This paper seeks to examine how the energy crisis impacts a Danish company in the plastic molding
industry through a cooperation agreement. The problem area is approached using a variety of
econometric and statistic modeling techniques to analyze the implications for the company’s revenue
and construct forecasts hereof.
Based on the fact that the company can change its product pool between every time step, the
paper has utilized proxies for its sales amount and price due to the heterogeneous structure of the
company.
A short- and long-term price elasticity is estimated to study the implications of the energy crisis on
the company’s revenue. The SVAR-model yields that a 1% increase in the firm’s sales price causes
the sales of plastic products to fall by 0,15% and 0,17%, respectively. The analytical results turn
out to be consistent with the hypothesis, which states the appearance of a relatively more elastic
estimate in the long run. Based on the inelastic results, the energy crisis has a positive impact on
the company’s revenue.
By considering the combination of RMSE and graphical illustration, the VAR-model provides the
best forecast of the company’s revenue. The forecast obtains a total average deviation of 9,7% from
the observed values, which is 0,5% point lower than the simpler univariate Prophet-model with
XGBoost Errors. Therefore, it can be discussed whether the company needs such complex models
for forecasting. Overall, the neural network models perform the worst, which might be due to the
simplistic nature of regression and the shorter length of the dataset.