the financial system on par with macroeconomic stability. Thus, newer approaches such as quantitative easing, enhanced regulation and supervision techniques, have taken the stage. In the developing world, we have seen in recent years policy strategies that include foreign reserve accumulation as a form of self insurance, active foreign exchange operations to mitigate excessive volatility, and the use of alternative mechanisms to provide liquidity when needed, recognizing the impact that the dynamics of financial conditions could yield on macroeconomic stability.
In emerging markets, we have to conduct monetary policy under deep uncertainty. Despite the considerable progress in many countries of the emerging world, it is undeniable that our macroeconomic behavior substantially differs from that of countries with higher relative development. Shock absorption mechanisms in developed economies (such as financial depth) are not only absent, but also work as amplifying forces, deepening the effects of shocks. All in all, the same external disturbance that might cause temporary internal imbalances in a developed country might lead in our economies to diverging paths in key variables, and this would imply a costly return back to equilibrium.
This paper embraces this unchartered territory. Section 2 reviews the recent contributions to the literature so as to identify the influence of uncertainty on monetary policy. Section 3 analyzes how uncertainty affects policy-making, particularly in emerging economies, while Section 4 includes the conclusions on the implications of uncertainty on monetary policy-making.2. Monetary policy and uncertainty
Until recently, the notion of uncertainty was not systematically embedded into the theoretical body of monetary policy. The “world”, as defined by a given model, was considered to be perfectly known by decision makers. Consideration was given, at the most, to the notion of risk which, unlike uncertainty, entails knowing the probability distribution function of an event.
However, thanks to the contributions made by Walsh (2003), research on the conditions under which the monetary policy develops has underlined uncertainty as a core issue.
Some time ago, research revealed a consensus that implicitly assumed that central bankers knew the true economic operating model. It assumed they were able to correctly follow all relevant variables and could accurately determine the different impulses affecting the economic system. Under this perspective, the only remaining “uncertainty” was, in fact, the actual realized value of these random disturbances.
However, monetary policy decisions are affected by different uncertainty sources. For instance, information problems hinder the determination of the different phases of the economic cycle. In addition, policy makers ignore the exact nature of economic relations and of the equilibrium values of the relevant variables. In turn, economic agents play a critical role when they build their expectations, thus impacting the effectiveness of policy decisions. Besides, the latter is compounded by structural conditions under which both policy makers and private agents must address a learning process about the new environment and modify their behavior accordingly.
The different uncertainty sources are reflected in the specific problems that Central Banks face when designing and implementing the monetary policy. Even more, they restrain the way information is processed and the procedures to determine the most adequate intervention and operation rules.
However, most of these developments are applicable to advanced economies, relatively less prone to a crisis (or, at least, to frequent crises) and characterized by a high degree of macroeconomic stability.
Despite the remarkable advances of many Latin American countries in recent times, it is irrefutable that the macroeconomic behavior of the region is considerably different from that of countries with a higher relative development. The recurrent shocks have disruptive effects on output evolution. Particularly, more volatile economies exhibit lower growth rates.
The shock absorption mechanisms operating in developed economies are not only absent but also operate as amplification mechanisms in the less developed countries. Thus, the impact of the shock is more severe. So, returning to balance tends to generate high costs. It is not a coincidence that in developing countries consumption is more volatile than output, exactly the opposite of what happens in an advanced economy.
As pointed out by Sargent (1993), uncertainty can be especially intense in transition economies. In these countries, the “right” model, the value of structural parameters, the transmission mechanisms or the nature of shocks are not accurately known. Under such circumstances, the lessons drawn by agents can sharply increase uncertainty, translating into adaptive responses that alter the economic structure on a permanent basis.
The uncertainty concept inherent to every economy feed one other in the emerging world. Thus, the idea that the uncertainty of the environment is crucial when analyzing monetary policy options does not come exclusively from methodological or epistemological considerations. It also stems from the concrete difficulties faced by policy makers under changing scenarios, which frequently occur, unfortunately, in Latin American countries.
Uncertainty about how the economy really works (about how monetary policy transmission mechanisms operate) leads, on many occasions, to implement a monetary policy more conservatively than if the structure of the economy were known in depth. Emerging countries’ uncertainty aversion is evident to some extent in what some authors have called “fear of floating”, in relation with the allegedly excessive exchange rate stability (several additional reasons account for this phenomenon of relative stability in the exchange rate, such as the scarce development of heading instruments), while the same phenomenon in industrialized countries is known as interest rate smoothing. An element referred to the Central Bank’s credibility (not to appear erratic in the reading of the cycle and, thus, in the modification of the rates) may be behind the interest rate smoothing cycles of rate increase and decrease.
Recent studies that economic agents go through a “learning period”, where their behavior is not necessarily compatible, in the limit, with that of agents with rational expectations.
In other words, economic players are forced to learn while interacting, whereby adding more complex behaviors which are not always captured by standard models.
In exploring the policy-makers’ task, several papers (such as Levin and Williams, 2003) have shown the significant challenges posed by the analysis under uncertainty. Thus, the authors have emphasized that optimal policies in some countries may perform poorly under different conditions.
This has led to the notion of “robustness”: it is highly desirable that monetary policy rules sustain themselves against changes in the economy’s behavior. For example, a given policy may be considered optimal and simultaneously have very negative consequences if the true model governing the behavior of the variables differs from the model assumed. Instead, an alternative policy could be somewhat less effective if the economic model coincides with the model undertaken. Simultaneously, it could also be less harmful if the operating conditions go against those initially assumed. Against this back, a second best –though not necessarily optimal under all potential circumstances– could be considered more robust than the first alternative.
Given the incomplete knowledge about some key structural aspects of the economy and the asymmetric distribution of the costs and benefits of specific outcomes, Greenspan has advocated a “risk management” approach for monetary policy definition (Blinder and Reis, 2005).
Under such circumstances, the risk management approach proposes a forecast-based policy, whose purpose is to combine economic models with the opinion of the experts to project scenarios. It is singular because it focuses on the analysis of the probability distribution of economic outcomes. Therefore, low-probability –but potentially harmful– events are included in the analysis. Under this approach, what matters is the distribution of them, and not just the average or most likely outcome, to decide monetary policy actions (Greenspan, 2004).
Consequently, the outcome of a low-probability event with severe adverse consequences may then be considered riskier than the costs of having insured against a contingency that does not occur.
Simple instrumental rules may have a good performance similar to that of much complex “optimal” reaction functions. There is consensus on the fact that pursuing such rules may provide an adequate reference framework for decision-making by monetary authorities. Unusual –and sometimes usual– circumstances require giving a preeminent role to the analysis and judgment of monetary policymakers, in line with the principles of the “risk management” approach. Model-based rules should thus be an important supplement to the judgment based on the careful analysis of empirical evidence and data, but they cannot replace it.
When designing and implementing monetary policy, it is necessary to take these considerations into account, as well as the characteristics of the local and international macroeconomic environment. Otherwise, the monetary policy will not only be inconsistent but will also become an additional source of uncertainty, as it occurred in Argentina.
Recurrent macroeconomic instability episodes have been one of the most distinctive features of the aggregate operating dynamics of this country. It is not a coincidence that, in the last 25 years, the Argentine economy has been off the dynamic economic stability path (defined as the range between two standard deviations from the long-term trend) one third of the time, against 18% for Australia or 25% for Brazil. These countries are comparable in terms of resources and position in the world; therefore, Argentina is expected to be somewhat symmetrical regarding the impact of external shocks.
These phenomena have been severely harmful for long-term performance and not for free in terms of welfare: excessive volatility is likely to be the main factor behind Argentina’s economic stagnation in the last decades of the past century. (See Figure 1.)3. Uncertainty and monetary policy implementation
The last 25 years have been prolific in terms of the lessons learned from crises of different origin and consequences for the monetary and financial system. The channels through which excessive volatility affects economic performance are different.
Argentina has undergone two types of crises:
• Crises arising from ex ante nominal volatility, characterized by a highly persistent variation in the nominal signals received by agents (high inflation), leading to a shortening of the horizon, conservative behavior, financial anemia and money functions' loss of value.
• Crises arising from nominal stability established by law (such as convertibility), which encourage a bold behavior by agents and “shortcuts” to financial depth but in the end cause mass defaults if the rule is not accompanied by consistent macroeconomic policies in the real economy.
In these economies, where society has developed a high risk aversion and the need to prevent the next crisis becomes a priority objective, the demand for macroeconomic policy coordination is more critical. If there are doubts about the inter-temporal solvency of any policy set, the monetary policy conventional room for maneuver can be limited. Likewise, the effectiveness of the traditional tools is affected. For example, Argentina’s credit channel is shallow, as shown in the last 25 years (See Figure 2).
When designing the monetary policy, it is relevant to consider the fiscal, financial and external conditions of the economy. An autistic monetary regime that ignores these issues, at the risk of becoming an additional source of uncertainty, would be useless.
In Sargent’s words (2007), there is no robust monetary regime for an inconsistent fiscal policy. The Argentine history is revealing particularly underlining the close link between the lack of fiscal solvency and inflation. Even though in recent years the restriction has eased due to budgetary surplus, the literature on fiscal dominance is not limited to the current period. In this context, it is well known that tax revenue structure depends on the current relative price structure, hence the need to address the issue under a general equilibrium approach. Similarly, any doubt about the financial system’s solvency or the external sustainability may restrain monetary policy room for maneuver.
Regarding external solvency, it would be inconsistent to tailor exchange rate policy to temporary trends: the pressure in favor of the appreciation of currencies was partly related to commodity prices (through the current account) and the investors’ risk appetite (through the capital account), which may clearly be temporary rather than permanent forces.
Rajan and Subramanian (2006) have highlighted the difference in treatment according to whether the trend is due to permanent or temporary factors. In a financial world characterized by capital flow volatility, a massive amount of capital may lead to some kind of “Dutch disease”, with unsustainable paths and high volatility of real variables.
These kind of problems may be more frequent in small economies, where the financial market is usually negligible as compared to international flows and where it is difficult to know the equilibrium level of real variables. Thus, temporary forces operating from the same side of the market may lead to nominal exchange rate overshooting (whether up or down). Far from adjusting towards the long-term equilibrium level, this may cause an excessive volatility and distort relative price signals for savings and consumption.
Ultimately, it is evident that the uncertainty faced by an economy such as Argentina's is not limited to the notion of risk. While in the latter case, the stochastic process to generate data is known, in the former the true model ruling macroeconomic function is unknown.
A frequency distribution bar chart of the Argentine GDP in the last 25 years would show that it is equally probable to grow or fall at a 10% rate; therefore, there are empirical problems to project the trend of fundamental variables. For instance, the 1998 national budget expected a 4.7% growth the following year and, in fact, in 1999 the decline exceeded 3%. This is an eloquent illustration of the problem in hand. Similar references could be mentioned with respect to the long-term equilibrium exchange rate or the consumption path.
Against an uncertain back, it is appropriate to prevent euphoria and adopt robust, gradual and coordinated policies to lead the economic convergence process towards its “cruising speed”. In this sense, making decisions but simultaneously keeping options to explore alternative paths is highly valuable. However, in a context of high nominal uncertainty, such as a widespread breach of contracts, there will be a demand on monetary authorities to provide clear and stable rules since the economic agents will relate these facts to an excessive past discretionality.
In this framework, the classical dilemma of rules versus discretion cannot be addressed with a “corner solution”. Extreme strategies, such as rigid and lock-in rules, would be useless and they would not even meet the initial objective of building fast credibility if agents do not perceive their consistency with the remaining policies.
On the other extreme, pure discretionary strategies cannot be pursued either. The demand for flexibility would be the perfect “excuse” for “not having a plan” and would validate the persistent trends in nominal variable imbalances.
Between both ends, the pillars of the domestic economy must be built progressively and patiently around a path that adequately combines the appropriate doses of flexibility and credibility.
In addition, this approach is not an isolated case in the world. Based on a study performed by Sturzenegger and Levy Yeyati (2005), only 50% of the countries that adopted the inflation-targeting model –where there is theoretically no intervention by the Central Bank in the exchange rate market– are effectively pursuing a free floating regime policy. Moreover, after the Lehman Brothers collapse almost all emerging markets with inflation targeting had to intervene in the foreign exchange market to avoid excessive volatility.
As regards the monetary policy, this resulted initially in the recovery of money basic functions, normalizing the financial system, and then in the rebuilding of the transmission channels of traditional policy tools so as to use them adequately when the economy is close to its steady state.
In general, five transmission channels can be identified: interest rate, credit, exchange rate, asset prices and expectations. Naturally enough, it is only meaningful to speak of monetary policy transmission mechanisms when there are nominal rigidities, both in goods and assets markets, which hinder an immediate accommodation of price levels. In a world characterized by flexible prices, all transmission mechanisms are alike and the question about the channels that might help the most to explain monetary policy transmission to the activity level becomes irrelevant.
In an economy in transition towards a steady state, not only the precise links among fundamental variables are unknown but also, and even more importantly, many policy tools that are typical of the steady state are unavailable.
A history of many decades of fixed exchange rate regimes – including a currency board that lasted more than ten years- followed by high devaluations and a hyperinflation processes make a difference between Argentina and neighboring countries in terms of absorbing significant swings in the exchange rate without threatening financial and monetary stability. The exchange rate cannot fully work as a “shock absorber”, as swings in the fx market boost uncertainty and substitution toward foreign currencies –channeling funds outside the financial system.
This peculiar economic history and idiosyncrasy not only affects the power of the instrument of the different branches of economic policy but also influences the way economic agents react to incentives and policies in a context where the precise connections among fundamental variables are unknown. Since the crisis, people markedly favor liquid forms of money to the detriment of wider monetary aggregates.
Therefore, in an economy like ours with precedents such as confiscation of deposits (1989, 2001), hyperinflation (1989, 1990), mega-devaluations (1989, 1990, 1991, 2002), and a default on the public debt (2001), the monetary system cannot set itself an exclusive goal and ignore the economy’s health and vulnerabilities. In these types of economies, where society has developed a high risk aversion, the need to prevent a crisis becomes a priority.
The past ¨haunts us¨ and leads us to adopt a prudent and asymmetrical strategy in response to exogenous forces: until enough evidence proves otherwise, we are forced to “dramatically shift the burden of proof” and assume that every negative shock is permanent and that, a priori, positive forces are temporary. The buildup of liquidity buffers in good times allows us to weather the storms in an unprecedented way for our history.
In the specific case of emerging countries that are very prone to financial and fiscal crises (such as Argentina), where dollarization levels are high, we need to design a monetary regime that is sufficiently robust to weather the challenges that we face in the transition phase towards a long term equilibrium.
And, to have a sizable amount of foreign reserves is of the essence. In the absence of an international lender of last resort, there is no substitute for a sound strong domestic liquidity policy. Many emerging market economies have been accumulating large stocks of foreign reserves, which allow them to be better prepared to face the global crisis. International reserves have reached a degree of coverage that was previously unseen in Argentina and they proved to be very useful to preserve monetary and financial stability by providing dollars to the market and to strengthen the demand for local currency in times of turbulence.
In particular, the four-pillar risk management strategy of the Central Bank during my tenure (the convergence between money supply and demand through yearly targets, a managed floating exchange rate regime, a counter-cyclical policy and liquidity networks in local and foreign currency, and an adequate banking regulation and supervision) allow to overcome every stress episode and minimize the impact on the real economy. It also prevents inconsistencies that could weaken its sustainability over time.
In recent times, the Argentine economy has undergone periods of remarkable tensions in the local market and the Central Bank’s policies responded as expected. The fact that four episodes in the last two years (July-October 2007, April-June 2008, September-November 2008, and March 2009) could be successfully addressed shows in a conclusive manner the soundness of the approach adopted by the Central Bank.
In each episode, the first action was to provide the means required to normalize the demand for money and stabilize the exchange rate market. Then, with simple tools, the injection of liquidity to guarantee the systemic stability was ensured.
Some measures were taken to neutralize the impact of a higher demand of foreign currency on money supply through liquidity injection, mainly redemption of Central Bank's Bonds and Notes (LEBAC and NOBAC) in the secondary market and partial renewals. This prevented an excessive pressure on interest rates and simultaneously ensured the fulfillment of monetary targets. Also, the Central Bank offered swaps between fixed and variable interest rate instruments so as to create a reference for the temporary structure of nominal interest rates in longer periods than those currently traded in the market. In addition, and with the objective of preventing an excessively cautious behavior by banks and fostering the use of their foreign currency excess resources to finance the business sector, a new option to access rediscounts in foreign currency was made available.
A non-flexible exchange rate and monetary regime was applied, on the basis of a managed float approach and a strict follow-up of the relevant monetary aggregates. This regime has allowed ensuring stability and at the same time build a healthy financial system (without currency mismatches or an excessive exposure to the public sector) that has progressively resumed its role as households’ savings intermediator (see Figure 4).
The managed floating exchange rate regime does not offer an “exchange rate guarantee” favoring short-term capital inflows and is intended to prevent high volatility from distorting consumption, saving and investment decisions. The objective is not to inhibit the convergence of variables to their long-term values but to avoid an excess volatility likely to cause unnecessary disturbances in economic decision making.
The orderly evolution of the exchange rate market in those episodes was enabled by the prudential reserves accumulation made by the Central Bank during the upswing of the international financial cycle (see Figure 5). By definition, any counter-cyclical policy must begin during a boom, and thinking that this favorable scenario could last for ever would have been a short-sighted reading of the recent Argentine history.
By respecting the special characteristics of the argentine economy that has suffered several financial crisis in the last decades, the monetary and financial policy options of the Central Bank of Argentina are framed within the context of a thorough review of the balance among the different policy objectives by Central Banks. Special consideration is given to the role and weight of price stability, output stability and financial stability.4. Conclusions
Monetary policy is managed under a high degree of uncertainty about the real economic structure and the way in which specific policy actions affect price evolution and the output level. This is particularly remarkable in emerging economies.
Against a back of high uncertainty, in countries witch are still in transition towards a steady state and trying to overcome decades of a sharp decline, the monetary authority must provide and ensure monetary and financial stability. In the case of Argentina, uncertainty is compounded by a long tradition of macroeconomic instability. Monetary regimes have unsuccessfully gone from one extreme to the other. The periods of relative stability have been short-lived, due to inconsistencies in the approaches that should have ensured macroeconomic solvency.
In these cases, the monetary regime cannot address an excluding objective and ignore the condition of the economy and its vulnerabilities. Therefore, in order to reach long-term, this historical evolution must be taken into account, particularly its effects on the behavioral pattern of economic agents, in addition to the uncertainty factors inherent to monetary policy management.
In this context, it is crucial to combine the use of models with the judgment of the experts to facilitate a thorough interpretation of economic phenomena. In turn, simple rules may provide valuable information to compare results and determine different policy options.
To this end, the robust strategy added to the theoretical consistency shown in this paper has proved to be the best approach for the current stage of the argentine economy. Its effectiveness was evident in the last quarters when a quick response was provided to preserve the monetary and financial stability in the face of a new international context. Undoubtedly, this scheme reinforces the need for a permanent and comprehensive work to reconcile tools and objectives. This is also the challenge faced to build institutions and simultaneously contribute to growth with stability and social inclusion which is, after all, the ultimate goal of every economic authority with a strategic vision.
Otherwise, all efforts will be in vain and there will be another fall in the classical Argentine pendulum, going endlessly from one extreme of likely policies to the other. In fact, there is a twofold challenge ahead, which is synchronic rather than sequential: making progress in the fulfillment of policymakers’ objectives and simultaneously building institutions for the benefit of future generations.
Blinder, A. and R. Reis (2005). “Understanding the Greenspan Standard”, at the Conference “The Greenspan Era: Lessons for the future”, Federal Reserve Bank of Kansas City, Jackson Hole, Wyoming, August, pp. 25-27.
Bernanke, B. S. (2004). “The Logic of Monetary Policy”. Remarks by Governor before the National Economists Club, Washington, D.C.
Greenspan, A. (2003). Opening Remarks at “Monetary Policy under Uncertainty”, symposium sponsored by the Federal Reserve Bank of Kansas City, Jackson Hole, Wyoming.
Greenspan, A. (2004). “Risk and Uncertainty in Monetary Policy”, remarks at the Meetings of the American Economic Association, San Diego, California, January 3.
Levin, A.; V. Wieland and J. Williams. (2003). “The Performance of Forecast-based Monetary Policy Rules under Model Uncertainty”, The American Economic Review, Vol. 93, pp. 622-645.
Levy Yeyati, E. and F. Sturzenegger (2005). “Classifying Exchange Rate Regimes: Deeds vs.Words”. European Economic Review, Vol. 49. August.
Rajan, R and A. Subramanian (2006). “Aid, Dutch Disease, and Manufacturing Growth”, Mimeo, IMF.
Sargent, T. (1993). “Bounded Rationality in Macroecnomics”. Oxford. Clarendon Press.
Sargent, T. (2007). “Evolution and Intelligent Design”, Eleventh Annual Conference, Santiago, Chile.
Walsh, C. (2003): “Implications of a Changing Economic Structure for the Strategy of Monetary Policy”, in Monetary Policy and Uncertainty: Adapting to a Changing Economy, Jackson Hole Symposium, Federal Reserve Bank of Kansas City, pp. 297-348.