Accounting requires the ability to forecast conduit administration fees. a simple stationary lognormal model for the fees is presented. Initially, the stationarity of the sweep fees is tested by measuring the level of mean reversion. Using a Dickey-Fuller statistical test the conduits are checked for approximate stationarity. Next, assuming the raw fee data is stationary, the proposed lognormal form of the model is investigated. A Jarque-Bera statistical test formalizes this step.