# [R-SIG-Finance] time series regression (demand for higher education)

Spencer Graves spencer.graves at pdf.com
Mon Apr 14 22:28:03 CEST 2008

```      1.  Did you not receive a reply yesterday from Matthieu Stigler,

For the analysis of multivariate time series use package vars for VAR
models and urca for VECM models, unit root and cointegration tests.
The author of these package wrote also a  book "analysis of integrated
and cointegrated time series with R" which can be usefull. See
http://pfaffikus.de/

posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.  If
you provide a self-contained example, you increase the pool of potential
respondents by a factor of 10 or 100, because even people who don't know
to the next step.

Hope this helps.
Spencer

bereket weldeslassie wrote:
> Dear All,
> I am doing a time series regression with one dependent time series variable,
> 7 independent time series variables and 32 annual observations in an attempt
> to model the demand for higher education. The dependent variable is
> Enrollment and the independent variables are like tuition, income and so on.
> The main purpose of my analysis is to investigate the impact of economic
> factors (like tuition and income) on enrollment. I am considering an error
> correction model of the form:
> diff(lnY(t))=a+b1*lnY(t-1)+b2*lnX(t-1)+b3*diff(lnX(t))+error
> to model the demand and solve the problem of cointegration, autocorrelation
> and multicollinearity. But this is not been able to solve all these
> problems. Is that a right way to estimate the elacticities?
> Any suggestion how to built a good model that solves these problems? I
> Thanks,
> Bereket
>
> 	[[alternative HTML version deleted]]
>
> _______________________________________________
> R-SIG-Finance at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-finance
> -- Subscriber-posting only.
> -- If you want to post, subscribe first.
>

```