• NEW! What Every Credit Union Should Have:

    COVID-19 Recession Sensitivity Report

  • A MUST WATCH Webinar:

    COVID-19 Stress Testing Update
    April 16, 2020

  • FOR IMMEDIATE RELEASE (2020-01-11)

    Minnesota Credit Union Network and
    Deep Future Analytics
    Announce Their Partnership

    FOR IMMEDIATE RELEASE: 11 January 2020

    Minnesota Credit Union Network and Deep Future Analytics Announce Their Partnership

    11 January 2020

    St. Paul, MN (January 11, 2020) – Minnesota Credit Union Network (MnCUN) and Deep Future Analytics (DFA) announced their partnership today. MnCUN will make available DFA’s Prescient Manager™ software to its membership. Prescient Manager™ is an easy-to-use, web-based credit risk forecasting and stress testing solution for credit unions and community banks. The software’s functionality includes:
    • Accurate, scenario-based, account-level FAS 5 ALLL and CECL forecasts including discounted cash flow functionality.
    • New loan pricing optimization leveraging the same cash flow model as for CECL.
    • Scenario-based loan valuations for purchases and sales of loan participations.

    Joseph Breeden, founder and CEO of Deep Future Analytics said, “We are excited to be partnering with MnCUN. We share a common vision that our accurate, scenario-based, account level cash flow models can create value across many functions in the FI. These solutions are integrated and coordinated in a way that a collection of independent models cannot be. MnCUN will be a great partner for bringing this capability to Minnesota credit unions.”

    “Deep Future Analytics will help best position Minnesota credit unions to manage and anticipate risk. The all-in-one software calculates the necessary lifetime loss forecasts for CECL, but also provides accurate and actionable information for portfolio management, account management, and loan pricing,” said John Ferstl, Chief Operations Officer for MnCUN.

    ABOUT DEEP FUTURE ANALYTICS
    Deep Future Analytics is a joint operational venture of Prescient Models, LLC and Nuvision CUSO Holdings, LLC, a CUSO operated by Nuvision FCU. Dr. Joe Breeden, founder of Prescient Models, brings more than 20 years of experience leading financial institutions through predictive financial modeling, allowing clients to achieve a real understanding of portfolio dynamics for retail lending. Nuvision FCU was founded nearly a century ago as the credit union of Douglas Aircraft, its values were forged in the factories and plants that made the region prosper. Now with assets well-over $2B, Nuvision is a multi-state Credit Union, with branches in Southern California, Arizona, Wyoming, Alaska and Washington.


    About the Minnesota Credit Union Network

    The Minnesota Credit Union Network is the statewide trade association that works to ensure the success, growth and vitality of Minnesota credit unions. With approximately $25 billion in assets, Minnesota credit unions are local, trusted financial cooperatives that serve more than 1.8 million members at nearly 400 branch locations around the state. As not-for-profit institutions, credit unions give back to the communities they serve. For more information, visit www.mncun.org.

    ###

    Media Contact:
    Charles Hoy, Director of Business Development
    Deep Future Analytics LLC
    choy@prescientmodels.com
    (505) 690-7195

  • Prepare for COVID-19

    View last month's webinar:
    Stress Testing for Pandemics

Mass Extinctions: Natural and Financial

Written by: Dr. Joseph Breeden | Posted on:

by Joseph L. Breeden, 27 Apr 2020, because the world can change in a day...

Natural Mass Extinctions Evolution is my favorite paradigm for explaining almost everything in our world. To a data scientist, evolution is just optimization via an ensemble of models over a time-varying fitness landscape. Paperclips evolved under the selection pressure of consumers to fill various niches, both functional and frivolous. Automobiles evolved to be safer, more efficient, and ever more specialized, but when oil prices or consumer interest shifts, the fitness landscape shifts and new winners emerge. Even religions evolve to satisfy the needs of their followers with some added twists to further propagate the religion.

Of course, evolution was first developed to explain biology. Although the mechanisms have been found to be more complex than originally thought (swapping genes between species, inheriting gene activation levels from one’s parents, retrovirus sequences in human DNA, etc.), the notion remains that the fittest survive, and if the environment changes, then those able to adapt the quickest will be the founders of the next generation.

Except when it doesn’t. A mass extinction event breaks evolution. Mass extinctions occur when change occurs so rapidly that a majority of the species do not have time to adapt to the new situation.

In the history of the earth, mass extinction events have been caused by extreme volcanism (Permian-Triassic), by asteroids or comets (Cretaceous–Paleogene), perhaps by not-so-distant supernovae (yet to come), and many other causes. These events are indiscriminate killers. Were all trilobites so unfit that they had to go extinct at the end of the Permian? (I quite like trilobites. It’s a shame they’re gone.) Were the dinosaurs all so unfit that they should go extinct at the end of the Cretacious, and can we really assume that all mammals were more fit than all dinosaurs (excepting those that we know as birds)?

In fact, no. Some quirk of mammals and birds allowed them to survive a disaster when the dinosaurs could not, but it does not say that dinosaurs were unfit. In fact, we have no reason to think that earth would not be ruled by neo-dinosaur megacities today if not for one errant asteroid. Birds like ravens and parrots are shockingly intelligent, and I would assert have a degree of consciousness, if such things could be measured on a continuum rather than a yes / no test that admits one winner.

Financial Mass Extinctions When a “normal” recession occurs, for example the 2001 recession, many economists recommend minimal interference, because it is seen as an opportunity for evolution to kill off the weak companies and for stronger, newer companies to emerge. The process can be painful for the human employees who make up these companies, so it’s fair for the government “of the people” to help them find work in new companies. Nevertheless, when future homes are 3-D printed and we no longer need brick layers, we have no reason to prop up construction companies that specialize in masonry work.

A financial mass extinction is different. Wars, famines, volcanic eruptions, and tidal waves, all bring financial mass extinctions. Sumatra, Indonesia in 2004 and Hurricane Katrina in Louisiana and Mississippi in 2005 were examples of natural disasters that caused financial mass extinctions. The Great Depression and the devastation of Europe in WWII were human-caused financial mass extinction events that were met with massive investments to restore the financial ecosystems.

Surviving Mass Extinctions The Old Testament story of Noah’s Ark is actually relevant to our current situation. Noah is reported to have built an ark to save a breeding pair of every type of animal in order to quickly restart the ecosystem after the waters receded. He couldn’t save all the animals, just representatives. However, he did not discriminate, because all of the species were interwoven in an ecosystem. Of course, practically, it would be impossible to fit breeding pairs of every species of animal onto an ark. The ark would sink under the weight of just the insects, but one assumes he chose which were most necessary.

This is the same task before our government. A natural disaster is threatening a mass extinction of our economy. The victims were randomly chosen from an evolutionary perspective. The flood waters of our disaster happened to hit certain kinds of businesses most. Restaurants are not less fit to survive than hardware stores. However, if all of those are left to go extinct, the whole financial ecosystem could collapse – a depression from which we only recover by evolving entirely new species (businesses). Rather than starting over this way, we are trying to keep the old ecosystem intact, but we lack the resources to save all the entities, so we are just trying to preserve enough species in each phylum to be able to restart once the waters recede.

Challenges of Managing a Crisis Noah did not have to deal with the problem of his decisions affecting the sea level. We do. A feedback loop exists between our balancing of disease response and business response. No previous natural disaster has had this feedback loop.

Unfortunately, this is a time-delayed feedback loop – one of the best ways to create a chaotic system and a situation where humans as a species are notoriously bad managers. In fact, psychological studies in test environments, research into disasters like Chernobyl, and research into the control of chaotic systems all highlight the importance of using models to make effective control decisions. Point-in-time observations are insufficient. You must have a model which predicts where the system is heading and act to modify that future state.

Current estimates are that the median time to die from coronavirus is 23 days. If our elected officials take actions based upon fatalities, they are reacting to the state of the world 3 weeks ago. The remedy is to test at a high enough level to ascertain the current state of the system, which means high levels of random testing. Then we can guess what is happening today, use models of where things are headed in the future, and take actions that will optimize the future outcomes.

How This Ends With sufficient data and effective models, can we control the system to prevent a financial mass extinction? Not exactly. Mathematically, this crisis is defined by exponential growth. Periods of exponential growth are both a blessing and a curse from a control perspective. In space flight, a gravitational assist maneuver involves a well-timed application of energy to take advantage of a perfect control point. In a pandemic, the perfect control point is the beginning, before it becomes a pandemic. Any subsequent control is exponentially more expensive until the disease has reached saturation and control no longer matters.

Given where we are today, we do not have enough resources to completely reset our financial system to a pre-disaster state. Some businesses will fail that were not unfit, but simply in the path of the storm. However, we may be able to limit the effects to a minor extinction event, like the Quaternary extinction when we lost the mammoths. (I wish we still had mammoths around, too.)

We can reasonably expect that long-term changes in consumer behavior will occur as a result of coronavirus fears. That shift in the environment and the gaps left by businesses that don’t survive will create an evolutionary opening for new or surviving businesses to adapt and propagate.

This is not meant to sound like the happy ending to a Hollywood disaster movie. Evolution is a messy process and the new companies of the new environment will not necessarily be better than what we would have had without the disaster. It is just a statement of where we are headed given the natural laws of the universe.

Also, I believe that putting things in the proper context of extinctions and evolution serves to provide a scientific basis for policy decisions being made now. Government controls are the opposite of natural evolution, but when evolution is broken, the value of government becomes clear.

Future of Lending In the lending industry, we will also need to carefully consider how we assess credit worthiness. Business owners or consumer borrowers who failed in this crisis should not be viewed as unfit (bad credit risks). Some may prosper because they were more nimble, more creative, or with better resources than their peers, but most who fail do not bear full responsibility. That a business missed government assistance, was not selected for survival by the government, is not a mark against that business. Having been the owner of a New York City restaurant cannot be viewed in the future as a bad risk. This will create a unique situation for lenders where recent performance is only a mild indicator of future performance. A 2019 bureau score or agency rating may be more predictive of future performance than a 2021 score.

For data scientists, these are unprecedented times. Our data will have the 2009 and possibly 2001 recessions, which were human-caused economic cycles from which we do seek to score fitness. The COVID-19 recession will generate data of a fundamentally different type. It will reveal more about how the hand of government seeks to prevent mass extinctions, but how should we consider this data when predicting the fitness of businesses and consumers? Will this become a time period we simply skip over – the lost year?

I am beginning to formulate modifications to our modeling to allow for both evolutionary and extinctionary events, but this question is broader than one analyst or one company. Across the industry, analysts will need to explore and test the possibilities. 2021 should be an interesting year in credit risk research. 2020 will be a year of model overrides and human judgment. Yes, we still need humans in lending. Humans may cause mass extinctions, but we are also our own best tool for stopping them.

Report this Published by Joseph BreedenStatus is reachable Joseph Breeden Chief Executive Officer at Prescient Models LLC Published • 17h 20 articles

Following Here are some recent musings. More philosophical than practical. I hope you find it interesting.


SCHEDULE A DISCOVERY CALL

Learn More About How DFA's Software Can Help Your Institution?


Recent Publications

All publications below were authored by DFA's own, Dr. Joseph Breeden

  • Image

    Sample Chart

    Reserves: All Loans vs. RE Loans

    Image

    CECL (Current Expected Credit Loss) is the new accounting standard for estimating loss reserves on loan portfolios. The CECL guidance provides a great amount of flexibility in which models are used and a range of other choices that may impact the calculations. This book provides details of a study on how to apply CECL to US mortgage data. It seeks to disclose as many modeling details, results, and validation tests as possible so as to provide a reference for comparison and best practices. Because CECL is so similar to IFRS 9 Stage 2, this can also serve as a benchmark for implementing the new international account standards. The book is organized into three parts. Part I: Study Summary provides an overview of CECL, the design of the mortgage study, and the key comparative results across the models tested. Part II: Model Details provides in-depth discussions of how the models were designed and estimated, the coefficients, and the validation. Part III: Background provides additional conceptual material. Chapters 11 and 12 may be particularly useful to those new to modeling, and Chapter 13 puts CECL modeling in the context of lending analytics overall.

    • Date // May 2018
    Image

    Sample Chart

    Vintage Performance

    Image

    Building on the solid foundation of the previous bestselling first impression, this extended updated impression walks through the various issues of retail lending and develops approaches to address the interaction between economic cycles and retail lending. The complexity of time is extensively explored: vintages, current time and maturity. Reinventing Retail Lending Analytics, Second Impression covers complex issues such as scenario based forecasting, stress testing, volatility analysis, economic capital and portfolio optimisation, credit scoring and last, but not least, model risk.

    The book ends by providing examples of the application of nonlinear decomposition. These examples will provide you with rich data sets for exploring portfolio dynamics and improving portfolio management using nonlinear decomposition techniques.

    • Date // March 2019
    Image

    Sample Page

    Preface

    Image

    The new loan loss accounting rules for CECL and IFRS 9 require thousands of organizations to learn about modeling. Likewise, accountants and others in finance are now required to learn about statistical modeling concepts. This book is intended to define terms in a manner consistent with decades of academic literature on statistical modeling and hopefully reduce some of the noise and confusion just around definition of terms. It may also serve as a useful guide to analysts new to the field tasked with IFRS 9 compliance, the international loss accounting rules, and credit risk modeling in general.

    Each chapter of this book is a term that one might encounter when discussing creating lifetime loss forecasting models for CECL or IFRS 9. Not every term is a model, and some models listed are being mentioned only to explain why they are not likely to be used for loss forecasting. The CECL guidelines and subsequent FAQs have given examples of modeling techniques. Some people new to loss forecasting have assumed that those are all the available or applicable methods. This book is meant in
    part to dispel that misconception.

    The definitions and descriptions provided here are meant to provide an intuitive understanding across a range of modeling techniques. Mathematical derivations are kept to a minimum. The references listed will provide all the necessary details for an eager analyst.

    • Date // June 2018
© 2020 Deep Furture Analytics Contact Us