Our Services

We build credit risk and loss forecasting model suites for banks and credit unions — from the first data conversation to production-ready output.

Our model development engagements start with the people who manage the portfolios. Before estimating a single parameter, we meet with line-of-business managers to understand the portfolio’s actual behavior, underwriting history, and risk characteristics. That meeting is standard practice on every Spero engagement. We have been told no other outside firm makes it standard.

Recent work includes:

• A complete loss forecasting model suite for every major consumer and wholesale portfolio at a $75 billion bank — built for both CECL and stress testing
• Custom CECL models for solar panel loans, home improvement loans, and refinanced student loans at a credit union exceeding $25 billion, where standard approaches and short sample histories made off-the-shelf methodologies unusable
• Risk rating model implementation in a leading third-party platform for a credit union exceeding $20 billion

Every development engagement includes an independent parallel verification process — a second set of calculations to confirm model outputs before anything goes to a board, an examiner, or a regulator.

From initial data infrastructure through model development, documentation, ongoing monitoring, and platform migration — we cover the full CECL lifecycle.

We have supported CECL implementations across a wide range of platforms and institution types, including full builds, platform migrations, and documentation-only engagements when that is all that is needed. We do not arrive with a predetermined methodology and force the data into it.

Our CECL work includes:

• End-to-end builds from data warehouse design through model suite development and parallel verification at institutions including a $75 billion bank
• Platform migrations from Sageworks to Moody’s ECL and from legacy systems to SAS ECL — full model rebuilds, not just documentation transfers
• Ongoing monitoring frameworks, backtesting, and audit-response support
• Regulatory documentation packages covering policy, ACL process narrative, methodology summaries, and production control workbooks
• Bespoke frameworks for run-off portfolios and non-standard lending products where commercial platforms provide no usable starting point

We also maintain a proprietary parallel loss-forecasting framework to independently verify vendor platform outputs — a check we apply on every engagement where a black-box platform is in use.

Independent model validation that answers two questions: Is the math right? Is it the right math?

We validate credit risk models, CECL methodologies, ALM/IRR models, and vendor platform outputs. Our validation work is fully documented to SR 11-7 standards and designed to hold up under examiner review — not just satisfy internal audit.

Our validation services include:

• Credit risk and loss forecasting model validation, including CECL methodology review and challenger model development
• ALM and IRR model validation, including QRM — covering NII, EVE, and rate sensitivity across retail portfolios, deposits, and the full balance sheet
• Independent vendor model challenge — we build parallel frameworks to verify what black-box platforms are actually computing
• Model risk function reviews and gap analysis against SR 11-7 expectations
• Findings documentation, remediation support, and regulatory examination preparation

We have established model risk management programs that received the highest possible rating in Federal Reserve horizontal reviews. We know what examiners look for because we have been on the other side of those reviews.

Validation and governance of ALM and IRR models — including QRM — at banks and credit unions where balance sheet risk is a first-order concern.

Our ALM and IRR work spans model validation, governance documentation, policy development, and stress testing. We have validated QRM models for institutions exceeding $25 billion in assets, confirming NII, EVE, and rate sensitivity calculations following major model updates.

Our ALM/IRR services include:

• QRM model validation — NII, EVE, and sensitivity analysis across retail portfolios and deposits in an IRRBB framework
• Validation of changes to index rate assumptions and model recalibrations
• Custom ALM policy development aligned to regulatory expectations
• IRR stress testing and scenario analysis
• Governance documentation and examination support

Interest rate risk sits at the intersection of model risk and regulatory scrutiny. Our work is designed to produce findings and documentation defensible to NCUA, the Fed, the OCC, and state examiners.

MRM function reviews, policy redesign, and governance frameworks built to SR 11-7 standards — at institutions that need to get their model risk house in order before the next exam.

We have built model risk management functions from the ground up and reviewed existing functions at institutions ranging from $25 billion to $75 billion in assets. Our work goes beyond gap analysis — we participate in implementation, including hiring.

Our MRM governance services include:

• SR 11-7 gap analysis via stakeholder interviews and documentation review
• Model inventory development and risk tiering
• Policy, procedure, and governance framework redesign
• Examination readiness assessment and regulatory communication support
• Direct participation in hiring senior MRM staff — an engagement that goes beyond consulting

One engagement highlight: while leading model risk management at Regions Bank, our team received the highest rating in the Federal Reserve’s 2013 horizontal review of model risk management across major U.S. financial institutions. We know what a well-run MRM function looks like because we built one under maximum regulatory scrutiny.

Custom training for boards, audit committees, ALCO members, and modeling staff — built on the same first-principles approach we use in our consulting work.

We teach what we build. Our training is designed by practitioners who have built and validated models at CCAR institutions, not by instructional designers working from textbooks. Every session is calibrated to the audience — a board session on CECL governance is structured very differently from a technical workshop for a modeling team.

Our training offerings include:

• Board and audit committee education on model risk governance, CECL, ALM/IRR, and AI risk in financial services
• ALCO-level workshops on interest rate risk governance and model oversight
• SR 11-7 training for model developers, validators, and MRM staff
• CECL methodology and governance training for risk and finance teams
• AI risk training for banks and credit unions — including what AI risk actually means for financial institutions versus the broader hype

Training can be delivered as live sessions, custom workshops, or eLearning modules. We can also design role-specific programs for new MRM hires or validation staff.

Executive search and professional coaching for model risk, credit risk, and quantitative finance functions — including direct support in hiring senior MRM staff.

Finding genuinely qualified model risk and credit risk talent is hard. The pool of people who can build, validate, and govern models at the standards regulators expect is small, and most of them are not actively looking. We know this market from the inside.

Our talent and coaching services include:

• Executive search for MRM managers, credit risk model developers, and quantitative analysts
• Direct hiring support — we have participated in the interview and selection process for MRM managers at institutions exceeding $35 billion in assets
• Professional coaching for risk leaders navigating regulatory pressure, team building, or institutional change
• Capability assessment — evaluating whether an existing team has the skills and structure to meet regulatory expectations
• Succession planning and leadership development for model risk and credit risk functions

Production platforms and benchmark tools for CECL and loss forecasting — with full visibility into methodology, automated workflows, and no black boxes.

We believe opacity is a model risk. When an institution cannot explain what its loss forecasting platform is computing, that is not just a validation problem — it is an examination problem. Our platform work is designed to give institutions control over their own models.

Our platform and automation services include:

• Proprietary parallel loss-forecasting frameworks that independently verify vendor platform outputs — a benchmark check we apply on every engagement where a black-box tool is in use
• Snowflake migration and consolidation — we migrated and automated CECL analytical workflows for a $35 billion credit union, consolidating multiple consumer portfolio processes into a single unified automated workflow
• Python-based process automation — eliminating manual data transfer steps between model outputs and reporting, reducing operational risk and accelerating the monthly CECL cycle
• Production platform development for CECL and loss forecasting, including data integration and ongoing monitoring automation
• Black-box challenger builds — custom parallel models that give institutions an independent benchmark against commercial platforms

Original research on credit and market risk questions — designed for institutions that need a credible, documented analytical answer, not a vendor pitch.

We approach research the same way we approach model development: start with the economic question, not the available data. Our principals have conducted this kind of work at major financial institutions — including investigating the relationship between oil prices and corporate bond defaults in the energy sector while leading strategic risk at Regions Bank — as well as independently for publication on sperorisk.com.

We use whatever data best answers the question. For publicly available work we draw on FDIC call report data, Fannie Mae and Freddie Mac loan-level datasets, FRED, and other government sources. For client engagements we use the institution’s own loan-level data, supplemented with public sources where sample histories are short. In one recent engagement, a bank with a growing ARM portfolio lacked sufficient loss history to build a reliable model; we used Freddie Mac data with appropriate calibration to supplement the internal history.

Recent published research:

• Fort Myers mortgage market: Modeled how mortgage losses would perform if home prices continued declining along the same path as the 2006–2011 crisis
• Hurricane effects on mortgage losses: Analysis of how hurricane events affect mortgage performance in affected regions
• Private credit loss records: Structural analysis of why private credit’s historical loss record may understate downside risk in a genuine credit downturn
• Unemployment and commercial loan losses: Research on the relationship between unemployment dynamics and C&I and CRE loss rates

Whether you need a quick analytical answer or a fully documented research study, we can scope the engagement to fit your timeline and budget.