I am currently working with a time series database, and I would like to study the effects of a sudden increase in the price of a good. I have some dates and some specific shocks ( an increase in 7% of the price, for example) . I do not know how to code this shocks into my model in order to determine if they where significant.
I would like to be able to regress:
Price_of_lemonade= Bo +B1Price_of_water + B2Price_of_lemons +B3shocks + e
Hello,
I may suggest you to create dummy variables for shocks, and then you can test for the significance. To create a dummy variable for shocks as you explained, you may try this codes. Good Luck.
sort price
gen shock=.
replace shock=1 if price>7
(price refers to a change of price in %)
Generate and add the dummies as Dilesha Rathnayake suggested. In a further step, you could also include time dummies (e.g. days or years, depending on your data structure) to form a period-fixed effects model. What you could also consider is to test your time series for one or multiple structural breaks after performing a regression (this can be done with estat sbknown; pay attention to the break() and breakvars()-options).
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Regression