Model for electricity consumption

[Ellizer Cabahug]

Hello!

I am conducting a study on the causal relationship between my explanatory variables (electricity consumption of the previous month, temperature, sunshine duration, purchasing power, and electricity price) and electricity consumption. The data I have is reported monthly between August 2009 to December 2017.

So far in my class we were only taught panel data analysis, therefore time series is very new to me. I don't know what model to employ after I've tested for the stationarity and integrated in the necessary order. I've tried to use the ARDL model since it works with variables integrated at l(0) and l(1). The results indicate that the explanatory variables are not statistically significant in explaining electricity consumption in the long run; I expected the result so it was not surprising. Given my explanatory variables, I think I can only explain the short run effects.

Will an OLS model be sufficient for me to achieve my goal?

Would really love your input on this

Best,
Ellizer Cabahug

[Phil Bromiley]

You didn't get a quick answer. You'll increase your chances of a useful answer by following the FAQ on asking questions - provide Stata code in code delimiters, readable Stata output, and sample data using dataex.

There is no general answer to your question. Different disciplines have markedly different preferences for statistical analysis. They also have very different perceptions of what is testing "a causal relation". You need to talk to your adviser.

[Marcos Almeida]

You presented electric consumption as Y and Xvar at the same. I assume it is just a mistype. That said, if you wish to spot time trends such as the ones you underlined, time-series analysis shall chime in fine. But this is not saying that forecast will forcefully be ‘significant’. A cautious note on ‘predictions’ in time-series analysis permeates the introductory texts, and this may come to the foreground. More so when it comes to ‘long term’ effects.