Bond issues

[jameszhow1985]

Dear all

I am building a two-country DSGE model in which domestic household can hold home bond issued by goverment and U.S. bond, but U.S household can only hold its own bond which is in zero net supply. when I run my model, dynare report BK conditon is satisfied, but the rank condition isn't verified. I have checked the timing of the model over and over again, but still can not fix this problem. can someone tell me what's is it ? Thank you guys.

[jpfeifer]

This sounds not like a timing problem, but rather one where one variable cannot be determined from within the model. Did you impose that the US bond in zero net supply by dropping this variable from the model? Because usually there is not market clearing condition for this one, leading to a problem. If that is not the source of the problem, you need to start looking for a missing equation that needs to replace a redundant equation still in the model. With two-country models, this is always tricky.

[jameszhow1985]

Hi, professor Pfeifer

It runs when I follow your sugggestion. I did not fix this problem in the past few days by myself, but under your suggestion I fix the problem in a short time. I am so excited.

By the way, why dropping the bond clearing condition can cause the rank failure?

[jpfeifer]

That is hard to say without knowing the model. But it seems you were instead having a separate equation that was redundant.

[jameszhow1985]

That is hard to say without knowing the model. But it seems you were instead having a separate equation that was redundant.

Hi,Professor Pfeifer
I want to analyse the stabilising effects of a model to some variables, such as output gap and trade balance deficit, under different policy settings. Under each scenario, I keep the steady state of the model constant. What should I do in order to accomplish the exercise? May I run my mod file under different policy setting each time, and save the irfs. Then I plot the irfs for different policy setting. Is that correct?

[jpfeifer]

That sounds correct.

[jameszhow1985]

That sounds correct.

Hi, Professor Pfeifer

I have a question that can I assume the goods price of the home country or the foreign country equal to unity in the steady state file? As doing this can solve the steady state of the model easily. I can solve the prices endogenously in my prior steady state file, but some steady state values of variables of interest are not fit with literature. Does assuming the price cause indeterminacy?

[jpfeifer]

Sorry, but I don't know your model, so I am afraid I won't be able to help you here. Only the economic intuition behind your model can tell you whether you are allowed to normalize the prices.

[jameszhow1985]

Sorry, but I don't know your model, so I am afraid I won't be able to help you here. Only the economic intuition behind your model can tell you whether you are allowed to normalize the prices.

Hi, Professor Pfeifer

I have a question that what should I do if two countries(e.g China and U.S) have significantly different growth rate of output in a two-country DSGE model ? I read some literature in which authors treat both growth rate of technology as the same, so they have the same growth rate of output in the steady state , thus the two-country system is stable. But if they have different technological progress rate, the country with lower progress rate of technology will vanish eventually. Back to my question, how should I treat this problem?

[jpfeifer]

That is hard to say. What you have in mind when you say one country is going to evenutually vanis is a different steady state growth rate. But coming from the Solow model, people would argue that the long-run growth rate will be the same and that the short-run difference in the growth rate is mostly due to physical and human capital accumulation.

For your model, it depends on whether you want to estimate or just simulate the model. In case of a simulated model, you could explicitly model this catch up behavior by choosing a starting value away from the balanced growth path. In case of an estimated model, things are trickier. Here you need to decide whether you want to have shocks explain the different growth rates (see my remark on imposing cointegration relationships in my Guide to Observation Equations).

In a nutshell: you as the model builder need to take a stand on where the differences in growth rates come from and whether you want to add features to the model that are capable of explaining them. If not, you might want to abstract from the differing growth rates by working with detrended data.

[jameszhow1985]

Thanks for your reply. Really appreciate your help. I will try the approach you suggest.