1
$\begingroup$

I tried looking this question up on google and didn't find material that answered my question. But my questions are:

(1) Is there a method to determine how long it takes a leading indicator to affect a variable ? So if we are looking at the affects of oil production on sales, when oil drops how long does it take to affect sales.

Could I use survival analysis for this? This seems related but in a biological context

(2) Can we measure the degree to which oil production affects sales? If oil production drops by 10% it affects sales by 17%.

(3) What's the best way to determine the most important leading indicator? Univariate regression and compare models?

(4) Is there a package in R that could be used for this?

$\endgroup$

2 Answers 2

0
$\begingroup$
  1. Go to Autobox.com and review the discussion about transfer functions.
  2. The is a question about elasticities. Review any principles of economics text. A new version of Autobox calculates elasticities.
  3. If your data is time series, it is inappropriate to use classical regression analysis. Applying classical regression analysis to summarize time series data will result in a violation of one or more of the underlying assumptions about the error terms. Use time series techniques to summarize time series data, and classical regression to summarize cross sectional data.
  4. If you focus on R, then you will spend an inordinate amount to time learning how to program rather than on analysis. Investigate the advantages of Autobox.
$\endgroup$
1
  • 2
    $\begingroup$ What model does Autobox use to determine price elasticity for time series data? $\endgroup$
    – DataTx
    Jun 15, 2016 at 19:57
0
$\begingroup$

(1)To find lead/lag indicator time Cross Correlation Function can be used in R with the ccf command.

(2)To find the effect of the indicator $x_{(x+h)}$ on $y_{x}$ you can regress on the lag function and use the coefficient $\beta_{1}$ in the regression as the magnitude. As found here

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.