ARCSys Model Analysis back-tests actual results against model predictions, benchmarks against peers, and runs sensitivity analysis to show your models hold up over time. The same framework drives stress testing, running your portfolio through adverse economic scenarios to quantify the capital and earnings impact before examiners ask.
Back-tests and stress tests run on the same loans, forecasts, and pools as your allowance, so stress results agree with the numbers your board already sees.
Model performance should be monitored continuously, not reconstructed in a scramble before the exam.
ARCSys makes model monitoring routine, and reuses the forecasts and segmentation already built for your allowance so stress results stay consistent with the CECL and ALM numbers your board already sees.
Compare your allowance, loss rates, and assumptions against institutions of similar size and portfolio mix.
Measure forecast loss and prepayment rates against actual results across short- and long-term horizons.
Quantify how the allowance moves when key assumptions, forecasts, or prepayment speeds shift.
Run multiple economic scenarios side by side, from baseline through severe recession.
Translate adverse-scenario losses into capital and earnings impact for board and ALCO reporting.
Documented, traceable output for examiners and auditors, generated on demand every quarter.
Board-ready summaries and examiner-ready detail from the same run, with consistent assumptions, severity you define, and defensible documentation.
Forecast loss and prepayment rates measured against actuals.
Allowance, loss rates, and assumptions compared with similar institutions.
How the allowance moves when key assumptions shift.
Routine monitoring reports instead of a pre-exam scramble.
Stress scenarios draw on the same economic covariates and segmentation as your CECL and ALM models.
Rate shocks, localized CRE stress, prolonged unemployment, and the scenarios that matter to your footprint.
Adverse-scenario losses translated into capital and earnings impact.
Every input, assumption, and calculation traceable under examiner scrutiny.
Stress testing reuses the forecasts and segmentation already built for your allowance, so adverse-scenario results stay consistent with the CECL and ALM numbers your board already sees. One model, multiple audiences: board-ready summaries and examiner-ready detail from the same run.
▸ Request a scope callMonitoring that is routine, with the analysis ready when the question arrives.
Adverse scenarios quantified on the same basis as the allowance and ALM.
Board-ready output that reconciles to the examiner-ready detail.
Bring last quarter's back-test, or the scenario your board keeps asking about.
