I'm doing an econometrics paper analyzing the impacts of oil price shocks on GDP growth. In one of my models, I use change in nominal oil prices, a dummy variable representing negative oil price shocks, and an interaction term between the dummy variable and change in nominal oil prices. How would I interpret the coefficients on these values in an ARDL time series model? Specifically the coefficient on nominal oil prices is negative and statistically significant, the coefficient on the dummy is negative and not statistically significant, while the coefficient on the interaction term is positive and statistically significant.