Microsoft Word Thesis Gent (1). doc

6.1.1 Impulse response function 
When the impulse response function (IRF) to structural shocks is calculated a certain 
recursive structural VAR could be the adequate model. To proceed with the impulse 
response function of dynamic equation, in general the output of the the impulse 
response function is a dynamic reaction of particular variable (e.g. NPL ratio) in 
response to other macroeconomic variable (e.g exchange rate). The impulse response 
function describe the reaction of variables in equation as function of time of 
explanatory variables that may cause the dependent variable. 
These structural shocks impact the variables present in the model. And through the 
model impulse graph can be constructed as response to a certain shock. 
In my case, my aim is to see the response of credit risk to macroeconomic shocks. 
Defining credit risk, I used the ratio of non-performing loan to total loans and as key 
macroeconomic variables I have collected variables such as long-term interest rate, 
inflation, exchange rate (czk/usd), GDP, and unemployment rate. Every structural 

---

47 
shock affects every other variable. Thus, I can construct an impulse graph for credit risk 
variable as the response to a certain macroeconomic shock.
§
The impulse response of credit risk in response to a long-term interest rate 
shock 
§
The impulse response of credit risk in response to a inflation shock 
§
The impulse response of credit risk in response to a exchange rate shock 
§
The impulse response of credit risk in response to a GDP shock 
§
The impulse response of credit risk in response to a unemployment shock 
The variables in equation tend to have the same kind of dynamics, which means that 
they have same dynamic equation representation, thus, the shape of impulse response 
functions tend to be the same in general. Furthermore, the impulse responses will 
become zero if the system is stable, while the timing effect will make a difference.
The impulse responses begin with the first period of a particular shock in relation to the 
first variable, and continue as a casual chain. Furthermore, the shocks sizes are usually 
set to standard deviation of the independent variables in to understand its reaction of the 
impulse responses clearer.