 |
4/2007 |
| Much of the recent macroeconomic research on monetary policy relies on New Keynesian business cycle models where there is little or no role for banks or similar specialized financial institutions in the transmission mechanism for monetary policy. As the saying goes, banks are not special for the transmission of monetary policy. The underlying reason is not too difficult to find, once it is recognized that financial market behaviour in these models, always described as an integral part of the microeconomics of these models, is not subject to the type of frictions that are, on the other hand, much emphasized by research on the effect of informational asymmetries on credit market and also more generally on financial market relationships. New Keynesian models for monetary policy analysis put a lot of emphasis on the forward looking expectations in affecting the monetary policy transmission process and the implications this has both for policy design and central bank communication. It is because expectations matter that the systematic behaviour of the central bank plays a critical role in determining the real effects of monetary policy, as only the systematic part of policy will affect future expectations. This is in sharp contrast to the conclusions drawn by many researchers back in the 1970's and early 1980's that what matters for the real effects of monetary policy is the ability of central banks to surprise the private sector. Modern New Keynesian policy models furthermore bring the interest elasticity of aggregate demand right to the core of the transmission mechanism of monetary policy, at the same time de-emphasizing the role of monetary aggregates for policy effects. More specifically, by being able to systematically affect the real interest rate, the central bank is in these models able to affect the cyclical behaviour of aggregate demand and, hence, output. Since the central bank is modelled as following an interest rate rule – Taylor rule – the amount of money is determined endogenously by the demand for money. This is the reason for the part of the individuals' optimization problem pertaining to the determination of optimal money holdings to be put into the background of the main analysis, at least in the case of additively separable per period utility function in consumption and real money holdings. |
| |
| So in order for bank lending or, more generally, credit market conditions to affect the monetary policy transmission process, banks or financial intermediation needs to be special. Economists have presented a number of good arguments for banks to be special. Some suggest that bank balance sheets effects, for example the interaction of bank liquidity or net worth with loan interest rate elasticities, magnify the impact of monetary policy. However, it is only recently that economists have introduced banks into monetary business cycle models, or dynamic stochastic general equilibrium models of the sort used for monetary policy analysis. In their forthcoming Bank of Finland discussion paper "The lending channel under optimal choice of monetary policy" Juha Kilponen and Alistair Milne also take a step towards this direction to incorporate banks into an otherwise standard monetary business cycle model of the New Keynesian type. Monetary policy is optimized through the central bank seeking to minimize a quadratic loss function in the output gap and inflation. Their analysis has two distinctive features. The first is that they allow banks to hold capital buffers in excess of the minimum regulatory capital requirements and for the magnitude of this capital buffer to affect the supply of bank lending. Consequently, the authors take into account the possibility, often pointed out in the relevant reference literature, that bank capital regulations have undesirable macroeconomic impacts. Many commentators also fear that the New Basel II accord will exaggerate cyclical fluctuations of output, since under the advance IRB calculations of the new accord regulatory capital requirements are increased in cyclical downturns, which may contribute to constraining bank lending. However, as Kilponen and Milne rightly note, most previous analysis of the macroeconomic impact of bank capital makes the mechanical and counterfactual assumption that regulatory capital requirements directly constrain the volume of bank lending. But banks dynamically optimize the level of their capital buffers and, as several recent studies have pointed out, banks almost always operate with buffers of capital in excess of minimum regulatory requirements. Hence, in the model used by Kilponen and Milne, a shortfall of this buffer reduces bank loan supply and increases bank loan margins. |
| |
| The second distinctive feature of their model is the allowance for the possibility of a 'cost channel' of monetary policy in the sense that either or both market and bank interest rates directly affect marginal costs of production. This is an interesting extension of existing models, as in the presence of a cost channel a tightening of monetary policy has the effect not just of reducing output and, hence, lowering future inflation, but also of raising current inflation. The existence of a cost channel thus constrains the ability of the central bank to respond to shocks. As Kilponen and Milne show in their analysis, the presence of a cost channel is an important determinant of the macroeconomic impact of bank capital. Without a cost channel, monetary policy can be re-adjusted to offset the effect of changes of bank capital or capital regulation on output and inflation. Although their model is in other respects a standard reduced form New Keynesian monetary business cycle model, where the dynamics of output gap and inflation depend upon both expectations of future output gap and inflation as well as lagged output gap and inflation, they rightly point out that the set up is subject to some inherent limitations. The model considers only linearised dynamics for small deviations of the endogenous variables around their trend values. However, the theory of precautionary bank capital suggests that there should be a nonlinear relationship between bank net worth and bank lending, with relatively small macroeconomic effects of bank net worth when bank capital is close to desired levels. Consequently, their results should be interpreted with some care. These concerns may actually show up in their simulation results, which suggest that with reasonable parameterization of the model, the bank lending channel has only small effect on output-inflation trade-offs. The authors are able to extend some of the previous research, as they find that the response of output and inflation to both demand and supply shocks are completely unaffected by the bank lending channel, as long as interest rates operate 'the conventional way', that is entirely through intertemporal trade-offs and aggregate demand. Consequently, Kilponen and Milne find that the bank lending channel can alter output-inflation trade-offs if there is a cost channel, meaning if interest rates alter marginal costs and thus directly affect inflation through the hybrid Phillips curve describing the dynamic relationship between inflation and marginal costs, where the latter is approximated by the output gap. Their results also show that the magnitude of the resulting bank loan channel impact on output and inflation depends on the relative contribution of bank loan rates, in comparison to market rates, on aggregate demand and on marginal costs. More specifically, if bank loan rates have a relatively large impact on aggregate demand and relatively small effect on marginal costs, then larger increases in interest rates are required to achieve a given reduction of aggregate demand and output-inflation trade-offs deteriorate. As such these results are interesting and may indicate that the bank lending channel plays, after all, an important role in the transmission of monetary policy. However, the analysis of Kilponen and Milne is eminently important also from another perspective too. Understanding the role of banks or more generally specific financial intermediaries and institutions for business cycle fluctuations may actually require us to shuffle our toolkit of models for business cycle fluctuations to make room for new models that are more efficient in enhancing our understanding of the relationship between financial intermediation and business cycle fluctuations. We as economist should pick up the gauntlet and put this in the agenda for future research. |