The consistent theme of recent weeks on this blog has been the topic of financial market imperfections. We have looked at this area from a policy perspective (see here and here), and also an underlying theoretical perspective (see here). The challenge for macroeconomics will always be how to incorporate these financial market imperfections into macro models, so that we can further our analysis and understanding of these issues. This is exactly what Vasco Curdia and Michael Woodford (2010, hereafter C&W) have done with their Dynamic Stochastic General Equilibrium (DSGE) model. While controversial (see anything by Australian economist Steve Keen), the work of C&W in a DSGE context has provided valuable insights into the possible conduct of monetary policy in a world of imperfect financial markets.
The DSGE model of C&W – A New Keynesian model with financial frictions
Even a relative simple DSGE model can be quite complicated, so a full explanation will not be provided here. Instead, the main parts of the model will be highlighted before focusing on the financial aspects of the model.
- C&W depart from the traditional representative household and assume that households differ in their preferences: some are borrowers and some are savers. This basically separates households on their impatience to consume.
- Households can only spend a different amount to their income by borrowing or lending to financial intermediaries. This relationship provides the key financial friction in the model.
- ‘Borrower’ households choose always to borrow, while ‘saver’ households choose always to save.
- Labour supply will be the same for households of each type, dependent upon the wage markup, marginal utility of real income and price level.
Financial disturbances in the credit market are found in the credit spreads, which is the difference between borrowing and lending rates in this model. Firstly, real resources are consumed in the process of originating loans. Secondly, some portion of the loans made will not be repaid, representing a loss to the financial intermediary. These two sources of credit spreads are exogenous.
The first financial disturbance mentioned here requires more explanation. C&W refer to ‘convex technology’ in the origination of these loans. As ‘real’ resources are required here, this technology refers to the notion that there is a finite capacity to lend at any point in time. This capacity is limited due to expertise available and necessary capital. This means that an increase in the volume of lending will lead to an increase in the equilibrium credit spread.
Spread adjusted Taylor Rules
To test the model’s ability to respond to financial frictions, the following spread adjusted Taylor rule was implemented for the central banks reaction function:
This has the usual inflation (pi) and output gap terms (Y), but the crucial term is on the right – the omega parameter. This is the interest rate spread and the coefficient on this term represents the influence that this spread can have on the main interest rate of the central bank.
The results of this DSGE analysis are very interesting, but the key point from this paper is that a single recommendation in terms of monetary policy and financial frictions is not forthcoming.
Beginning with financial disturbances, the results show the benefit of a spread-adjusted Taylor rule over an adjusted one. With the coefficient on the spread parameter set to zero (the unadjusted Taylor rule), the increase in the spread sees a fall in aggregate credit, a fall in real activity and a drop in inflation. This part of the analysis finds an optimal adjustment coefficient greater than 0.5 but less than 1.
The results relating to financial shocks are not clear-cut however. In the case of less convex intermediation technology, the optimal spread adjustment is very sensitive to the persistence of the shock. Thus, C&W state that a ‘contemporaneous spread adjustment cannot be found that will have desirable consequences in all circumstances.
Contrasting results are also found when the model examines non-financial disturbances. When C&W examined various shocks to aggregate expenditure, a positive spread adjustment leads to looser monetary policy and an even less ideal response than the baseline Taylor rule. This is due to the shock increasing credit demand and hence the credit spread, resulting in the monetary response that is too loose. This finding highlights the general point of this study in that the type of spread adjustment that is desirable in some circumstances will actually be problematic in others.
Implications for monetary policy: Rules versus flexible inflation targeting
The Taylor rule adjusted shocks have been very insightful in this paper, but C&W take the opportunity to highlight the primacy of flexible inflation targeting over such an instrument rule. The authors cite an example of a flexible approach with the following criterion:
Where pi is inflation and x refers to the output gap. With this approach to monetary policy, policy is only optimal if this criterion is satisfied in each period. The author’s previous work has shown that this provides a good approximation to optimal policy even in the presence of heterogeneity and credit frictions. Moreover, this is not dependent upon the type of disturbance and is also robust to the persistence of the shock. Hence, it provides a more appropriate framework to monetary policy than even the spread adjusted Taylor rule.
The key point here is that a flexible monetary approach allows more discretion in these complicated areas than an instrument rule. Central banks will need to consider the source of the disturbance, as well as multiple credit spreads and the likely degree of persistence of any shock. Overall, if a simple rule worked for all disturbances, the job of the central bankers would be much easier. As this is not the case, it reinforces the work currently going on to understand and incorporate financial market imperfections into both macro models and macroprudential policymaking.