TFP shock in a NK-DSGE model with stocks

General DSGE Modeling Replication Attempts
#1

Hi there,

I am actually replicating the results of the paper by Challe and Giannitsarou (2012). Briefly, the model is similar to a standard Christiano et.al (2005) or Smets and Wouter (2007), except that stock prices are included. Capital is owned by firms and therefore they have another layer of optimization in addition to the standard optimization (optimal pricing scheme and minimizing cost). However, with a representative framework the set of equations are exactly the same where capital is owned by households and we add the stock market equations in an ad-hoc manner (Please let me know if I am wrong on this).

I have the following questions:

  1. I have attached my code. But the main issue I have is regarding the response of TFP shocks. Specifically, output decreases with a certain level of price stickiness. I am not sure on why this is happening. This should not be the case with a low Calvo probability. I am not sure if it has something to do with labor dynamics. Would be great to get some feedback or comment on where the code or I am wrong.

  2. Also, with no price stickiness, the model is not able solve for the steady state. Even though, I have analytically solved the steady state and is shown in the “parameters” block. I am not sure why this is the case either.

Thank you. Would highly appreciate if anyone could provide any help. Code attached.SP_CA.mod (4.2 KB)

#2

No idea about number 1 as it flies in the face of any intuition I can offer.

For number 2, just spitballing here but maybe there’s a unit root in one of the exogenous processes?

#3

I have solved the second issue. It was a silly typo from my end! But still having problem with 1.

#4

My guess would be a mistake somewhere in the mod-file. Are you able to exactly replicate the results of the original paper before changing parameters?

#5

Hi Johannes. I could replicate their results for the original paper. Since the paper’s main focus is on monetary policy shock they do not analyze TFP shocks. I will still double check my code. It just doesn’t fit my intuition of firms producing less when there is a TFP boom with reasonable parametrization.

#6

Hi Johannes. I believe that the impulse response functions looks ok and more intuitive to me. But I had a quick question about the response of inflation (wage and marginal costs) to government spending shocks: I see that the response of inflation hinges on two parameters: the inverse of Frisch elasticity and the variable for output used in the Taylor rule. In particular, using output growth leads to an increase in inflation while it decreases if I use output deviation from steady state. Any idea on the intuition?

Attached my corrected version of the code. Sorry it’s very messy.

Thank you!

SP_CA.mod (5.1 KB)

#7

It’s hard to tell without knowing the model, but a deviation from steady state is very different than a growth rate. In one case, interest rates are high if output is high, in the other case, it interest rates can be low if output is high, but already going back to steady state. So the interest rate response can be very different.

#8

Got it. Thank you very much.