Statistic help need...urgent...

Posts 2 · Viewed 4311 · Go to Last Post

avii

· Snapshot

Like · Likedby · 0

Dear all

I am doing my dissertation on the effect of interest rate and inflation in Dow Jones and FTSE 100.

I am thinking to do in other way around ie effect of US interest & inflation rate on FTSE100 and effect of UK interest & inflation rate on Dow Jones Industrial average

Now my problem is I am little bit confuse with the statistic tools. Simple correlation analysis will show the relationship between these macroeconomic factors with closing stock indices

My concern is how I can do multiple regression analysis using these data. My data will be from 1999 jan to 2009 dec on monthly basis.

Hence anyone with knowledge on above topic and statistic analysis (multiple regression analysis) please share your view and suggestion.

Thanks in advance.

Gajedi

· Snapshot

Like · Liked by · 0

Is this really for your dissertation? I just can't believe it, dissertation is usually a part of your Ph.D. program and you are doing something like a multiple regression for your dissertation.

Anyway, here is what you should do:

1. Find some research articles that have provided some insights on your topic. Why did you choose interest rate and inflation for predicting ^DJI and ^FTSE. You can't just say that they are highly correlated and you assume that they can predict the variation in ^DJI or ^FTSE.

2. Check if you have high correlation between interest rate and inflation (remember Fisher's equation). If they are highly correlated then you have a problem of multicollinearity in your equation.

3. Now run equation for multiple linear regression using statistics tools such as eviews, R, etc.

EVIEWS (very expensive):

ls ftse c interest inflation

# Check for t-stat, R squared, etc.

R (freely available)

fit <- lm(y ~ x1 + x2 + x3, data=mydata)

summary(fit) # show results

layout(matrix(c(1,2,3,4),2,2))

plot(fit)

Last edited: 22-Mar-10 05:37 PM

Last edited: 22-Mar-10 05:38 PM

Please log in to reply to this post

You can also log in using your Facebook