A0557
Title: Modelling the demand of photovoltaic panels using mixed-causal autoregression
Authors: Alain Hecq - Maastricht University (Netherlands)
Lenard Lieb - Maastricht University (Netherlands) [presenting]
Abstract: Renewable energies have been heavily subsidized over the last decade. In particular, subsidies for solar panels to enhance (private) demand have had a strong impact on the market, and triggered irregularities and phenomena such as increasing volatility and/or bubbles. We propose to use mixed-causal autoregressions to model the demand of solar panels in Belgium over the last decade. It has been shown previously that mixed-causal and noncausal models can parsimoniously capture some of those (nonlinear) features. We review the mostly theoretical literature on mixed-causal regressions and provide the applied researcher with a guideline for practical implementation for estimation and inference. We assess the finite sample properties of existing estimation strategies and suggest a robust alternative based on self-weighted least absolute deviation, allowing for the existence of infinite variance processes. A MATLAB toolkit is provided allowing the applied researcher to straightforwardly specify, estimate, and test mixed-causal autoregressive models.
