Search code examples
time-seriesregressionlinear-regressionforecastingdummy-variable

Linear Regression with time variant independent variables

In this model I would be trying to predict HDI of a country.

Model looks something like this: HDI(t) ~ Life Expectancy(t) + GNI (t)

However life expectancy and gni themselves are time variant independent variables. Is there anyway to account for this in linear regression such that i can predict

HDI(2018) using data from years before 2017? Is there anyway for linear regression to capture this?

Would be creating dummy variables like country be of any help?

Solution

  • As you mentioned HDI(t) is depending on Life Expectancy and GNI at the same time t so if you don't have their values in t it would be a guess. By the way what are you looking for something like a and b in HDI = a * Life Expectancy + b * GNI?

    You can assume a and b are time invariant and solve the problem for 2017 (or for several years and calculate mean) then you'll have them :)