Monte carlo simulations with one of the indexed pre-defined

[Eyal Rittberg]

Hi,
Good morning.
I came to this wonderful forum after watching MIT 18.S096 Topics in Mathematics with Applications in Finance, Fall 2013- lecture 7 VAR by Kenneth Abbott.

I am trying to model mortgage portfolio behavior in the Israeli market.
A mortgage portfolio is built from different loans. Each loan's monthly payment, interest and principal is dynamic and depends on different indicators which are:
Consumer pricing index, Israeli government bond rates etc..

Following Ken's lecture, I obtained a correlated synthesized spreadsheet using the time series data of the last 10 years.

Now, I want to check extreme scenarios. For example how the Israeli government bond rate will change from a steep incline of inflation rate.
Therefore, I want to simulate a synthesized time series with a predefined consumer index and get the correlated bond rate with respect to that.

Any Idea how this can be done?
From portfolio management point of view, is this the right thing to do (As I am trying to assess the risk 20 years from now to a mortgage given today )?

Thank you for your comments,
Eyal

 

[Ken Abbott]

You can build an increase into the outcomes, but it would be deterministic. I'm not sure the covariance structure gives you enough information about directional moves to divine anything useful.

 

[Eyal Rittberg]

Dear Mr Abbott,
Thank you so much for your response!

I decided to follow your insights and drop this drift issue.

Another question if I may:
The Israeli Benchmark Interest Rate is currently at 0.1 Percent .
I don't believe (based on news and current economic situation) that the interest rate will go any lower.
How would you recommend imposing a constraint on the monte carlo synthesized signal to prevent the base rate from going under while still maintaining the right value for the correlation matrix?

Is the (matlab code):
interest_rate = max(interest_rate,0.1)
a good idea?

Thank yoy very much
Eyal