In 1997, France Telecom went through a partial privatization. Using a database that tracks over 200,000 eligible participants, we analyze employees’ decisions whether to participate; how much to invest; and what stock alternatives to select. The results are broadly consistent with a neoclassical model of investing behavior. We report four anomalous findings: (1) The firm specificity of human capital has a negligible effect on employees’ investment decisions; (2) the amount invested seems driven by different forces than the decision to participate, and we attempt to measure an apparent “threshold effect”; (3) employees “left on the table” benefits worth one to two months’ salary by failing to participate; and (4) most participants underweighted the most valuable asset.

Cross-holdings are created when a shareholder of one firm holds shares in other firms as well, and cross-holdings alter shareholder preferences over corporate decisions that affect those other firms. Prior evidence suggests that such cross-holdings explain the puzzle of why shareholders allow acquisitions that reduce the value of the bidder. Conducting a shareholder-level analysis of cross-holdings, we instead find that cross-holdings are too small to matter in most acquisitions and that bidders do not bid more aggressively even in the few cases in which cross-holdings are large. We conclude that cross-holdings do not explain value-reducing acquisitions. Beyond acquisitions, we find that institutional cross-holdings between large firms have, in fact, increased rapidly over the last 20 years, but mostly due to indexing and quasi-indexing. As in acquisitions, cross-holdings by active investors are typically too small to matter.