RDP 2011-04: Assessing Some Models of the Impact of Financial Stress upon Business Cycles 3. Evidence of Financial Effects on Aggregate Activity

What is the evidence concerning the impact of financial factors upon the aggregate level of activity? Here we exclude questions relating to the impact of the short-term interest rate as these generally appear in the core model. Instead we ask about the evidence on the impact of credit conditions upon aggregate activity. A good summary of this evidence has been provided in sources such as IMF (2009), and here we select four conclusions from that document. There are more ‘stylised facts’ but these seem to be a useful starting point.

  1. In the first two years of an expansion after a financial crisis, real credit grows quite weakly, at a slower rate than output.[6]
  2. The probability of the economy entering a recession increases markedly once the external finance premium exceeds some ‘crisis level’.
  3. The probability that an economy will stay in a recession beyond a certain number of quarters is higher when the onset of a recession was accompanied by a financial crisis. A crude interpretation of this would be that recessions with a financial crisis are of longer duration.
  4. A measure of financial stress can help predict output growth. Moreover, real investment growth shows even greater predictability. The latter seems to imply that the effects of financial factors will be greater on investment than on aggregate economic activity. In particular, the cycle in investment expenditure should be more closely related to credit conditions.

To look at these outcomes in the context of a model, a measure of financial stress is needed. In the IMF (2009) work the dates of financial crises were taken to be those identified by Reinhart and Rogoff (2008), who used a ‘narrative’ approach to find them. Here we need a measure that can be generated by any model augmented with financial effects. Because the financial stress measure aims to quantify the extra costs that firms encounter if they are required to borrow, one guide would be the size of the external finance premium. Thus, ideally, a crisis would be defined as occurring when the finance premium gets above a certain level but, as this is unlikely to be easy to determine, we simply investigate relationships as the level of the premium rises.

The features noted above require that one locate turning points in the level of economic activity in order to compute the characteristics of the business cycle, so as to locate the dates when an expansion or a recession started. For this purpose we use the BBQ program, which is described in Harding and Pagan (2002) and is a quarterly version of the method for locating turning points set out in Bry and Boschan (1971).[7]


Strictly speaking, in IMF (2009) this is for the manufacturing sector rather than aggregate output. [6]

A modified version of the program, written by James Engel, is used and is available at http://www.ncer.edu.au/data/. [7]