Lag structure

As common for longitudinal analyses we assume that the independent variable affects the dependent variable immediately and with a certain delay. This relation is included in our model through a lag structure where we compare the respective time values of the DV with several different time values of the IV. This approach helps in this case to account for the time-dependent influence of policy measures on investor behaviour (Angrist & Pischke, 2008; Wooldridge et al., 2009). We decided to introduce a lag structure with a lag of the policy effectiveness within zero to three years (i.e. investments in capacity in year t are influenced by policy measures in years t,t-i,t-2,t-3). This means policies could have a direct effect on investor decisions or it could take up to three years for a policy to trigger a capacity addition. On the one hand, it is possible that investors anticipate the regulation and already have their projects ready when it is passed, as the regulatory process is mostly open. On the other hand, there are factors delaying the investment process such as the time needed to build the wind farm or solar park and to gain access to the grid. In general, changes in regulations evolve throughout time and are communicated prior to them being passed and set active.

< Prev   CONTENTS   Source   Next >