- Email: [email protected]
Transparency and central bank losses in developing countries
Research in Economics 62 (2008) 45–54 www.elsevier.com/locate/rie
Transparency and central bank losses in developing countries Osama D. Sweidan a,∗ , Benjamin Widner b,1 a Department of Economics, P.O. Box 27272, University of Sharjah, Sharjah, United Arab Emirates b Department of Economics and International Business, New Mexico State University, P.O. Box 30001/MSC3CQ, Las Cruces,
NM 88003-8001, United States Received 2 April 2007; accepted 6 December 2007
Abstract Recent evidence shows central banks suffering from losses in some developing countries. This is a surprise to economists and policymakers. At the same time, these banks are asked to conduct monetary policy within a more transparent framework. Therefore, this paper seeks to answer the following question: Would more transparency in developing countries suffering from central bank losses be beneficial? This paper shows that the cost constraints of conducting monetary policy, central bank losses, in both transparency and opacity alike is significant and affects positively the error of the private sector in expected inflation rate and the output gap. In a country suffering from central bank losses, the expected benefits of transparency and the existence of cost constraint move in two opposite directions. As a result, it is unwise for developing countries suffering from central bank losses to focus on transparency. Priority should be given to fixing monetary policy and to developing financial markets. c 2007 University of Venice. Published by Elsevier Ltd. All rights reserved.
Keywords: Transparency; Central bank losses; Monetary policy
1. Introduction Central banking has experienced considerable change during the last decade. The recent paradigm in monetary policy is to have independent and transparent central banks. The argument among economists about transparency is a continuation to the debate that tries to answer the following question: Should monetary policy be made by rule or discretion? This dispute is related more to the institutional characteristics of the central banks and their effect on the monetary policy performance. Recent studies in this area have attempted to focus more on the optimal degree of transparency. The proponents of transparency agree that it is a self-evident necessity in order to attain sound monetary policy. As well, it affords better control of the public’s expectations by the central bank. Besides, it helps the actions adopted by the central banks to be more accountable. The technical effect of transparency to achieve price stability at lowest cost comes from the following process: the central bank should preannounce a low inflation target and commit to achieving it. In addition, the central bank should make the target reasonable for the public to let the announcement influence their expectations. ∗ Corresponding author. Tel.: +971 6 5053572; fax: +971 6 5585050.
E-mail addresses: [email protected] (O.D. Sweidan), [email protected] (B. Widner). 1 Tel.: +1 505 646 5989; fax: +1 505 646 1915. c 2007 University of Venice. Published by Elsevier Ltd. All rights reserved. 1090-9443/$ - see front matter doi:10.1016/j.rie.2007.12.002
46
O.D. Sweidan, B. Widner / Research in Economics 62 (2008) 45–54
Some economists went further to describe transparency as a part of democratic society. They assure that transparency should be a feature of the economy to guarantee successful economic policy. Stiglitz (1998) describes transparency as a central aspect of the democratic process, and a necessary condition for democratic governance to survive. Furthermore, he believes that the existence of transparency and democratic society reduces the concentration of power in the hands of single individuals, and is able to control human fallibility. Likewise, Mishkin (1999) demonstrates that transparency not only increases the ability of the central banks to conduct an effective monetary policy, but also helps to align central banks with democratic principles. Moreover, it promotes public support for the goals of monetary policy and reduces political pressures on the central banks. On the other hand, in a vital work Cukierman (2005) does not agree with full transparency in all cases. He assures that full transparency is not necessarily an optimal solution. Consequently, he distinguishes between areas in which high transparency is desired and areas in which an intermediate level of transparency is desired. He corroborates that it is highly risky and counterproductive to publish signals about severe problems in the financial system. This behaviour might lead to an unanticipated reaction from the public. Within the context of this debate, the central banks of developing countries are asked to present increased transparency. However, most of the theoretical analysis of transparency never pays no attention to the features of developing counties. Recent observations from developing countries submitted by economists such as Dalton and Dziobek (2005), Chandavarkar (1996) and Leone (1993) show that many central banks are suffering losses. It reads like a novel having losses in the income statement of a central bank. Thus, the current paper tries to address a specific question which is: would more transparency in developing countries be beneficial within this restriction? To answer this question, we use a known model from monetary economics. This model is used by Tarkka and Mayes (1999), and we upgrade part of it to shed light on the uniqueness of developing countries. Our results show that the cost constraints of conducting monetary policy, the central bank’s losses, in both transparency and opacity alike are significant and affect positively the error of the private sector’s expected inflation rate, and the output gap. In a country suffering from central bank losses, the benefits of transparency and the effect of cost constraint move in opposite directions. As a result, it is unwise for developing countries suffering from central bank losses to focus on transparency. Priority should be given to fixing monetary policy and to developing financial markets. The result of the current study raises an important implication and even a solution to why developing countries do not switch to using required reserve ratio as a monetary tool to guarantee a minimum cost of short-run operations of monetary policy. Despite the fact that this tool has criticisms2 it is still a good choice to help monetary authorities in these developing countries to catch a balance in their income statements. The rest of the paper is organized as follows: Section 2 presents the literature review about both transparency of monetary policy and central bank losses. Also, it describes the problem facing developing countries and its effect on conducting monetary policy. Section 3 introduces the model of the paper. Sections 4–6 bring in the calculations of both inflation error and output gap under transparency and opacity. Conclusions are drawn in the last section. 2. Literature review 2.1. Transparent monetary policy Let us confirm that we are talking about two important issues; transparency is a vital and necessary process which we try to apply, and central bank losses which is a matter of fact in some developing countries. Hence, the goal of the paper is to connect them by understanding the effect of the former on the economy within the existence of the latter. Also, we would like to confirm, based on our knowledge, that there are no previous studies making this connection. Therefore, we utilize a model used by other economists to study transparency. Transparency can be defined as the existence of symmetrical information between the central bank and the economic agents.3 Some economists consider that adopting a framework of inflation targeting and publication of inflation forecasts is an example of transparent monetary policy. This means that transparency reduces uncertainty. For the purpose of this study, we define transparency as an easy flow of information concerning economic variables and expectations between the central 2 For more details about required reserve ratio see Mishkin (2004). 3 For more details see Enoch (1998), Jensen (2000), Faust and Svensson (2001) and Gerrats (2002).
O.D. Sweidan, B. Widner / Research in Economics 62 (2008) 45–54
47
bank and the private sector. However, information associated with the central bank’s problems is not subject to any kind of release. Other economists such as Tarkka and Mayes (1999) show that transparency improves macroeconomic performance. Blinder et al. (2001) advocate the idea that nowadays central banks should work toward greater openness and transparency. This helps the public to understand monetary policy and respond more efficiently.4 Gerrats (2001) identifies several benefits of transparency i.e. it improves the central bank’s ability to build reputation and reduces inflation bias. Besides, she states that when a central bank is given a chance to select a regime, transparency need not be the outcome. Faust and Svensson (2001) conclude that more transparency increases credibility and improves policy performance, and makes the central bank’s optimal policy closer to the social optimum one. Chortareas et al. (2002) conduct an empirical study by using cross-sectional data of 87 countries over the period 1995–1999. Their results show a statistically significant negative relationship between transparency and inflation rate, in particular, in countries adopting flexible exchange rate. Also, their results do not support the proposition that there is a positive relationship between transparency and output volatility. Jensen (2000) studies the optimal degree of transparency in a simple model of monetary policymaking. He finds that the optimal degree of transparency is related to the trade-off between credibility and flexibility. This implies a high degree of transparency is desirable for a central bank with poor credibility. However, transparency could provide less flexibility to a central bank with high credibility. On the other hand, many economists such as Thornton (2002) state that transparency is neither a necessary nor sufficient condition for credibility. He argues that if policy effectiveness is enhanced by transparency, policymakers should be obliged to be more transparent, otherwise policymakers should be more secretive. He concludes that choosing transparency depends on both the economic environment and policy goals. Cukierman (2002) states that western central banks are opaque about their economic models and objective function. Hahn (2006) illustrates that transparency is preferred if it values the employment targets, whereas society prefers opacity if it estimates the inflation target. Cukierman (2005) probes the limits of transparency in the monetary policymaking area along with feasibility and desirability. On the feasibility side, he states that limited knowledge about the economy affects performance and expectations of monetary policy. On the desirability side, he distinguishes between areas in which high transparency is required from areas in which an intermediate level of transparency is wanted. The different points of view mentioned above deal generally with developed countries. In these countries, money and financial markets are sophisticated, well developed and well organized. Therefore, with these conditions some academics see it is wise to shorten the distance between the players by creating more confidence and transparency. 2.2. Central bank losses Discussing profits and losses of central banks in developed countries has not been a significant issue. This is probably due to the fact that central banks in these countries carry out strictly monetary functions and operate under stable macroeconomic conditions with basic financially sound institutions. However, in many developing countries, the central banks have usually conducted inappropriate monetary policy that has had a significant impact on their financial position. In recent years severe central bank losses have occurred in several developing countries. Different reasons have emerged to explain this unexpected result, of which the important ones are an absence of efficient and well-developed financial markets in developing countries.5 This outcome is a surprise to economists and policymakers because there is a limited analysis of this phenomenon in the literature. In addition, central bankers do not have enough provisions to deal with such an issue. Within the context of transparency an important question arises in developing countries; do these countries need to focus more on transparency or to fix monetary policy and to develop their financial market? A typical central bank should be able to issue noninterest- bearing debt (currency) and use the seignorage to make investments. Moreover, the central bank can invest in buying assets to earn interest, i.e. government securities and discount loans. Under normal circumstances, the revenue generated by a central bank should be able to finance its debt. However, from a technical point of view, central bank losses can arise in two ways: first, engaging in different activities such as open market operations, domestic and foreign investments, currency inflows and restructuring weak commercial banks. Second, valuation losses arise from revaluation of assets and liabilities. Central bank losses denote 4 For more details about the consequences of transparency see Gerrats (2002). 5 For more details see Leone (1993), Chandavarkar (1996) and Sweidan and Maghyereh (2006).
48
O.D. Sweidan, B. Widner / Research in Economics 62 (2008) 45–54
a negative net worth or the capital of the bank is eroding. In the current paper, we are interested in the operational losses that occur because of open market operations. Quintyn (1994) says the main advantage of issuing central bank securities to absorb excess reserves is the separation of monetary management from debt management. This guarantees an operational independent central bank. However, issuing central bank securities has a shortcoming in the high cost which leads to central bank losses.6 Axilord (1997) confirms that the cost of issuing central bank securities depends on the amount of excess reserves and the level of interest rate. Chandavarkar (1996), Dalton and Dziobek (2005) and Sweidan and Maghyereh (2006) provide evidence that central banks of several developing countries suffered from losses because of issuing central bank securities. Accordingly, the evidence shows that adopting a monetary policy that goes much beyond conventional central banking functions causes high cost or losses to the central bank. If this is the case, we might ask about the transparency of monetary policy when financial markets are not yet fully developed or have constraints. To cover these losses the International Monetary Fund (IMF) recommended the intervention of the government through recapitalizing the central bank by injection of either government securities or cash. The IMF is interested in facing this phenomenon and proposes legislative and accounting practices for the coverage of central bank losses.7 The current paper focuses mainly on the losses that arise from issuing central bank securities. In a developing country, if the financial market is underdeveloped and the central bank is unable to bring in genuine open market operations, then it will engage in a primary market through establishing its own securities. This technique implies that the securities should be on the liabilities side not on the assets side of the balance sheet. Therefore, the central bank bears high interest cost when adopting a tight monetary policy. On the other hand, more losses mean the central bank is biased to adopt an easy monetary policy. Central bank losses might harm the effectiveness of monetary policy. Most developing countries adopt a fixed exchange rate which requires accumulating reserves, and a high flexibility to adjust domestic interest rate in consistent with international interest rate. Operational losses in the central bank’s balance sheet prevent monetary policy from performing its task efficiently. All these pressures might spillover to the fiscal policy and generate stress on the government budget. What is more, in the case of capital inflows, indecisive monetary policy leads to a serious problem in the economy. Under the pressure of losses these banks are not capable to absorb the excess liquidity from the market, and the only solution is to encourage private firms to generate more underwritings which might cause a shortage in the quantity of the money supply. As a result, the economy and the financial market face a critical situation. Evidence from Jordan8 shows that in summer of 2006 Amman Stock Exchange (ASE) had a severe decline in its trading activity, and the reason was lack of liquidity because of the extensive underwritings. Accordingly, the contribution of the current paper comes from connecting both transparency and central bank losses to inflation error and output gap. In general, based on the above-mentioned literature review, we can state of the problem as follows. Central banks under the concept of transparency are required to disseminate data, facts and forecasts about their monetary policy. However, monetary policy in some developing countries has constraints i.e. central bank losses. In the current paper, we are interested in the losses that occur when a central bank issues its own securities to engage in a tight monetary policy. We are trying to build a scenario to explore a case in which a central bank publishes the information and expectations it has, whereas it is secretive about its costs or losses. Therefore, we try to answer the following question: Would more transparency in developing countries be beneficial within this restriction of conducting monetary policy? 3. The model To achieve the goals of the current study, we use a standard model of analysing monetary policy involving expectations. This model allows for uncertainty through the difference between central bank and private sector expectations. The model constructs a story about how the economy is evolving by using sequence of events. We choose such a model because it is a flexible form and helps us to model central bank costs or losses. Many economists such as Walsh (1998) and Tarkka and Mayes (1999) used this standard model. Our contribution to this model is that we added three additional equations to present the central bank’s losses. In addition, we adjust the sequence of events 6 For the purpose of this study high cost and losses are synonyms. 7 For more details about remedies see Dalton and Dziobek (2005), Stella (1997) and Leone (1993). 8 The Central Bank of Jordan (CBJ) which had losses in some years during the period (1996–2003).
O.D. Sweidan, B. Widner / Research in Economics 62 (2008) 45–54
49
Fig. 1. Sequence of events in repeatedly cycle.
to take these equations into consideration. The sequence of events creates one repeated cycle. The sequence of events and structure of information is as follows: first, at the beginning of period 0, the private sector forms its expected inflation rate (π0e ). At the end of period 0, the central bank observes these expectations, and at the end of this stage the central bank determines its exact inflation target (π¯ 0 ). Second, at the beginning of period 1, the central bank sets its monetary policy. The monetary authority bases its decision on how to set monetary policy on the past information it has, the new facts it receives and on its expectation of the future. In this period of time, we assume it is required by the central bank to respond by adopting a contractionary monetary policy. Then, the central bank realizes that it faces losses or high cost because of issuing its own securities. As a result, the central bank adjusts its money supply to reduce the losses. We assume the central bank gives the priority to heal its losses compared with inflation rate. The monetary authority decision’s maker can assess the right situation after the shock itself. The size of the losses depends on factors such as the level of the interest rate and the amount of the excess reserve the central bank intend to withdraw from the market. At this stage, the central bank observes the control error and as a result actual inflation rate. Third, in period 2 stochastic shocks of both demand and supply are realized. The central bank will take this new information set and influence its policy in the next cycle. Thus in period 2, the final effects on the output and inflation can be seen by the private sector and the central bank. Fig. 1, shows the decision making problem as a sequence of events. Let us assume that the aggregate supply of the economy is described by Lucas supply equation, as follows: Yt = Y¯ + a(πt − π0e ) + ε s t
(1)
where Yt is the real aggregate supply, Y¯ denotes the natural rate of output, πt stands for the inflation rate, π0e is the level of expected inflation rate by the public in period zero. ε s t is an aggregate supply shock. The aggregate demand of the economy is described by the following equation: Yt = Mt − πt + ε d t
(2)
Yt denotes the real aggregate demand, Mt is the money supply, πt stands for inflation rate and ε d t denotes the aggregate demand disturbance other than shifts in money supply. Following Tarkka and Mayes (1999) we assume that the aggregate demand disturbance has a mean (η) and a random error (φ) as follows: εdt = η + φ.
(3)
According to the loss function, it is a well-known consensus that the main goal of monetary policy should be price stability.9 The central bank decision maker seeks to minimize the difference between the actual inflation rate (πt ) and its inflation target (π0∗ ). The central bank sets its monetary policy by minimizing the following loss function: L C B = E C B,1 (πt − π0∗ )2 .
(4)
We will assume that the central bank has imperfect control over the inflation rate. The monetary authority will announce this fact, but they will not justify this imperfection. This implies that the central bank knows that the target cannot be achieved exactly. Also, we assume that the central bank responds to the losses in the sense that it appears in the determinants of the target variable.10 For this reason the central bank’s inflation target is as follows: π0∗ = π¯ 0 + µt 9 For more details see Cukierman (2002). 10 For more details about the difference between “target variable” and “responding to a variable” see Razin (2004).
(5)
50
O.D. Sweidan, B. Widner / Research in Economics 62 (2008) 45–54
where π¯ 0 is the central bank’s exact target of inflation rate which is announced at the end of period 0. π0∗ is observed in period 2. However, π¯ 0 is observed in period 0. µt denotes a monetary policy problem i.e. the expected losses facing central banks. We assume that the error (µt ) is not observed by the public. Therefore, we can rewrite Eq. (5) as π0∗ = π¯ 0 + λCt
(6)
where Ct stands for the cost of adopting a contractionary monetary policy. λ is the responsiveness of the central bank’s inflation target to a change in the central bank’s cost of adopting a tight monetary policy. Eq. (6) illustrates that the central bank’s inflation target varies with the cost constraint of the adopted monetary policy. Eq. (6) justifies and supports a conclusion by Masson et al. (1998) which says that in most developing countries the preconditions for adopting inflation targeting are not yet present. Based on the description presented above, we can write the relationship between the money supply and the cost to the central bank as follows: Ct = −Π Mt + ε ct
(7)
Mt denotes the money supply i.e. open market operations procedures. Π stands for the responsiveness of the central bank’s cost to a change in the money supply. εct is a cost shock i.e. quasi-fiscal activities.11 Eq. (7) shows a negative relationship running from the money supply to the cost of the policy. To complete the information about the model, we assume the expectations of the economic units in the model to different shocks as follows: 1- Expectations in period 0 E P S,0 (εdt ) = 0
E P S,0 (εst ) = 0
E P S,0 (φ) = 0.
2- Expectation at the beginning of period 1 E C B,1 (Y¯ ) = Y¯
E C B,1 (εdt ) = η
E C B,1 (φ) = 0
E C B,1 (εst ) = 0
E C B,1 (εct ) = 0
where PS,0 denotes private sector in period 0, and CB,1 stands for central bank in period 1. 4. Solving the model In this model, transparency is connected to the ability of both central bank and private sector to predict the expected inflation of the other part. However, the central bank’s losses are strategic and private information, and no one is going to release them at the present time. We do agree with Cukierman (2005) who considered that it is highly risky and counterproductive to disseminate information about a serious problem in the financial system. Distributing such a fact may cause runs on the banks or any other unexpected movement that threatens the stability of the financial system. As a result, the concept of transparency might not be available in some economic systems or environments. Moreover, Leone (1993) states that the income statement of the central banks in developing countries shows considerable profits despite the fact that the monetary authorities have significant problems. Also, he confirms that the method of registering the income and expenses of the central banks in these countries guarantees fictional profits. We do believe that the central bank’s losses in developing countries started to be an obvious global phenomenon after adopting the economic adjustment programmes with the IMF. The evaluation in the current model is in each single cycle not over the cycles. Based on the above-mentioned model and its sequence of events, we can derive the central bank’s expected inflation rate for period 2 at period 1 by minimizing the loss function, Eq. (4), subject to Eq. (6): E C B,1 (π2 ) = π¯ 0 + λE C B,1 (C E,1 )
(8)
where E,1: denotes expectation at end of period 1. Eq. (8) shows that the central bank’s expected inflation of period 2 in period 1 depends on two factors the inflation target set in period 0, (π¯ 0 ), and the expected cost or losses in period 1, {E C B,1 (C E,1 )}. Eq. (8) says that the cost of monetary policy has a positive effect on the central bank’s actual inflation target. Further, this equation is an indicator of the effect of the operational procedures of the central bank
11 For more details about those activities see Leone (1993).
O.D. Sweidan, B. Widner / Research in Economics 62 (2008) 45–54
51
on its expected actual inflation rate. At this point of time, the private sector believes that the central bank is working toward achieving its expectations in period 0, which is E P S,1 = π¯ 0 .
(9)
Eqs. (8) and (9) show the first effect of the central bank’s losses by creating a gap between the expectations of both the central bank and the private sector in period 1. By using Eqs. (1) and (2) we can solve the model to find out the equilibrium of the actual inflation rate in period 2, which is as follows: a 1 πe + (M1 − Y¯ + εdt − εst ). (10) 1+a 0 1+a The central bank sets the suitable monetary policy to affect the aggregate demand and to achieve its goals. When the central bank sets its monetary policy in period 1 it uses all the available information from the following sources: period 0, the new information and the expectations of the future. Hence, taking expectations of Eq. (10) conditional on the central bank’s available information, and using Eq. (8) gives: π2 =
E C B,1 (M1 ) = (1 + a)E C B,1 (π2 ) − a E C B,1 (π0e ) + E C B,1 (Y¯ ) + E C B,1 (εst ) − E C B,1 (εdt ) M1 =
(1 + a)π¯ 0 + (1 + a)λE C B,1 (C E,1 ) − a E C B,1 (π0e ) + Y¯
− η.
(11) (12)
Eq. (12) shows that the central bank sets its monetary policy based on five factors. The cost of adopting a contractionary monetary policy {E C B,1 (C E,1 )} is one of these factors and plays an important role in determining the money supply. To understand the behaviour of the central bank in Eq. (12), it is important to use Eq. (7) and the sequence of events. Assume the central bank in period 1 expects the private sector’s expected inflation rate will increase; therefore the money supply should decrease as it is obvious from Eq. (12). Consequently, Eq. (7) states that the decrease in money supply increases the cost of central bank. For this reason, the central bank adjusts to increase the money supply to prevent higher cost and more losses. From the other side, presume the central bank observed a positive aggregate demand shock (η) in period 2. Hence, the central bank, in period 1, intervenes by reducing the money supply. Consequently, Eq. (7) demonstrates again that this action boosts up the cost of the monetary authority. As a result, the money supply is adjusted upward. This mechanism helps to understand the effect of the central bank’s losses on monetary policy in the long run or over the repeatedly cycles by substituting Eq. (7) in Eq. (12), we get the following equation: M1 =
a 1 (1 + a) π¯ 0 − E C B,1 (π0e ) + (Y¯ − η). (1 + λΠ + aλΠ ) (1 + λΠ + aλΠ ) (1 + λΠ + aλΠ )
(13)
Eq. (13) confirms that the existence of the central bank’s losses in the long run reduces the value of coefficients affecting the money supply. The central bank will be more cautious when it reacts to any shock in the economy. With a monetary policy given by Eq. (12) and with an equilibrium inflation rate given by Eq. (10), actual inflation rate in period 2 is calculated as follows: 1 a {E C B,1 (π0e ) − π0e } + (φ − εs2 ). (14) 1+a 1+a Eq. (14) presents the sources of actual inflation in such an economy, which they are; the central bank’s exact inflation target (π¯ 0 ), the cost of adopting a contractionary monetary policy {E C B,1 (C E,1 )}, the error made by the central bank in period 1 in predicting the private sector’s expected inflation rate in period 0 {E C B,1 (π0e ) − π0e }, and errors of anticipating aggregate demand and aggregate supply disturbances (φ − εs2 ). Central bank loss is a constraint on monetary policy action to target the private sector expected inflation rate. That’s why the existence of central bank losses in each single cycle forces the inflation rate to increase. This result is consistent with the previous finding which is that losses increase the money supply in each single cycle. π2 = π¯ 0 + λE C B,1 (C E,1 ) −
5. Opacity The opacity concept in this model means that the central bank does not release its own information and forecasts regarding the demand shocks and any problem facing the monetary policy i.e. central bank losses. Thus, the private
52
O.D. Sweidan, B. Widner / Research in Economics 62 (2008) 45–54
sector does not have access to information and is unable to make assessments and projections in period 0 for period 1.12 If the private sector has a full sight in period 0, the expected inflation of the private sector is π0e = E P S,0 (π2 ) = E P S,0 (π¯ 0 ) + λE P S,0 {E C B,1 (C E,1 )} −
a 1 E P S,0 {E C B,1 (π0e ) − π0e } + E P S,0 (φ − εs2 ). 1+a 1+a
(15)
Based on the above definition of opacity, the two terms E P S,0 {E C B,1 (C E,1 )} and E P S,0 {E C B,1 (π0e ) − π0e } are equal to zero. Accordingly, we can rewrite Eq. (15) as π0e = E P S,0 (π¯ 0 ).
(16)
Eq. (16) illustrates that in the case of opacity the private sector is just able to see the exact target of the central bank which is announced in period 0. The error of the private sector expected inflation rate can be expressed by the difference between the actual inflation rate in period 2 and expected inflation rate in period 0. We can express this error by the following formula: Error P S,2 = π2 − π0e π2 − π0e = {π¯ 0 − E P S,0 (π¯ 0 )} + λE C B,1 (C E,1 ) −
(17) a 1 {E C B,1 (π0e ) − π0e } + (φ − εst ). 1+a 1+a
(18)
Eq. (18) shows that the absence of transparency guarantees the existence of error made by the private sector in predicting inflation rate. This error depends on the following sources: (1) (2) (3) (4) (5)
The failure of private sector to predict the central bank’s exact target of inflation rate. The cost of adopting a tight monetary policy or the cost constraint. The failure of the central bank to forecast the private sector expected inflation rate. The aggregate supply shock. The random error of the aggregate demand shock.
Eq. (18) is similar to Eq. (3.15’) derived by Tarkka and Mayes (1999), the new thing we have in our formula is the effect of the cost of adopting a contractionary monetary policy. Moreover, we can calculate the fluctuation in output under opacity by substituting Eq. (18) in Eq. (1) as follows: a2 1 a Y2 − Y¯ = a[{π¯ 0 − E P S,0 (π¯ 0 )} + λE C B,1 (C E,1 )] − {E C B,1 (π0e ) − π0e } + φ+ εs 2 . 1+a 1+a 1+a
(19)
Eq. (19) confirms the role of the cost constraint of adopting a contractionary monetary policy on the departure of actual from potential output. 6. Transparency As defined above, transparency in the current paper means that the central bank releases all the information and predictions it has regarding the macroeconomic variables. However, the central bank will not release any information regarding the central bank losses. Contrary to the opacity case, the private sector is able to use the information and expectations starting from period 1. Thus, the expected inflation rate made by the private sector looks like Eq. (15). The error made by the private sector in predicting inflation rate is expressed by π2 − π0e = {π¯ 0 − E P S,0 (π¯ 0 )} + λ{E C B,1 (C E,1 ) − E P S,0 {E C B,1 (C E,1 )}} −
a 1 {{E C B,1 (π0e ) − π0e } − E P S,0 {E C B,1 (π0e ) − π0e }} + (φ − εs2 ). 1+a 1+a
(20)
Based on the definition of transparency, the term E P S,0 {E C B,1 (C E,1 )} is equal to zero. The private sector is unable to perceive the problem of losses facing the central bank. Moreover, transparency enhances the forecasting power of 12 If the private sector is unable to make projection from period 0 this implies that π e = 0. 0
O.D. Sweidan, B. Widner / Research in Economics 62 (2008) 45–54
53
the private sector. Thus, the private sector is more professional in anticipating the central bank’s projected action. The private sector can predict the error made by the central bank in predicting the private sector expected inflation rate. That is why the term {{E C B,1 (π0e ) − π0e } − E P S,0 {E C B,1 (π0e ) − π0e }} is equal to zero. The formula [Eq. (20)] of the error made by the private sector in predicting inflation rate is reduced to 1 (φ − ε s 2 ) 1+a a 1 Y2 − Y¯ = a[{π¯ 0 − E P S,0 (π¯ 0 )} + λE C B,1 (C E,1 )] + φ+ εs 2 . 1+a 1+a
π2 − π0e = {π¯ 0 − E P S,0 (π¯ 0 )} + λE C B,1 (C E,1 ) +
(21) (22)
Eq. (21) shows that the departure of the private sector expected inflation rate from the actual inflation rate is less in transparency case compared to opacity case in Eq. (18). This result is confirmed by Tarkka and Mayes (1999).13 Further, the current paper brings four main conclusions: First, the cost constraint of adopting a tight monetary policy in both transparency and opacity alike is significant and affects positively the error of the private sector expected inflation rate [Eqs. (18) and (21)], and the output gap [Eqs. (19) and (22)]. Technically, the results of our paper show that the cost constraint of monetary policy influences the actual inflation rate, but does not affect the private sector expected inflation rate. Second, it is obvious from the above results that the benefits of transparency and the effect of cost constraint on the conduct of monetary policy move in two opposite directions; transparency reduces fluctuation in both inflation error and output gap, while the cost constraint leads to more fluctuation. Three, based on the information they have, the private sector understands that the policy of the central bank is enhanced by transparency. However, the private sector at the end of each single cycle notices the gap between actual and predicted macroeconomic variables i.e. output and inflation rate, therefore the private sector will be surprised. Consequently, the private sector is unable to understand the reason for the gap. This contradiction between what is expected and what is seen in reality might put a big question on the accuracy, credibility and accountability of the transparency process. Four, the results of the current paper point to the possibility that inflation rate acceleration might be a serious problem in developing countries especially those with constraints on conducting their monetary policies. 7. Conclusions We believe that developing countries are not subject to all theories applied to developed countries. The privacy of the problems in developing countries plays a crucial role in this mismatch. The current paper focuses on the transparency of central banks in developing countries. This issue looks important because recent evidence shows that the central banks in some developing countries suffer from losses. We think this issue is unique, important and never expected to happen. We try to answer the following question: Would more transparency in developing countries suffering from central bank losses be beneficial? To achieve our goal, we utilize a model used by other economists to study transparency. However, we upgrade part of the model to shed light on the uniqueness of developing countries. This paper illustrates that the cost of adopting a contractionary monetary policy has a positive relationship with the central bank’s inflation target and actual inflation rate. In addition, in the long run, monetary policy will be less sensitive to any shock in the economy. These results create a gap between the actual macroeconomic variables i.e. inflation rate and output and its expected value by the private sector. If the private sector believes it is living within a transparent economic system, many questions arise concerning the accountability of monetary forecasts. Therefore, we believe that the existence of both transparency and cost constraint on the conduct of monetary policy will not stay forever. The constraint works to terminate transparency. The outcome of the current paper proves that the effect on inflation error and output gap moves in two opposite directions. Therefore, transparency in developing countries with central bank losses could be beneficial for a very short period of time, and in the long run we expect that transparency vanishes. It is a necessary condition to have a suitable environment for transparency, and it is unwise for developing countries suffering from central bank losses to focus on transparency; priority should be given to fixing monetary policy and to develop the financial markets. We would like to confirm that repairing monetary policy in developing countries should track two dimensions: first, central banks suffering operational losses because of adopting open market operations should switch to monetary 13 Comparing the difference between transparency and opacity is not a target of the current study.
54
O.D. Sweidan, B. Widner / Research in Economics 62 (2008) 45–54
policy that lowers the cost as required by the reserve ratio. Second, we reconfirm the procedures proposed by the IMF which enforce governments to recapitalize central banks facing losses, and commit governments to adopt this action in advance. Acknowledgements We would like to thank the editor and two anonymous referees of Research in Economics Journal for valuable and helpful comments. We are responsible for any remaining errors. References Axilord, S.H., 1997. Transformations to Open Market Operations. International Monetary Fund, January. Washington. Blinder, A., Goodhart, C., Hildebrand, P., Lipton, D., Wyplosz, C., 2001. How Do Central Banks Talk? International Center for Monetary and Banking Studies. Geneva, Switzerland. Chandavarkar, A., 1996. Central Banking in Developing Countries. St. Martin’s Press Inc. Chortareas, G., Stasavage, D., Sterne, G., 2002. Does it Pay To Be Transparent? International Evidence from Central Bank Forecasts. Review, Federal Reserve Bank of St. Louis. July/August. pp. 99–118. Cukierman, A., 2002. Are Contemporary Central Banks Transparent About Economic Models and Objectives and What Difference Does it Make? Review, Federal Reserve Bank of Saint Louis. July/August. pp. 15–35. Cukierman, A., 2005. The Limits of Transparency. Unpublished. Dalton, J., Dziobek, C., 2005. Central Banks Losses and Experiences in Selected Countries. IMF Working Paper No. WP/05/72, April. Washington. Enoch, C., 1998. Transparency in Central Bank Operations in the Foreign Exchange Market. IMF Paper on Policy Analysis and Assessment, PPAA/98/2, March. Washington. Faust, J., Svensson, L.E., 2001. Transparency and Credibility: Monetary Policy with Unobservable Goals. International Economics Review 42 (2), 369–397. Gerrats, P.M., 2001. Why Adopt Transparency? The Publication of Central Bank Forecasts. Working Paper No. 41, European Central Bank. Gerrats, P.M., 2002. Central Bank Transparency. The Economic Journal 112 (November), F532–F565. Hahn, V., 2006. The Transparency of Central Bank Preferences, June, Available at: SSRN: http://ssrn.com/abstract=592261. Jensen, H., 2000. Optimal Degree of Transparency in Monetary Policymaking. Working Paper, University of Copenhagen. Leone, A., 1993. Institutional and Operational Aspects of Central Bank losses. IMF Paper on Policy Analysis and Assessment, PPAA/ 93/ 14, September. Washington. Masson, P.R., Savastano, M.A., Sharma, S., 1998. Can Inflation Targeting Be a Framework for Monetary Policy in Developing Countries? Finance and Development, March. pp. 34–37. Mishkin, F.S., 2004. The Economics of Money, Banking, and Financial Markets, seventh edition. Addison-Wesley. Mishkin, F.S., 1999. Central Banking in a Democratic Society: Implications for Transition Countries. Zagreb Journal of Economics 3 (3), 51–74. Quintyn, M., 1994. Government Securities versus Central Bank Securities in Developing Open Market Operations. IMF Working Paper No. 94/62, May. Washington. Razin, A., 2004. Aggregate Supply and Potential Output. Working Paper 10294, National Bureau of Economic Research. Stella, P., 1997. Do Central Banks Need Capital. IMF Working Paper No. 97/83, July. Washington. Stiglitz, J., 1998. Central Banking in a Democratic Society. De Economists 146 (2), 199–226. Sweidan, O., Maghyereh, A., 2006. Monetary Policy and the Central Bank’s Securities. Applied Economics Letters 13 (9), 593–598. Tarkka, J., Mayes, D.G., 1999. The Value of Publishing Official Central Bank Forecasts. Discussion Paper 22/99, Bank of Finland. Thornton, D.L., 2002. Monetary Policy Transparency: Transparent About What? Federal Reserve Bank of St. Louis, Working Paper 2002-028B, November. Walsh, C.E., 1998. Monetary Theory and Policy. MIT Press, London.
Transparency and central bank losses in developing countries Osama D. Sweidan a,∗ , Benjamin Widner b,1 a Department of Economics, P.O. Box 27272, University of Sharjah, Sharjah, United Arab Emirates b Department of Economics and International Business, New Mexico State University, P.O. Box 30001/MSC3CQ, Las Cruces,
NM 88003-8001, United States Received 2 April 2007; accepted 6 December 2007
Abstract Recent evidence shows central banks suffering from losses in some developing countries. This is a surprise to economists and policymakers. At the same time, these banks are asked to conduct monetary policy within a more transparent framework. Therefore, this paper seeks to answer the following question: Would more transparency in developing countries suffering from central bank losses be beneficial? This paper shows that the cost constraints of conducting monetary policy, central bank losses, in both transparency and opacity alike is significant and affects positively the error of the private sector in expected inflation rate and the output gap. In a country suffering from central bank losses, the expected benefits of transparency and the existence of cost constraint move in two opposite directions. As a result, it is unwise for developing countries suffering from central bank losses to focus on transparency. Priority should be given to fixing monetary policy and to developing financial markets. c 2007 University of Venice. Published by Elsevier Ltd. All rights reserved.
Keywords: Transparency; Central bank losses; Monetary policy
1. Introduction Central banking has experienced considerable change during the last decade. The recent paradigm in monetary policy is to have independent and transparent central banks. The argument among economists about transparency is a continuation to the debate that tries to answer the following question: Should monetary policy be made by rule or discretion? This dispute is related more to the institutional characteristics of the central banks and their effect on the monetary policy performance. Recent studies in this area have attempted to focus more on the optimal degree of transparency. The proponents of transparency agree that it is a self-evident necessity in order to attain sound monetary policy. As well, it affords better control of the public’s expectations by the central bank. Besides, it helps the actions adopted by the central banks to be more accountable. The technical effect of transparency to achieve price stability at lowest cost comes from the following process: the central bank should preannounce a low inflation target and commit to achieving it. In addition, the central bank should make the target reasonable for the public to let the announcement influence their expectations. ∗ Corresponding author. Tel.: +971 6 5053572; fax: +971 6 5585050.
E-mail addresses: [email protected] (O.D. Sweidan), [email protected] (B. Widner). 1 Tel.: +1 505 646 5989; fax: +1 505 646 1915. c 2007 University of Venice. Published by Elsevier Ltd. All rights reserved. 1090-9443/$ - see front matter doi:10.1016/j.rie.2007.12.002
46
O.D. Sweidan, B. Widner / Research in Economics 62 (2008) 45–54
Some economists went further to describe transparency as a part of democratic society. They assure that transparency should be a feature of the economy to guarantee successful economic policy. Stiglitz (1998) describes transparency as a central aspect of the democratic process, and a necessary condition for democratic governance to survive. Furthermore, he believes that the existence of transparency and democratic society reduces the concentration of power in the hands of single individuals, and is able to control human fallibility. Likewise, Mishkin (1999) demonstrates that transparency not only increases the ability of the central banks to conduct an effective monetary policy, but also helps to align central banks with democratic principles. Moreover, it promotes public support for the goals of monetary policy and reduces political pressures on the central banks. On the other hand, in a vital work Cukierman (2005) does not agree with full transparency in all cases. He assures that full transparency is not necessarily an optimal solution. Consequently, he distinguishes between areas in which high transparency is desired and areas in which an intermediate level of transparency is desired. He corroborates that it is highly risky and counterproductive to publish signals about severe problems in the financial system. This behaviour might lead to an unanticipated reaction from the public. Within the context of this debate, the central banks of developing countries are asked to present increased transparency. However, most of the theoretical analysis of transparency never pays no attention to the features of developing counties. Recent observations from developing countries submitted by economists such as Dalton and Dziobek (2005), Chandavarkar (1996) and Leone (1993) show that many central banks are suffering losses. It reads like a novel having losses in the income statement of a central bank. Thus, the current paper tries to address a specific question which is: would more transparency in developing countries be beneficial within this restriction? To answer this question, we use a known model from monetary economics. This model is used by Tarkka and Mayes (1999), and we upgrade part of it to shed light on the uniqueness of developing countries. Our results show that the cost constraints of conducting monetary policy, the central bank’s losses, in both transparency and opacity alike are significant and affect positively the error of the private sector’s expected inflation rate, and the output gap. In a country suffering from central bank losses, the benefits of transparency and the effect of cost constraint move in opposite directions. As a result, it is unwise for developing countries suffering from central bank losses to focus on transparency. Priority should be given to fixing monetary policy and to developing financial markets. The result of the current study raises an important implication and even a solution to why developing countries do not switch to using required reserve ratio as a monetary tool to guarantee a minimum cost of short-run operations of monetary policy. Despite the fact that this tool has criticisms2 it is still a good choice to help monetary authorities in these developing countries to catch a balance in their income statements. The rest of the paper is organized as follows: Section 2 presents the literature review about both transparency of monetary policy and central bank losses. Also, it describes the problem facing developing countries and its effect on conducting monetary policy. Section 3 introduces the model of the paper. Sections 4–6 bring in the calculations of both inflation error and output gap under transparency and opacity. Conclusions are drawn in the last section. 2. Literature review 2.1. Transparent monetary policy Let us confirm that we are talking about two important issues; transparency is a vital and necessary process which we try to apply, and central bank losses which is a matter of fact in some developing countries. Hence, the goal of the paper is to connect them by understanding the effect of the former on the economy within the existence of the latter. Also, we would like to confirm, based on our knowledge, that there are no previous studies making this connection. Therefore, we utilize a model used by other economists to study transparency. Transparency can be defined as the existence of symmetrical information between the central bank and the economic agents.3 Some economists consider that adopting a framework of inflation targeting and publication of inflation forecasts is an example of transparent monetary policy. This means that transparency reduces uncertainty. For the purpose of this study, we define transparency as an easy flow of information concerning economic variables and expectations between the central 2 For more details about required reserve ratio see Mishkin (2004). 3 For more details see Enoch (1998), Jensen (2000), Faust and Svensson (2001) and Gerrats (2002).
O.D. Sweidan, B. Widner / Research in Economics 62 (2008) 45–54
47
bank and the private sector. However, information associated with the central bank’s problems is not subject to any kind of release. Other economists such as Tarkka and Mayes (1999) show that transparency improves macroeconomic performance. Blinder et al. (2001) advocate the idea that nowadays central banks should work toward greater openness and transparency. This helps the public to understand monetary policy and respond more efficiently.4 Gerrats (2001) identifies several benefits of transparency i.e. it improves the central bank’s ability to build reputation and reduces inflation bias. Besides, she states that when a central bank is given a chance to select a regime, transparency need not be the outcome. Faust and Svensson (2001) conclude that more transparency increases credibility and improves policy performance, and makes the central bank’s optimal policy closer to the social optimum one. Chortareas et al. (2002) conduct an empirical study by using cross-sectional data of 87 countries over the period 1995–1999. Their results show a statistically significant negative relationship between transparency and inflation rate, in particular, in countries adopting flexible exchange rate. Also, their results do not support the proposition that there is a positive relationship between transparency and output volatility. Jensen (2000) studies the optimal degree of transparency in a simple model of monetary policymaking. He finds that the optimal degree of transparency is related to the trade-off between credibility and flexibility. This implies a high degree of transparency is desirable for a central bank with poor credibility. However, transparency could provide less flexibility to a central bank with high credibility. On the other hand, many economists such as Thornton (2002) state that transparency is neither a necessary nor sufficient condition for credibility. He argues that if policy effectiveness is enhanced by transparency, policymakers should be obliged to be more transparent, otherwise policymakers should be more secretive. He concludes that choosing transparency depends on both the economic environment and policy goals. Cukierman (2002) states that western central banks are opaque about their economic models and objective function. Hahn (2006) illustrates that transparency is preferred if it values the employment targets, whereas society prefers opacity if it estimates the inflation target. Cukierman (2005) probes the limits of transparency in the monetary policymaking area along with feasibility and desirability. On the feasibility side, he states that limited knowledge about the economy affects performance and expectations of monetary policy. On the desirability side, he distinguishes between areas in which high transparency is required from areas in which an intermediate level of transparency is wanted. The different points of view mentioned above deal generally with developed countries. In these countries, money and financial markets are sophisticated, well developed and well organized. Therefore, with these conditions some academics see it is wise to shorten the distance between the players by creating more confidence and transparency. 2.2. Central bank losses Discussing profits and losses of central banks in developed countries has not been a significant issue. This is probably due to the fact that central banks in these countries carry out strictly monetary functions and operate under stable macroeconomic conditions with basic financially sound institutions. However, in many developing countries, the central banks have usually conducted inappropriate monetary policy that has had a significant impact on their financial position. In recent years severe central bank losses have occurred in several developing countries. Different reasons have emerged to explain this unexpected result, of which the important ones are an absence of efficient and well-developed financial markets in developing countries.5 This outcome is a surprise to economists and policymakers because there is a limited analysis of this phenomenon in the literature. In addition, central bankers do not have enough provisions to deal with such an issue. Within the context of transparency an important question arises in developing countries; do these countries need to focus more on transparency or to fix monetary policy and to develop their financial market? A typical central bank should be able to issue noninterest- bearing debt (currency) and use the seignorage to make investments. Moreover, the central bank can invest in buying assets to earn interest, i.e. government securities and discount loans. Under normal circumstances, the revenue generated by a central bank should be able to finance its debt. However, from a technical point of view, central bank losses can arise in two ways: first, engaging in different activities such as open market operations, domestic and foreign investments, currency inflows and restructuring weak commercial banks. Second, valuation losses arise from revaluation of assets and liabilities. Central bank losses denote 4 For more details about the consequences of transparency see Gerrats (2002). 5 For more details see Leone (1993), Chandavarkar (1996) and Sweidan and Maghyereh (2006).
48
O.D. Sweidan, B. Widner / Research in Economics 62 (2008) 45–54
a negative net worth or the capital of the bank is eroding. In the current paper, we are interested in the operational losses that occur because of open market operations. Quintyn (1994) says the main advantage of issuing central bank securities to absorb excess reserves is the separation of monetary management from debt management. This guarantees an operational independent central bank. However, issuing central bank securities has a shortcoming in the high cost which leads to central bank losses.6 Axilord (1997) confirms that the cost of issuing central bank securities depends on the amount of excess reserves and the level of interest rate. Chandavarkar (1996), Dalton and Dziobek (2005) and Sweidan and Maghyereh (2006) provide evidence that central banks of several developing countries suffered from losses because of issuing central bank securities. Accordingly, the evidence shows that adopting a monetary policy that goes much beyond conventional central banking functions causes high cost or losses to the central bank. If this is the case, we might ask about the transparency of monetary policy when financial markets are not yet fully developed or have constraints. To cover these losses the International Monetary Fund (IMF) recommended the intervention of the government through recapitalizing the central bank by injection of either government securities or cash. The IMF is interested in facing this phenomenon and proposes legislative and accounting practices for the coverage of central bank losses.7 The current paper focuses mainly on the losses that arise from issuing central bank securities. In a developing country, if the financial market is underdeveloped and the central bank is unable to bring in genuine open market operations, then it will engage in a primary market through establishing its own securities. This technique implies that the securities should be on the liabilities side not on the assets side of the balance sheet. Therefore, the central bank bears high interest cost when adopting a tight monetary policy. On the other hand, more losses mean the central bank is biased to adopt an easy monetary policy. Central bank losses might harm the effectiveness of monetary policy. Most developing countries adopt a fixed exchange rate which requires accumulating reserves, and a high flexibility to adjust domestic interest rate in consistent with international interest rate. Operational losses in the central bank’s balance sheet prevent monetary policy from performing its task efficiently. All these pressures might spillover to the fiscal policy and generate stress on the government budget. What is more, in the case of capital inflows, indecisive monetary policy leads to a serious problem in the economy. Under the pressure of losses these banks are not capable to absorb the excess liquidity from the market, and the only solution is to encourage private firms to generate more underwritings which might cause a shortage in the quantity of the money supply. As a result, the economy and the financial market face a critical situation. Evidence from Jordan8 shows that in summer of 2006 Amman Stock Exchange (ASE) had a severe decline in its trading activity, and the reason was lack of liquidity because of the extensive underwritings. Accordingly, the contribution of the current paper comes from connecting both transparency and central bank losses to inflation error and output gap. In general, based on the above-mentioned literature review, we can state of the problem as follows. Central banks under the concept of transparency are required to disseminate data, facts and forecasts about their monetary policy. However, monetary policy in some developing countries has constraints i.e. central bank losses. In the current paper, we are interested in the losses that occur when a central bank issues its own securities to engage in a tight monetary policy. We are trying to build a scenario to explore a case in which a central bank publishes the information and expectations it has, whereas it is secretive about its costs or losses. Therefore, we try to answer the following question: Would more transparency in developing countries be beneficial within this restriction of conducting monetary policy? 3. The model To achieve the goals of the current study, we use a standard model of analysing monetary policy involving expectations. This model allows for uncertainty through the difference between central bank and private sector expectations. The model constructs a story about how the economy is evolving by using sequence of events. We choose such a model because it is a flexible form and helps us to model central bank costs or losses. Many economists such as Walsh (1998) and Tarkka and Mayes (1999) used this standard model. Our contribution to this model is that we added three additional equations to present the central bank’s losses. In addition, we adjust the sequence of events 6 For the purpose of this study high cost and losses are synonyms. 7 For more details about remedies see Dalton and Dziobek (2005), Stella (1997) and Leone (1993). 8 The Central Bank of Jordan (CBJ) which had losses in some years during the period (1996–2003).
O.D. Sweidan, B. Widner / Research in Economics 62 (2008) 45–54
49
Fig. 1. Sequence of events in repeatedly cycle.
to take these equations into consideration. The sequence of events creates one repeated cycle. The sequence of events and structure of information is as follows: first, at the beginning of period 0, the private sector forms its expected inflation rate (π0e ). At the end of period 0, the central bank observes these expectations, and at the end of this stage the central bank determines its exact inflation target (π¯ 0 ). Second, at the beginning of period 1, the central bank sets its monetary policy. The monetary authority bases its decision on how to set monetary policy on the past information it has, the new facts it receives and on its expectation of the future. In this period of time, we assume it is required by the central bank to respond by adopting a contractionary monetary policy. Then, the central bank realizes that it faces losses or high cost because of issuing its own securities. As a result, the central bank adjusts its money supply to reduce the losses. We assume the central bank gives the priority to heal its losses compared with inflation rate. The monetary authority decision’s maker can assess the right situation after the shock itself. The size of the losses depends on factors such as the level of the interest rate and the amount of the excess reserve the central bank intend to withdraw from the market. At this stage, the central bank observes the control error and as a result actual inflation rate. Third, in period 2 stochastic shocks of both demand and supply are realized. The central bank will take this new information set and influence its policy in the next cycle. Thus in period 2, the final effects on the output and inflation can be seen by the private sector and the central bank. Fig. 1, shows the decision making problem as a sequence of events. Let us assume that the aggregate supply of the economy is described by Lucas supply equation, as follows: Yt = Y¯ + a(πt − π0e ) + ε s t
(1)
where Yt is the real aggregate supply, Y¯ denotes the natural rate of output, πt stands for the inflation rate, π0e is the level of expected inflation rate by the public in period zero. ε s t is an aggregate supply shock. The aggregate demand of the economy is described by the following equation: Yt = Mt − πt + ε d t
(2)
Yt denotes the real aggregate demand, Mt is the money supply, πt stands for inflation rate and ε d t denotes the aggregate demand disturbance other than shifts in money supply. Following Tarkka and Mayes (1999) we assume that the aggregate demand disturbance has a mean (η) and a random error (φ) as follows: εdt = η + φ.
(3)
According to the loss function, it is a well-known consensus that the main goal of monetary policy should be price stability.9 The central bank decision maker seeks to minimize the difference between the actual inflation rate (πt ) and its inflation target (π0∗ ). The central bank sets its monetary policy by minimizing the following loss function: L C B = E C B,1 (πt − π0∗ )2 .
(4)
We will assume that the central bank has imperfect control over the inflation rate. The monetary authority will announce this fact, but they will not justify this imperfection. This implies that the central bank knows that the target cannot be achieved exactly. Also, we assume that the central bank responds to the losses in the sense that it appears in the determinants of the target variable.10 For this reason the central bank’s inflation target is as follows: π0∗ = π¯ 0 + µt 9 For more details see Cukierman (2002). 10 For more details about the difference between “target variable” and “responding to a variable” see Razin (2004).
(5)
50
O.D. Sweidan, B. Widner / Research in Economics 62 (2008) 45–54
where π¯ 0 is the central bank’s exact target of inflation rate which is announced at the end of period 0. π0∗ is observed in period 2. However, π¯ 0 is observed in period 0. µt denotes a monetary policy problem i.e. the expected losses facing central banks. We assume that the error (µt ) is not observed by the public. Therefore, we can rewrite Eq. (5) as π0∗ = π¯ 0 + λCt
(6)
where Ct stands for the cost of adopting a contractionary monetary policy. λ is the responsiveness of the central bank’s inflation target to a change in the central bank’s cost of adopting a tight monetary policy. Eq. (6) illustrates that the central bank’s inflation target varies with the cost constraint of the adopted monetary policy. Eq. (6) justifies and supports a conclusion by Masson et al. (1998) which says that in most developing countries the preconditions for adopting inflation targeting are not yet present. Based on the description presented above, we can write the relationship between the money supply and the cost to the central bank as follows: Ct = −Π Mt + ε ct
(7)
Mt denotes the money supply i.e. open market operations procedures. Π stands for the responsiveness of the central bank’s cost to a change in the money supply. εct is a cost shock i.e. quasi-fiscal activities.11 Eq. (7) shows a negative relationship running from the money supply to the cost of the policy. To complete the information about the model, we assume the expectations of the economic units in the model to different shocks as follows: 1- Expectations in period 0 E P S,0 (εdt ) = 0
E P S,0 (εst ) = 0
E P S,0 (φ) = 0.
2- Expectation at the beginning of period 1 E C B,1 (Y¯ ) = Y¯
E C B,1 (εdt ) = η
E C B,1 (φ) = 0
E C B,1 (εst ) = 0
E C B,1 (εct ) = 0
where PS,0 denotes private sector in period 0, and CB,1 stands for central bank in period 1. 4. Solving the model In this model, transparency is connected to the ability of both central bank and private sector to predict the expected inflation of the other part. However, the central bank’s losses are strategic and private information, and no one is going to release them at the present time. We do agree with Cukierman (2005) who considered that it is highly risky and counterproductive to disseminate information about a serious problem in the financial system. Distributing such a fact may cause runs on the banks or any other unexpected movement that threatens the stability of the financial system. As a result, the concept of transparency might not be available in some economic systems or environments. Moreover, Leone (1993) states that the income statement of the central banks in developing countries shows considerable profits despite the fact that the monetary authorities have significant problems. Also, he confirms that the method of registering the income and expenses of the central banks in these countries guarantees fictional profits. We do believe that the central bank’s losses in developing countries started to be an obvious global phenomenon after adopting the economic adjustment programmes with the IMF. The evaluation in the current model is in each single cycle not over the cycles. Based on the above-mentioned model and its sequence of events, we can derive the central bank’s expected inflation rate for period 2 at period 1 by minimizing the loss function, Eq. (4), subject to Eq. (6): E C B,1 (π2 ) = π¯ 0 + λE C B,1 (C E,1 )
(8)
where E,1: denotes expectation at end of period 1. Eq. (8) shows that the central bank’s expected inflation of period 2 in period 1 depends on two factors the inflation target set in period 0, (π¯ 0 ), and the expected cost or losses in period 1, {E C B,1 (C E,1 )}. Eq. (8) says that the cost of monetary policy has a positive effect on the central bank’s actual inflation target. Further, this equation is an indicator of the effect of the operational procedures of the central bank
11 For more details about those activities see Leone (1993).
O.D. Sweidan, B. Widner / Research in Economics 62 (2008) 45–54
51
on its expected actual inflation rate. At this point of time, the private sector believes that the central bank is working toward achieving its expectations in period 0, which is E P S,1 = π¯ 0 .
(9)
Eqs. (8) and (9) show the first effect of the central bank’s losses by creating a gap between the expectations of both the central bank and the private sector in period 1. By using Eqs. (1) and (2) we can solve the model to find out the equilibrium of the actual inflation rate in period 2, which is as follows: a 1 πe + (M1 − Y¯ + εdt − εst ). (10) 1+a 0 1+a The central bank sets the suitable monetary policy to affect the aggregate demand and to achieve its goals. When the central bank sets its monetary policy in period 1 it uses all the available information from the following sources: period 0, the new information and the expectations of the future. Hence, taking expectations of Eq. (10) conditional on the central bank’s available information, and using Eq. (8) gives: π2 =
E C B,1 (M1 ) = (1 + a)E C B,1 (π2 ) − a E C B,1 (π0e ) + E C B,1 (Y¯ ) + E C B,1 (εst ) − E C B,1 (εdt ) M1 =
(1 + a)π¯ 0 + (1 + a)λE C B,1 (C E,1 ) − a E C B,1 (π0e ) + Y¯
− η.
(11) (12)
Eq. (12) shows that the central bank sets its monetary policy based on five factors. The cost of adopting a contractionary monetary policy {E C B,1 (C E,1 )} is one of these factors and plays an important role in determining the money supply. To understand the behaviour of the central bank in Eq. (12), it is important to use Eq. (7) and the sequence of events. Assume the central bank in period 1 expects the private sector’s expected inflation rate will increase; therefore the money supply should decrease as it is obvious from Eq. (12). Consequently, Eq. (7) states that the decrease in money supply increases the cost of central bank. For this reason, the central bank adjusts to increase the money supply to prevent higher cost and more losses. From the other side, presume the central bank observed a positive aggregate demand shock (η) in period 2. Hence, the central bank, in period 1, intervenes by reducing the money supply. Consequently, Eq. (7) demonstrates again that this action boosts up the cost of the monetary authority. As a result, the money supply is adjusted upward. This mechanism helps to understand the effect of the central bank’s losses on monetary policy in the long run or over the repeatedly cycles by substituting Eq. (7) in Eq. (12), we get the following equation: M1 =
a 1 (1 + a) π¯ 0 − E C B,1 (π0e ) + (Y¯ − η). (1 + λΠ + aλΠ ) (1 + λΠ + aλΠ ) (1 + λΠ + aλΠ )
(13)
Eq. (13) confirms that the existence of the central bank’s losses in the long run reduces the value of coefficients affecting the money supply. The central bank will be more cautious when it reacts to any shock in the economy. With a monetary policy given by Eq. (12) and with an equilibrium inflation rate given by Eq. (10), actual inflation rate in period 2 is calculated as follows: 1 a {E C B,1 (π0e ) − π0e } + (φ − εs2 ). (14) 1+a 1+a Eq. (14) presents the sources of actual inflation in such an economy, which they are; the central bank’s exact inflation target (π¯ 0 ), the cost of adopting a contractionary monetary policy {E C B,1 (C E,1 )}, the error made by the central bank in period 1 in predicting the private sector’s expected inflation rate in period 0 {E C B,1 (π0e ) − π0e }, and errors of anticipating aggregate demand and aggregate supply disturbances (φ − εs2 ). Central bank loss is a constraint on monetary policy action to target the private sector expected inflation rate. That’s why the existence of central bank losses in each single cycle forces the inflation rate to increase. This result is consistent with the previous finding which is that losses increase the money supply in each single cycle. π2 = π¯ 0 + λE C B,1 (C E,1 ) −
5. Opacity The opacity concept in this model means that the central bank does not release its own information and forecasts regarding the demand shocks and any problem facing the monetary policy i.e. central bank losses. Thus, the private
52
O.D. Sweidan, B. Widner / Research in Economics 62 (2008) 45–54
sector does not have access to information and is unable to make assessments and projections in period 0 for period 1.12 If the private sector has a full sight in period 0, the expected inflation of the private sector is π0e = E P S,0 (π2 ) = E P S,0 (π¯ 0 ) + λE P S,0 {E C B,1 (C E,1 )} −
a 1 E P S,0 {E C B,1 (π0e ) − π0e } + E P S,0 (φ − εs2 ). 1+a 1+a
(15)
Based on the above definition of opacity, the two terms E P S,0 {E C B,1 (C E,1 )} and E P S,0 {E C B,1 (π0e ) − π0e } are equal to zero. Accordingly, we can rewrite Eq. (15) as π0e = E P S,0 (π¯ 0 ).
(16)
Eq. (16) illustrates that in the case of opacity the private sector is just able to see the exact target of the central bank which is announced in period 0. The error of the private sector expected inflation rate can be expressed by the difference between the actual inflation rate in period 2 and expected inflation rate in period 0. We can express this error by the following formula: Error P S,2 = π2 − π0e π2 − π0e = {π¯ 0 − E P S,0 (π¯ 0 )} + λE C B,1 (C E,1 ) −
(17) a 1 {E C B,1 (π0e ) − π0e } + (φ − εst ). 1+a 1+a
(18)
Eq. (18) shows that the absence of transparency guarantees the existence of error made by the private sector in predicting inflation rate. This error depends on the following sources: (1) (2) (3) (4) (5)
The failure of private sector to predict the central bank’s exact target of inflation rate. The cost of adopting a tight monetary policy or the cost constraint. The failure of the central bank to forecast the private sector expected inflation rate. The aggregate supply shock. The random error of the aggregate demand shock.
Eq. (18) is similar to Eq. (3.15’) derived by Tarkka and Mayes (1999), the new thing we have in our formula is the effect of the cost of adopting a contractionary monetary policy. Moreover, we can calculate the fluctuation in output under opacity by substituting Eq. (18) in Eq. (1) as follows: a2 1 a Y2 − Y¯ = a[{π¯ 0 − E P S,0 (π¯ 0 )} + λE C B,1 (C E,1 )] − {E C B,1 (π0e ) − π0e } + φ+ εs 2 . 1+a 1+a 1+a
(19)
Eq. (19) confirms the role of the cost constraint of adopting a contractionary monetary policy on the departure of actual from potential output. 6. Transparency As defined above, transparency in the current paper means that the central bank releases all the information and predictions it has regarding the macroeconomic variables. However, the central bank will not release any information regarding the central bank losses. Contrary to the opacity case, the private sector is able to use the information and expectations starting from period 1. Thus, the expected inflation rate made by the private sector looks like Eq. (15). The error made by the private sector in predicting inflation rate is expressed by π2 − π0e = {π¯ 0 − E P S,0 (π¯ 0 )} + λ{E C B,1 (C E,1 ) − E P S,0 {E C B,1 (C E,1 )}} −
a 1 {{E C B,1 (π0e ) − π0e } − E P S,0 {E C B,1 (π0e ) − π0e }} + (φ − εs2 ). 1+a 1+a
(20)
Based on the definition of transparency, the term E P S,0 {E C B,1 (C E,1 )} is equal to zero. The private sector is unable to perceive the problem of losses facing the central bank. Moreover, transparency enhances the forecasting power of 12 If the private sector is unable to make projection from period 0 this implies that π e = 0. 0
O.D. Sweidan, B. Widner / Research in Economics 62 (2008) 45–54
53
the private sector. Thus, the private sector is more professional in anticipating the central bank’s projected action. The private sector can predict the error made by the central bank in predicting the private sector expected inflation rate. That is why the term {{E C B,1 (π0e ) − π0e } − E P S,0 {E C B,1 (π0e ) − π0e }} is equal to zero. The formula [Eq. (20)] of the error made by the private sector in predicting inflation rate is reduced to 1 (φ − ε s 2 ) 1+a a 1 Y2 − Y¯ = a[{π¯ 0 − E P S,0 (π¯ 0 )} + λE C B,1 (C E,1 )] + φ+ εs 2 . 1+a 1+a
π2 − π0e = {π¯ 0 − E P S,0 (π¯ 0 )} + λE C B,1 (C E,1 ) +
(21) (22)
Eq. (21) shows that the departure of the private sector expected inflation rate from the actual inflation rate is less in transparency case compared to opacity case in Eq. (18). This result is confirmed by Tarkka and Mayes (1999).13 Further, the current paper brings four main conclusions: First, the cost constraint of adopting a tight monetary policy in both transparency and opacity alike is significant and affects positively the error of the private sector expected inflation rate [Eqs. (18) and (21)], and the output gap [Eqs. (19) and (22)]. Technically, the results of our paper show that the cost constraint of monetary policy influences the actual inflation rate, but does not affect the private sector expected inflation rate. Second, it is obvious from the above results that the benefits of transparency and the effect of cost constraint on the conduct of monetary policy move in two opposite directions; transparency reduces fluctuation in both inflation error and output gap, while the cost constraint leads to more fluctuation. Three, based on the information they have, the private sector understands that the policy of the central bank is enhanced by transparency. However, the private sector at the end of each single cycle notices the gap between actual and predicted macroeconomic variables i.e. output and inflation rate, therefore the private sector will be surprised. Consequently, the private sector is unable to understand the reason for the gap. This contradiction between what is expected and what is seen in reality might put a big question on the accuracy, credibility and accountability of the transparency process. Four, the results of the current paper point to the possibility that inflation rate acceleration might be a serious problem in developing countries especially those with constraints on conducting their monetary policies. 7. Conclusions We believe that developing countries are not subject to all theories applied to developed countries. The privacy of the problems in developing countries plays a crucial role in this mismatch. The current paper focuses on the transparency of central banks in developing countries. This issue looks important because recent evidence shows that the central banks in some developing countries suffer from losses. We think this issue is unique, important and never expected to happen. We try to answer the following question: Would more transparency in developing countries suffering from central bank losses be beneficial? To achieve our goal, we utilize a model used by other economists to study transparency. However, we upgrade part of the model to shed light on the uniqueness of developing countries. This paper illustrates that the cost of adopting a contractionary monetary policy has a positive relationship with the central bank’s inflation target and actual inflation rate. In addition, in the long run, monetary policy will be less sensitive to any shock in the economy. These results create a gap between the actual macroeconomic variables i.e. inflation rate and output and its expected value by the private sector. If the private sector believes it is living within a transparent economic system, many questions arise concerning the accountability of monetary forecasts. Therefore, we believe that the existence of both transparency and cost constraint on the conduct of monetary policy will not stay forever. The constraint works to terminate transparency. The outcome of the current paper proves that the effect on inflation error and output gap moves in two opposite directions. Therefore, transparency in developing countries with central bank losses could be beneficial for a very short period of time, and in the long run we expect that transparency vanishes. It is a necessary condition to have a suitable environment for transparency, and it is unwise for developing countries suffering from central bank losses to focus on transparency; priority should be given to fixing monetary policy and to develop the financial markets. We would like to confirm that repairing monetary policy in developing countries should track two dimensions: first, central banks suffering operational losses because of adopting open market operations should switch to monetary 13 Comparing the difference between transparency and opacity is not a target of the current study.
54
O.D. Sweidan, B. Widner / Research in Economics 62 (2008) 45–54
policy that lowers the cost as required by the reserve ratio. Second, we reconfirm the procedures proposed by the IMF which enforce governments to recapitalize central banks facing losses, and commit governments to adopt this action in advance. Acknowledgements We would like to thank the editor and two anonymous referees of Research in Economics Journal for valuable and helpful comments. We are responsible for any remaining errors. References Axilord, S.H., 1997. Transformations to Open Market Operations. International Monetary Fund, January. Washington. Blinder, A., Goodhart, C., Hildebrand, P., Lipton, D., Wyplosz, C., 2001. How Do Central Banks Talk? International Center for Monetary and Banking Studies. Geneva, Switzerland. Chandavarkar, A., 1996. Central Banking in Developing Countries. St. Martin’s Press Inc. Chortareas, G., Stasavage, D., Sterne, G., 2002. Does it Pay To Be Transparent? International Evidence from Central Bank Forecasts. Review, Federal Reserve Bank of St. Louis. July/August. pp. 99–118. Cukierman, A., 2002. Are Contemporary Central Banks Transparent About Economic Models and Objectives and What Difference Does it Make? Review, Federal Reserve Bank of Saint Louis. July/August. pp. 15–35. Cukierman, A., 2005. The Limits of Transparency. Unpublished. Dalton, J., Dziobek, C., 2005. Central Banks Losses and Experiences in Selected Countries. IMF Working Paper No. WP/05/72, April. Washington. Enoch, C., 1998. Transparency in Central Bank Operations in the Foreign Exchange Market. IMF Paper on Policy Analysis and Assessment, PPAA/98/2, March. Washington. Faust, J., Svensson, L.E., 2001. Transparency and Credibility: Monetary Policy with Unobservable Goals. International Economics Review 42 (2), 369–397. Gerrats, P.M., 2001. Why Adopt Transparency? The Publication of Central Bank Forecasts. Working Paper No. 41, European Central Bank. Gerrats, P.M., 2002. Central Bank Transparency. The Economic Journal 112 (November), F532–F565. Hahn, V., 2006. The Transparency of Central Bank Preferences, June, Available at: SSRN: http://ssrn.com/abstract=592261. Jensen, H., 2000. Optimal Degree of Transparency in Monetary Policymaking. Working Paper, University of Copenhagen. Leone, A., 1993. Institutional and Operational Aspects of Central Bank losses. IMF Paper on Policy Analysis and Assessment, PPAA/ 93/ 14, September. Washington. Masson, P.R., Savastano, M.A., Sharma, S., 1998. Can Inflation Targeting Be a Framework for Monetary Policy in Developing Countries? Finance and Development, March. pp. 34–37. Mishkin, F.S., 2004. The Economics of Money, Banking, and Financial Markets, seventh edition. Addison-Wesley. Mishkin, F.S., 1999. Central Banking in a Democratic Society: Implications for Transition Countries. Zagreb Journal of Economics 3 (3), 51–74. Quintyn, M., 1994. Government Securities versus Central Bank Securities in Developing Open Market Operations. IMF Working Paper No. 94/62, May. Washington. Razin, A., 2004. Aggregate Supply and Potential Output. Working Paper 10294, National Bureau of Economic Research. Stella, P., 1997. Do Central Banks Need Capital. IMF Working Paper No. 97/83, July. Washington. Stiglitz, J., 1998. Central Banking in a Democratic Society. De Economists 146 (2), 199–226. Sweidan, O., Maghyereh, A., 2006. Monetary Policy and the Central Bank’s Securities. Applied Economics Letters 13 (9), 593–598. Tarkka, J., Mayes, D.G., 1999. The Value of Publishing Official Central Bank Forecasts. Discussion Paper 22/99, Bank of Finland. Thornton, D.L., 2002. Monetary Policy Transparency: Transparent About What? Federal Reserve Bank of St. Louis, Working Paper 2002-028B, November. Walsh, C.E., 1998. Monetary Theory and Policy. MIT Press, London.