surplus (risk reserve) process

Picture an insurer's bank account over time as a single running balance. It starts with some money set aside on day one. Then, steadily through the year, premiums flow in like a small stream raising the water level. But every so often a claim arrives and the level drops suddenly — a small dent for a fender-bender, a deep gash for a house fire. The surplus process is just the graph of that balance through time: a line that drifts gently upward between claims and jumps down each time the company has to pay one.

Precisely, write U(t) for the surplus at time t. The classical recipe is U(t) = u + c*t - S(t): you start with initial surplus u, you collect premium at a steady rate c per unit time so c*t has come in by time t, and S(t) is the total of all claims paid up to time t (a sum that sits flat and then jumps up at each claim). Subtracting claims from 'starting money plus premium received' gives the balance still standing. Between claims U(t) rises along a straight slanted line of slope c; at each claim it drops by the size of that claim. The whole point of risk theory is to study the shape and fate of this jagged line.

This matters because the surplus process is the single object that ties pricing to survival. If premiums are set too low, the upward drift is too gentle to recover from the downward jumps and the line tends to sink; if claims are unexpectedly large or bunched together, even a well-priced book can be driven below zero. Actuaries do not watch one real company's wiggly line — they study the probability distribution of all the lines that could happen, and ask how often such a line ever falls below zero. A common confusion: the real-world surplus also earns investment return and pays expenses and dividends; the clean U(t) = u + c*t - S(t) is a deliberately simplified model that strips those away to isolate the pure interplay of premium income and claims.