Week 10: Portfolio Monitoring: Behavioral Signals Over Self-Report

Why an ongoing position is monitored on behaviour, not on periodic self-report, and which signals are cheap to fake versus costly to fake

26 April 2027

Self-report is the least reliable monitoring channel a relationship has, not because people lie often, but because self-report is cheap to produce and mostly reflects how someone wants to be seen answering, not the underlying state. “Everything’s fine” is nearly free to say and carries almost no information. The model instead monitors behavioural signals that are costly to fake because faking them costs something real, and it prices seven of them, not one.

Three are about who does the work: initiation rate (who starts the plans), disclosure rate (who volunteers information without being asked), and reciprocity balance (whether that effort stays roughly symmetric over a long window, not just this week). Two are about whether a signal can be trusted at all: its cost (did producing it consume real time, money or opportunity, or was it free to say) and its specificity (was it directed at you, or is it generic courtesy that would have gone to anyone). The last two are about durability: persistence (does warmth survive past the novelty and convenience period, or does it only show up when nothing else is competing for the same hour) and repair speed (how quickly a conflict gets addressed once it surfaces). All seven degrade quietly, well before self-report catches up, which is why “monitoring” is a separate skill from “checking in.”

Outline#

• why self-report is a weak monitoring channel, and what makes a signal costly to fake
• seven behavioural signals, in three groups: effort (initiation rate, disclosure rate, reciprocity balance), trustworthiness (cost, specificity), durability (persistence, repair speed)
• reading a signal trend, not a signal snapshot — one bad week is noise, four is a trend
• reading for next week: what to do once a monitored signal actually crosses a threshold