نقدینگی اضافی، قواعد قیمت گذاری بانکی، و سیاست های پولی

This paper studies the implications of excess bank liquidity for the effectiveness of monetary policy in a simple model with credit market imperfections. The demand for excess reserves is determined by precautionary factors and the opportunity cost of holding cash. It is argued that excess liquidity may impart greater stickiness to the deposit rate in response to a monetary contraction and induce an easing of collateral requirements on borrowers – which in turn may translate into a lower risk premium and lower lending rates. As a result, asymmetric bank pricing behavior under excess liquidity may hamper the ability of a contractionary monetary policy to lower inflation.

A recurrent concern of central bankers is the possibility that an abundance of liquidity may hamper the ability of monetary policy to influence the level of economic activity and inflation. In most financial systems, excess (bank) liquidity can be defined as the involuntary accumulation of liquid reserves by commercial banks. 1 Thus, although banks may choose to hold reserves above and beyond what is required by the legislation (to satisfy unexpected withdrawals of cash from their clients, for instance), excess liquidity prevails only if they unwillingly hold more cash than desired. In a crisis environment, characterized by increased volatility, the demand for reserves may increase sharply – either for precautionary reasons or because (risk-averse) banks find it too risky to lend. 2 This observation, as argued by Agénor et al. (2004), provides a useful starting point for identifying the source of a credit crunch: if banks are unwilling, rather than unable, to extend loans, a contraction in credit can be attributed to supply, rather than demand, factors. To perform this test requires therefore estimating a demand function for excess liquid assets by commercial banks and examining its predictive capacity. Involuntary accumulation of liquidity can then be determined residually, and its statistical significance assessed by a variety of tests. From an analytical standpoint, it is convenient to classify factors leading to excess liquidity into structural and cyclical determinants. The first structural factor that is commonly identified is a low degree of financial development. In countries with less developed financial sectors, banks (and their customers) will tend to have a greater demand for liquidity. In particular, unreliable money payment systems may induce banks to choose to hold a relatively large buffer of reserves to help them regulate their liquidity needs. The costs of processing information, evaluating projects, and monitoring borrowers may also be relatively high; in turn, this may complicate liquidity management and may lead to an accumulation of reserves beyond desired levels. This is one of the main explanations of the high and persistent levels of bank reserves (above and beyond those required by regulations) in the low-income countries of Central Africa, where opportunities for portfolio allocation are limited (see Saxegaard, 2006). A second factor is a high degree of risk aversion, which leads to high risk premia and a low demand for credit. The degree of risk aversion, in turn, may be directly related to chronic macroeconomic instability, and this may explain a positive, long-run correlation between high inflation and excess liquidity. Inflation may also represent a cyclical cause of excess liquidity. To the extent that it is accompanied by higher volatility in relative prices (and thus an increase in the riskiness of investment projects characterized by a high degree of irreversibility), a surge in inflation may raise uncertainty about the value of collateral pledged by borrowers – leading banks, if confronted with adverse selection problems, to either charge a higher risk premium or increase the incidence of credit rationing. Because, in the former case, a higher lending rate would typically lead to a contraction in credit demand, both responses may translate into an involuntary accumulation of excess reserves. Another important cyclical factor is large capital inflows intermediated by the banking system. In the past two decades, a number of developing countries, low- and middle-income alike, have indeed implemented measures to foster an asymmetric opening of the capital account (that is, a lifting of restrictions on capital movements for non-residents, while retaining a wide array of controls on foreign exchange operations by residents). In many cases these measures led to large capital inflows, often associated with privatization of large-scale state enterprises. In Guyana for instance, external financial liberalization was accompanied in the late 1990s by a dramatic increase in excess reserves (see Khemraj, 2007). In Morocco, a number of large privatization operations and increased amounts of foreign direct investment led in recent years to a significant increase in liquidity in the banking system, prompting the central bank to raise reserve requirements sharply to avoid the development of inflationary pressures (see Agénor and El Aynaoui, 2007). Between 2006 and 2008, China, India, and Korea all raised required reserve ratios on bank deposits several times in an attempt to mop up excess liquidity. In countries operating a pegged exchange rate regime, upward pressure on the nominal exchange rate created by large capital inflows often leads to sustained central bank intervention and a build-up of official holdings of foreign exchange; in the absence of sterilization, surplus reserves translate into an expansion of the monetary base and rapid accumulation of excess liquidity by commercial banks – with possibly destabilizing macroeconomic effects. Similar outcomes have been observed in managed float regimes, where the central bank intervenes to maintain the exchange rate within a (more or less stable) target range. Surprisingly enough, and despite the importance of the issue for central banks, there have been few attempts to explore analytically the implications of excess liquidity for the effectiveness of monetary policy.3 This paper attempts to bridge this gap by proposing a simple macroeconomic model where banks’ pricing behavior and portfolio decisions are explicitly accounted for. Given the static nature of our framework, we do not specifically identify the source (structural or cyclical) of excess liquidity; instead, we focus on the case where the (voluntary) motive for holding excess reserves is uncertainty about cash withdrawals by the public, and define excess liquidity as a situation where actual excess reserves exceed the desired value. Although our framework is fairly general, the countries that we have in mind are middle-income countries, where the financial system is sufficiently developed to allow monetary policy to operate through the manipulation of a short-term interest rate whose “pass-through” effect on market rates is fairly rapid, as in more developed countries; in many low-income countries, by contrast, monetary policy is often based on indirect instruments. At the same time, however, we assume that the financial system is dominated by banks and that capital markets are either underdeveloped or illiquid – in line with the evidence for middle-income countries. Thus, firms in these countries (unlike their counterparts in more developed countries) have no real alternative but to either use retained earnings or borrow from commercial banks if they must cover production costs prior to the sale of output.4 The rest of the paper is organized as follows: Section 2 presents the basic model, which dwells on the framework for short-run monetary policy analysis developed in Agénor and Montiel, 2006 and Agénor and Montiel, 2008a. Because open-economy considerations are somewhat tangential to the issue at hand, we deliberately simplify the approach followed in those papers by focusing on a closed economy. Credit market imperfections are introduced by assuming that commercial banks set both deposit and lending rates, in the latter case as a premium over the cost of borrowing from the central bank. The premium itself depends on firms’ net worth, in the tradition of Bernanke and Gertler (1989). In addition, we also derive explicitly a demand function for excess reserves, which we relate to precautionary factors and opportunity cost variables. After characterizing the model’s solution in Section 3, it is used in Section 4 to examine the financial and real effects of a change in the official cost of borrowing and the required reserve ratio. The analysis is then extended in Section 5 to introduce an asymmetric effect of excess liquidity on bank pricing rules and determine how these asymmetries affect the effectiveness of monetary policy. This section represents the essential contribution of this paper; it has been argued in some previous contributions that excess liquidity may generate asymmetric pricing behavior, but (as far as we know) this has not been formally analyzed before. Our key insight is that excess liquidity may impart greater stickiness to the deposit rate in response to a monetary contraction and induce an easing of collateral requirements on borrowers – which in turn may translate into a lower risk premium and lower lending rates. Asymmetric price response under excess liquidity may therefore hamper the ability of a contractionary monetary policy to fight inflationary pressures. The last section offers some concluding remarks and suggests some directions for future research.

It has often been observed that when banks hold liquid funds in excess of their desired level, the transmission mechanism of monetary policy may be considerably altered. The purpose of this paper has been to propose a simple macroeconomic model that allows a formal analytical exploration of the implications of excess bank liquidity for the effectiveness of monetary policy in the presence of asymmetric bank response. In the model, a cost channel is introduced by assuming that firms must borrow to pay wages prior to the sale of output. Banks set both deposit and lending rates, in the latter case as a premium over the cost of borrowing from the central bank. The premium itself depends on firms’ net worth, in the Bernanke–Gertler tradition. Firms do not face binding credit constraints but are able to borrow funds only at ever higher interest rates if they cannot pledge sufficient collateral. A demand function for excess reserves is also explicitly derived, under the assumption that banks face uncertainty about cash withdrawals by the public. The basic model is used to examine the financial and real effects of changes in the cost of borrowing from the central bank and the required reserve ratio, in a “standard” case where excess liquidity (defined as a situation where actual excess reserves exceed the desired value) exerts no direct effect on bank pricing behavior. It is shown, in particular, that because aggregate supply and aggregate demand both fall, prices may either increase or fall. In the former case, a familiar Bernanke–Gertler financial accelerator effect was identified, as a result of endogenous changes in net worth in the risk premium. In the latter, however, a financial “decelerator” may be at play, for similar reasons. The analysis is then extended to consider the case where excess liquidity (defined as involuntary holdings of cash) prevails and induces asymmetric bank pricing rules. In the first case that we consider, deposit rates are assumed to exhibit upward rigidity only, in response to changes in monetary policy instruments: banks may be more willing to cut deposit rates following either a reduction in the refinance rate or an increase in the required reserve ratio, given that both policies lead to a reduction in the supply of deposits by households. The model therefore provides a rationale for the decision (often observed in practice) to raise required reserve ratios to “sterilize” excess liquidity. In addition, we find that if banks choose not to adjust deposit rates at all following an increase in the refinance rate under excess liquidity, a contractionary monetary policy may be less effective in reducing inflation. We also consider the case where excess liquidity leads to asymmetric behavior in setting lending rates, by inducing banks (in an attempt to stimulate the demand for credit) to relax credit standards. We argue that such behavior could bring about a risk premium that is lower than otherwise (or totally unresponsive), and this can mitigate the direct upward effect of an increase in the refinance rate on the lending rate. In such conditions, a contractionary monetary policy may again be less effective in reducing price pressures. Thus, our analysis supports the view adopted (less formally) by various observers, according to which the effectiveness monetary policy may indeed be limited in an environment where excess liquidity prevails in the financial system. The analysis presented in this paper can be extended in several fruitful directions. First, although increases in the reserve requirement rate are shown to be an effective monetary policy instrument under excess liquidity, the scope for using it may be limited if reserves held at the central bank must be remunerated – even at below-market interest rates. In addition, if the increase in liquidity is due to cyclical factors (such as large increases in capital inflows, for instance), the possible adverse long-term effect of higher required reserve ratios on financial intermediation may create dynamic trade-offs and further constrain the use of these ratios as instruments of short-term macroeconomic management. Second, relaxing credit standards to stimulate the demand for loans may also end up lowering the quality of banks’ portfolios – thereby making them more vulnerable to default. In that sense, excess liquidity can make the financial system more prone to crisis. An important issue therefore is to analyze what policies, in a second-best environment, can effectively mitigate increased vulnerability. This may offer a different perspective on the role of reserve requirements. Third, a broader menu of collateralizable assets in the economy could be considered, to better account for possible portfolio shifts induced by excess liquidity. As noted earlier, excess liquidity may push banks toward riskier uses of deposits – not only lending to firms (as discussed earlier), but also investment in assets whose collateral value may be highly volatile, such as real estate. Greater exposure to risk could weaken banks’ balance sheets and increase their vulnerability in a downturn. Fourth, in the model the asymmetric nature of the bank’s pricing rules was assumed, rather than derived from profit maximization. Although the rules specified here are plausible, it would be worth deriving them more rigorously. One possibility is to explore how the dynamic, stochastic models of asymmetric price behavior for “real sector” firms developed by Ball and Mankiw, 1994 and Ravn and Sola, 2004, and others, which account for menu costs, can be adapted to bank optimal pricing behavior under excess liquidity. In the spirit of the Ball–Mankiw approach, for instance, it could be assumed that (monopolistically competitive) banks can costlessly set interest rates every second period, but are subject to a menu cost if they change them between periods. If banks face menu costs in changing interest rates, small deviations of actual reserves from desired levels will not result in a price adjustment, whereas larger deviations that exceed some threshold will. An alternative (and perhaps more directly related) approach is to adapt the model presented in Florio (2006), which incorporates matching frictions on the credit market, investment irreversibility, and endogenous bargaining between lenders and borrowers. In the model, an asymmetric effect of monetary policy results from an asymmetric response of loan rates, due to changes in borrowers’ balance sheets.34 To adapt Florio’s model to our purpose here, one avenue worth exploring is to assume that lenders’ bargaining power falls in situations of excess liquidity (as a result of a reduction in their perceived net worth) and induces them to accept a larger proportion of weaker projects to finance – in effect, weakening collateral requirements and loan covenants, and therefore reducing the risk premium, as discussed earlier. Finally, it is also important to expand empirical research by considering a broader group of countries when testing for whether excess liquidity leads indeed to asymmetric bank pricing behavior. In particular, it is important to establish whether the speed of adjustment in deposit rates is significantly different depending on whether excess liquid reserves are above or lower an estimated equilibrium (desired) value. As noted earlier, we are aware of only one study, for Morocco, that explicitly supports this conclusion (Lebedenski, 2007). The error-correction approach used by Scholnick, 1996, Sander and Kleimeier, 2002 and Kwapil and Scharler, forthcoming, and others, but focusing specifically on episodes of excess liquidity, could be useful for that purpose. Such studies are essential for strengthening our understanding of how asymmetric bank pricing behavior affects the monetary transmission mechanism.