Digital Orientation Factor

This dataset contains monthly returns for the Digital Orientation (DO) factor, which captures the return premium associated with firms' digital orientation based on long-short portfolios. The factor is constructed by sorting common stocks listed on NYSE, AMEX, and NASDAQ into quintiles based on their digital orientation scores from the previous fiscal year, then taking long positions in the top quintile and short positions in the bottom quintile. Digital orientation is measured using a text-based bag-of-words approach applied to four dimensions of digital innovation (device, network, service, and content) based on the layered modular architecture by Yoo et al. (2010) and extracted from the MD&A section of firms' 10-K filings. Because a large proportion of stocks has a DO measure of zero, we group these stocks in the first quintile. The dataset spans July 1997 to May 2020 (275 months) and includes returns under two weighting schemes: equally weighted (EW) and value-weighted (VW), with the latter capped at the 80th percentile to limit large-firm influence. Portfolios are rebalanced annually each July, and monthly raw returns are reported without factor controls. Data sources include SEC EDGAR (10-K filings), Compustat, and CRSP.