Hi, could someone confirm if my understanding is correct.
Assume we are trying to explain P/E in terms of the Dividend Payout Ratio (DPO). We have these data for 10 companies across 10 years (annual intervals).
In a cross sectional regression, we are only concerned with a specific point in time. So we would regress the P/E ratios of all 10 companies against the DPOs of the 10 companies for a specific year only.
In a time series regression, we would only be interested in one company’s data. So we plot that company’s P/E ratio against its DPO for each of the 10 years.