Products
Consulting
Built from your loans.
Not industry averages.
Banks and credit unions trust ARCSys to build their CECL, ALM, and Analytics from their own loan-level data, so when a number moves, they can see exactly why.
Most CECL and ALM platforms start with third-party industry data. ARCSys starts with your loans.
Pooled loss history, industry prepayment speeds, balances by category: most platforms start there and fit your institution to it. It's a reasonable shortcut. But when your number moves, the reason is buried in someone else's data.
ARCSys starts with every loan you hold, reconciled to your general ledger every month. Your models are built from how your own borrowers have actually behaved. Where your history is thin, outside data can fill the gap, and you'll see exactly where it entered. It never becomes the foundation.
Built from your loans, not industry averages. Here is what that changes.
With third-party data, the forecast is built once, and every institution gets fitted to it. With your loan-level data, we forecast every single one of your pools.
When CECL arrived, the industry took four reasonable shortcuts.
None of them is wrong. Each one made a platform faster to build and easier to roll out. And each one moved your model a step further from your own loans.
Pooled loss history.
Fast to stand up. Built on someone else's borrowers.
It gets you a compliant number on day one. But the default and recovery rates behind it come partly from other institutions, so when that history shifts, your allowance shifts with it, and you can't see why.
Industry prepayment speeds.
Easy to plug in. Averaged across everyone.
Prepayment decides how long a loan is exposed to loss and how your balance sheet reprices. When the speeds come from an industry curve, your life-of-loan estimate and your rate-risk results rest on how the average borrower behaves, not how yours do.
Balances by category.
Light to load. Blind to the loan.
Loading summary balances is quicker. But a category total can't tell you which loans moved the number, so every variance becomes an investigation.
Separate data for CECL and ALM.
Each sound on its own. Together, two balance sheets.
When credit risk and rate risk are modeled from different data, with different assumptions about the same loans, ALCO and the audit committee end up looking at two pictures that don't reconcile.
ARCSys took none of these shortcuts. That's the whole difference, and it's why everything below is possible.
Every number names the loans behind it.
When your allowance moves, ARCSys shows you which portfolio, which pool, which loans, and which assumption moved it, in plain English.
As rates rose, borrowers prepaid more slowly, so loans stay on the books longer and expected lifetime losses rise.
Three loans were downgraded after their annual reviews.
The unemployment path was revised upward.
Losses really did fall. You can see why the allowance rose anyway, because prepayment is measured on your own loans.
Two sets of data give you two balance sheets. ARCSys gives you one.
Because it's one record, everything connects. Your CECL cash flows become your ALM model, so your rate risk starts with how your borrowers have actually prepaid, and your deposits and investments sit in the same model. When ALCO asks how your mortgages will prepay if rates fall, the answer starts with how your mortgages have prepaid, by pool and by loan type, over time. Assumptions stay consistent and documented across both models.
Load your loans once. Everything after that is already connected.
The same loans feed ALCO, the board report, your disclosures, and a data warehouse you can query yourself, included in your fee rather than sold as an add-on. That's why some institutions now run whole teams, not just accounting, from their ARCSys data.
The same record of your institution is extending into loan participations, consumer-credit decisions, and early portfolio risk signals as each becomes available. Each one arrives on data you've already loaded, not as a new data project for your team.
every loanRECONCILED TO YOUR GL
EVERY MONTH
every loanRECONCILED TO YOUR GL
EVERY MONTH
- ARC-CECLAVAILABLEYour allowance, built from your own loans. Calculate, investigate, report, and defend expected credit losses.
- ARC-CECLReviewAVAILABLEYour model, reviewed every year by the people who know it.
- ARC-ALMAVAILABLERate risk across your loans, deposits, and investments, starting from the same loans as your allowance.
- ARCAnalyticsAVAILABLEEvery report, from the same reconciled loans. A queryable warehouse, included.
- ARCPeerAnalyticsAVAILABLECustom peer groups and a CAMELS review from public call report data, ready for the board and exam prep.
Building from every loan is harder. That part is our job.
Starting from your loans takes more work than plugging in an industry curve, and we do that work. First we learn what you hold, how your teams use the data, which questions you need to answer, and how much history is available. You send your loan files. We load, format, and reconcile years of history, build your pools, forecast every one of them, run every method on your loans, and sit with your team through the first close. What reaches your desk is the answer and the reason behind it.
When someone asks why, you talk to the people who built your model: accountants, data scientists, and statisticians working inside the same data you are.
Three things.
- 01Tell us what you hold and which questions you need to answer.
- 02Send your loan files.
- 03Sit with us through the first close.
Everything between your files and your number.
- 01Learn what you hold, how your teams use the data, and how much history is available.
- 02Load, format, and reconcile years of history to your general ledger.
- 03Build your pools and forecast every one of them.
- 04Run every method on your loans, side by side.
- 05Measure each loan's prepayment, every month.
- 06Sit with your team through the first close.
- 07Write down the answer and the reason behind it.
- 08Pick up the phone when a validator, auditor, or board member asks why.
Four questions worth asking any CECL or ALM provider.
Many institutions have never had a direct answer. Ask us the same four, and we'll show you on your own loans.
▸ Ask us on your loans- 01Where does the loss history in our model come from, and how much of it is ours?
- 02Whose prepayment speeds are in our CECL and ALM models?
- 03Can you show me the specific loans that moved last quarter's allowance?
- 04What in the system was built around our institution, rather than configured after the fact?
What institutions tell us after they switch: they can finally see their data.
The path often runs in three steps, and it starts before anyone switches anything.
It often starts with an independent validation.
Some institutions first meet ARCSys through an independent CECL or ALM validation of the model they already run, on any platform. The findings show what the number rests on and where the history behind it comes from.
Then the model is rebuilt from their own loans.
When an institution moves to ARCSys, we load and reconcile its history, build its pools, and forecast every one of them. The first close happens with our team beside yours, and the first variance comes with the loans that moved it.
Then more of the institution runs from the same record.
Once the loan record is in place, ALCO, the board report, the disclosures, and the data warehouse draw from the same loans. Over time, more of the institution works from that one record.
See how it fits your institution.
A 30-minute walkthrough on the questions you’re carrying today, with a specialist who works inside the same loan-level data you do.
Your institution isn’t average. Your models shouldn’t be either.