Get Started
Category: Webinar Recording

Webinar: Machine Learning in Building a Prepayment Model

webinar

Machine Learning in Building a Prepayment Model

Join RiskSpan financial model experts Janet Jozwik, Fan Zhang, and Lei Zhao to discuss how machine learning can help simplify prepayment models. They will discuss

  • Data:  Preprocessing the data and determining which variables are important to include in prepayment models
  • Modeling Approach:  Evaluating machine learning approaches
  • Model Performance: Opening the black box and tuning the model to improve performance

About The Hosts

Janet Jozwik

Managing Director – RiskSpan

Janet Jozwik helps manage quantitative modeling and data analysis groups at RiskSpan. Janet has a background in mortgage credit modeling, loss forecasting, and data analysis. Since joining RiskSpan, Janet has focused on loss forecasting and mortgage portfolio analytics for a key client as well as building a credit model using GSE loan-level data. Prior to joining RiskSpan, Janet was a financial economist at Fannie Mae where she specialized in single family credit pricing. Her work directly impacted the national guarantee fee pricing scheme and government programs to support the housing market during and after the financial crisis. Janet has extensive experience in analyzing massive datasets, a deep understanding of the drivers of credit risk, and an expertise in modeling mortgage cash flows. Janet holds an MBA from the University Of Chicago Booth School Of Business and a BA in Economics from Johns Hopkins University. 

Fan Zhang

Director of Model Development

Fan Zhang has 12 years of quantitative finance experience specializing in behavioral modeling, fixed income analysis and, machine learning. At RiskSpan, Fan leads the quantitative modeling team where he is currently driving improvements to prepay modeling and application of cutting edge machine learning methods. Fan was a senior quantitative manager at Capital One where he worked on prepayment, deposit, MSR, auto, interest rate term structure, and economic capital modeling. He was also a senior financial engineer at Fannie Mae managing a team to validate model implementation and risk analytics. Fan holds an MBA from the University of Maryland and a BA in Economics from the University of Michigan.

Lei Zhao

Quantitative Modeling Analyst

Lei Zhao is a key member of the quantitative modeling team at RiskSpan. Lei has done extensive research on clustering methodologies and his postdoctoral research paper has been cited over a hundred times in scholarly publications. Lei holds a Master of Science degree in Financial Engineering from University of California, Los Angeles, and a PhD in Mechanical Engineering from Zhejiang University, China. 


Webinar: Managing Down Model Validation Costs

webinar

Managing Down Model Validation Costs

Learn how to make your model validation budget go further for you.  In this webinar, you’ll learn about:  Balancing internal and external resources, prioritizing models with the most risk, documenting to facilitate the process.


About The Hosts

Timothy Willis

Managing Director – RiskSpan

Timothy Willis is an experienced engagement manager, financial model validator and mortgage industry analyst who regularly authors and oversees the delivery of technical reports tailored to executive management and regulatory audiences. Tim has directed projects validating virtually every type of model used by banks. He has also developed business requirements and improved processes for commercial banks of all sizes, mortgage banks, mortgage servicers, Federal Home Loan Banks, rating agencies, Fannie Mae, Freddie Mac, and U.S. Government agencies.

Nick Young

Director of Model Risk Management

Nick Young has more than ten years of experience as a quantitative analyst and economist. At RiskSpan, he performs model validation, development and governance on a wide variety of models including those used for Basel capital planning, reserve/impairment, Asset Liability Management (ALM), CCAR/DFAST stress testing, credit origination, default, prepayment, market risk, Anti-Money Laundering (AML), fair lending, fraud and account management.


Webinar: Mortgage Insurance and CECL Presented by MGIC with RiskSpan

webinar

Mortgage Insurance and CECL – Presented by MGIC with RiskSpan

Mortgage insurance is typically purchased to protect mortgage investors from credit risk. Under the new “Current Expected Credit Loss” (CECL) accounting standard, mortgage insurance provides a secondary benefit: a lower allowance for credit losses.

This webinar will:

  • Quantify the impact of MI on CECL under a range of macroeconomic scenarios
  • Introduce a way of measuring MI “value” in a CECL context, namely, a premium-to-allowance reduction ratio
  • Under a mainstream set of macroeconomic assumptions, analyze various coverage levels to search for best value

Webinar: Practical Approaches for Debt Securities Accounting

webinar

Practical Approaches for Debt Securities Accounting

Join RiskSpan allowance expert David Andrukonis for lessons learned from early Current Expected Credit Loss standard (CECL) adopters. 2020 CECL adopters are ready for the new loan accounting, but many are scrambling to meet the new requirements for their HTM and AFS debt securities.

This session will give you:

  • Concrete, practical approaches to solve for HTM and AFS credit loss accounting – approaches that can still be implemented in time for the 2020 adoption deadline and parallel runs
  • CECL implementation experiences from small banks up to $150bn firms, with both 2020 and 2023 implementation dates
  • Solutions for all security types, across a range of budgets
  • Q&A with the host, David Andrukonis

About The Hosts

David Andrukonis, CFA

Managing Director

David Andrukonis has technical and managerial experience in banking, credit risk, and valuation. At RiskSpan, David performs non-traditional ABS valuations and has validated a wide range of financial forecasting models, including models that estimate return on equity, capital levels, asset/liability valuations, and loan losses.



Webinar: Using Machine Learning in Whole Loan Data Prep

webinar

Using Machine Learning in Whole Loan Data Prep

Tackle one of your biggest obstacles: Curating and normalizing multiple, disparate data sets.

Learn from RiskSpan experts:

  • How to leverage machine learning to help streamline whole loan data prep
  • Innovative ways to manage the differences in large data sets
  • How to automate ‘the boring stuff’

About The Hosts

LC Yarnelle

Director – RiskSpan

LC Yarnelle is a Director with experience in financial modeling, business operations, requirements gathering and process design. At RiskSpan, LC has worked on model validation and business process improvement/documentation projects. He also led the development of one of RiskSpan’s software offerings, and has led multiple development projects for clients, utilizing both Waterfall and Agile frameworks.  Prior to RiskSpan, LC was as an analyst at NVR Mortgage in the secondary marketing group in Reston, VA, where he was responsible for daily pricing, as well as on-going process improvement activities.  Before a career move into finance, LC was the director of operations and a minority owner of a small business in Fort Wayne, IN. He holds a BA from Wittenberg University, as well as an MBA from Ohio State University. 

Matt Steele

Senior Analyst – RiskSpan

LC Yarnelle is a Director with experience in financial modeling, business operations, requirements gathering and process design. At RiskSpan, LC has worked on model validation and business process improvement/documentation projects. He also led the development of one of RiskSpan’s software offerings, and has led multiple development projects for clients, utilizing both Waterfall and Agile frameworks.  Prior to RiskSpan, LC was as an analyst at NVR Mortgage in the secondary marketing group in Reston, VA, where he was responsible for daily pricing, as well as on-going process improvement activities.  Before a career move into finance, LC was the director of operations and a minority owner of a small business in Fort Wayne, IN. He holds a BA from Wittenberg University, as well as an MBA from Ohio State University. 


Webinar: Estimating Credit Losses in the COVID-19 Pandemic

webinar

Estimating Credit Losses in the COVID-19 Pandemic

Business-as-usual macroeconomic scenarios that seemed sensible a few months ago are now obviously incorrect. Off-the-shelf models likely need enhancements. How can institutions adapt? 

In this webinar, we will discuss:   

  • How to incorporate COVID-19-driven macroeconomic scenarios in an allowance model (for CECL and OTTI attributions)
  • Making necessary enhancements and tunings to credit and prepayment models
  • Model overrides and user-defined scenarios
  • Situations in which management “Q-factor” or qualitative adjustments may be called for


Webinar: Modeling Techniques for Hard-to-Value Bonds

webinar

Modeling Techniques for Hard-to-Value Bonds

Learn from leading practitioners as they discuss how to model bonds whose market values do not reflect their underlying fundamentals.

The market continues to punish every category of structured finance product. Even the highest-rated securities are not immune, but the further a bond moves down the ratings scale, the greater the uncertainty around what its real valuation is. 

Mark-to-model is fast becoming the new normal as the Covid-19 crisis is causing investors to become less and less comfortable relying on normal pricing service output. But transitioning from Level 1 to Level 2 (and even sometimes Level 3) assets brings with it a host of internal compliance and other challenges.  

Modelers must be able to demonstrate that their assumptions are defensible and their techniques are sound. 

Join Bill Moretti, Scott Carnahan, and Joe Sturtevant as they discuss “Modeling Techniques for Hard-to-Value Bonds” 

Key Topics:  

  • Overview of recent cross-sector performance 
  • Considerations when having to adapt from a market-based approach to a model-based one 
  • Example illustration of how to value a CLO security using mark-to-model. 


Get Started