I have intermediate expertise of R and limited knowledge in statistics. I run the following simplified (removed control variables) regression:

model <- feols(RET ~ EUA_change * CE |
datadate + country + industry, data = dataset, cluster = c(“gvkey”,”datadate”))

where RET is the daily stock return (%), EUA_change is the daily carbon price change (%), and CE is a continuous carbon exposure variable (%). As a result, I obtain a highly significant coefficient for the interaction term of -0.005. Because I control for the day as a fixed effect, the EUA_change variable itself is removed due to collinearity.

I would like to plot the impact for a 1% change in the EUA price for a continuum of carbon exposure levels, i.e. simply the line with a slope of -0.005 and coefficient of 0. Should I just do this manually, or can I mirror the attached image and plot this relationship with a confidence interval?

I tried so far various functions e.g. the plot_slopes function from the package marginaleffects:
plot_slopes(model, variables = ‘EUA_change’, condition = ‘CE’)

Or also the plot_predictions function. I am aware that is not solely a code issue, but more understanding the logic respectively the statistics.

Many thanks for any support! (First time stack overflow user)