A workplace relationship policy is a short document that attempts to govern one of the oldest human practices using language designed by lawyers to be defensible in litigation. The typical policy is two pages. It defines what counts as a relationship—often "romantic or sexual," sometimes more elaborately. It identifies which relationships must be disclosed, usually those involving direct supervisory authority. It prescribes the remedy, usually reassignment of one party to break the reporting line. It states the company's reservation of rights to act in any case that creates conflict, embarrassment, or risk. It is signed at orientation, forgotten by lunch, and recalled only when something goes wrong.
The policies exist because the alternatives have failed. The mid-twentieth-century alternative was a tacit understanding: bosses dated secretaries, executives married assistants, and the costs were absorbed by the women involved through career truncation and by their colleagues through warped meritocracy. The late-twentieth-century alternative was strict prohibition: no romantic relationships at work, terminable for cause. Strict prohibition produced its own pathology—relationships went underground, disclosure never happened, the company learned of liaisons only when one party filed a complaint after the breakup. The contemporary disclosure-and-reassignment model is the third try, and it is the least bad option presently available, which is not the same as being good.
The policies do real work. The volume of consensual relationships at any large workplace is high; surveys consistently show that around a third of employees have dated a colleague at some point, and that a meaningful fraction of marriages began at work. A policy that addresses zero of these relationships is fiction. A policy that addresses all of them is unenforceable. The disclosure-and-reassignment design accepts that most peer relationships do not require company involvement and that all supervisory relationships do. The design works when the company can actually reassign people without career damage; it fails when the org chart is too small to allow reassignment, which is the situation in most startups and most departments below a certain size.
The romantic stakes are particular. The workplace is a place where people see each other in cognitive flow, demonstrate competence under pressure, and reveal character over months rather than minutes. This is precisely the setting in which durable attraction tends to form, which is why workplace relationships, when they survive, have famously low divorce rates compared to relationships originating in other contexts. A policy regime that recognizes this fact treats consensual peer dating as a normal phenomenon to be lightly governed. A regime that does not treat it this way pushes the dating underground, where the company has no visibility and the parties have no recourse.
The supervisory case is different and the policies are mostly right about it. A reporting relationship is, by definition, an asymmetric-power relationship. The subordinate's consent is structurally compromised even when subjectively given, because the subordinate cannot freely refuse without career risk. This is not a moral claim about either party. It is a structural observation that the consent in such a relationship has been pre-shaped by employment dependence. Catharine MacKinnon's argument in 1979 that workplace sexual interaction is sex discrimination because of this asymmetry was treated as radical for a decade and is now the implicit basis of every supervisory-relationship policy in the country.
The harder cases are the cases the policies do not cleanly address. Cross-departmental relationships where no formal reporting exists but informal influence does. Relationships between people of different seniority levels who do not report to each other but operate in the same political field. Relationships that begin as peer and become supervisory through promotion. Relationships between people who interact through clients, vendors, or board service. Each of these has a real risk profile and almost none of them are governed by the standard policy text. The policy assumes a clean org chart that does not exist.
The collective question is what HR policy is actually for. If it is for litigation defense, the current policies are roughly fit for purpose. If it is for fostering an environment where consensual adult life can occur without producing institutional damage, the current policies are an underinvestment. The companies that handle this best treat the policy as a starting point, supplemented by manager training that addresses the cases the policy does not cleanly cover, and by a culture in which disclosure is met with practical accommodation rather than punishment. The companies that handle it worst treat the policy as the entire program and discover periodically that the policy has not prevented any of the things it was supposed to prevent.