Category: Auto Finance

Applied AI for Auto Finance

Maybe it’s my “availability bias,” but AI seemed to be the theme of last week’s Auto Finance Summit.  There was one dedicated session while others, like residual risk and subprime credit, had AI in the background.

The exhibit hall featured the usual AI-based businesses: Upstart, Zest, SAS, et al.  In today’s post, I’ll summarize what I learned from the conference.

One panelist framed AI as “the same thing we’ve done with credit scores for the last thirty years.”  While technically incorrect – no one would describe a scorecard as AI – this framing has merit.  As with credit scoring, any AI model must be monitored for “drift,” and continually retrained.

Welcome to ML Ops

This brings us to MLOps or, as Experian calls it, Model Ops – and it’s not easy.  Experian reports that 55% of models never make it into production.  Their survey, which I have been excerpting on my Twitter feed, is filled with stats like this.

MLOps is like Dev Ops, only you have to version the data as well, and the code is guaranteed to rust.

Here is how I described MLOps to an engineering manager: Think of the work your team does to control code and manage a pipeline.  MLOps is like that, only you have to version the data as well, and the code is guaranteed to rust.

There are good no-code AI tools, like Google AutoML, which I wrote about here, SageMaker, and SAS Viya.  As an old-school Python coder, I was gratified to see that these tools are in the minority.

Not Just for Credit Scoring Anymore

The “same as credit scores” framing is instructive in another way, too.  This is where lenders first learned the power of predictive analytics, and began to build a capability that now includes targeted marketing, fraud detection, behavioral scoring, and more.

While generative AI is an exciting and rapidly advancing technology, the other applications of AI … continue to account for the majority of the overall potential value – McKinsey

There was a “gotcha” moment in one session, where the panelists had to admit they’re not doing anything with Generative AI.  But, why would they?  There are at least a dozen low-hanging use cases for AI classifiers and regressors, as I mentioned in What Is Real AI?

The AI/ML Development Journey

The most practical advice came from CUDL President Brian Hamilton.  Brian reminded the audience not to overlook Robotic Process Automation.  This was the only reference to RPA that I heard, and it got me thinking of the broader AI context.  A typical journey might look something like this:

  1. Data Wrangling
  2. Predictive Analytics
  3. Process Automation
  4. Data Engineering
  5. Machine Learning
  6. Model Ops
  7. Generative AI

Gen AI might appear in an automation role, as chatbot or search engine, but – certainly for financial services, which is swimming in metrics – the early use cases will be predictive.

While McKinsey estimates the impact of Gen AI around $3 trillion per year, this is on top of $11 trillion for non-gen AI and analytics.  The McKinsey study is here, and the Microsoft maturity framework is here.

Conferences like this are a great way to see what other people are doing with AI but, in the end, you must decide how best to deploy it toward your own business needs.  It’s a vehicle, not a destination.

Biweekly Payment Magic

A while back, I did some foundational work for a leading biweekly payment service.  That is, the math part, which I will reprise here.  Biweekly works best in a climate of high interest rates and, unfortunately, soon after this project, the Federal Reserve dropped their reference rate to zero.  The Fed has not been persistently above 2% until recently, and biweekly is once again looking good.

The featured chart shows a scenario first constructed by my erstwhile partner Phil Battista.  I call it the “magic trick” because the customer in this scenario has financed an extra $3,250 with no change to the term, APR, or payment.  Before presenting the trick, here are some basics about biweekly.

Biweekly Payment Plan Basics

In Canada, the banks offer loans with native biweekly payment schedules, and dealers feature them in their advertising.  Here in the States, you have to use a service.  The service collects payments biweekly via direct debit and manages the lender to accelerate the amortization.

Here is an example.  According to recent Cox data, the average price of a new car is now above $49,500 with an APR of 7.0% and a 72-month term.  By the way, this survey does not include luxury brands, and some people are financing up to 84 months.

Below, I have modeled this “average” loan showing monthly versus biweekly payment schedules.  This is showing the amortization only, omitting whatever fees the biweekly service may charge.  You can see that the loan is paid off seven months early.

If you’re using longer terms to fit customers into payments, biweekly will shorten the trade cycle a bit.  Also, credit-challenged buyers may be better off with direct debit synched to their paychecks.

Nostalgia Alert: coding for the U.S. Equity project was originally done in C# by my son, Paul, who would have been around fourteen at the time.  We were making an OO model to include all loan and lease instruments as subclasses.  Coding for this article was done by me, in Python, which is 10X easier.

The Magic Trick

If you compare the two charts above, you can see graphically how Phil’s trick works.  Instead of starting your biweekly loan at the same amount and having it end earlier, you start it higher and aim to end on the same date.

The trick works because half the monthly payment is higher than a native biweekly payment would be – by $33 in this example.  The customer makes the equivalent of thirteen monthly payments per year, and the bank loses a little bit of interest income.  Here are the steps:

  1. Increase the amount financed, which will increase the monthly payment.
  2. Increase the term until the monthly payment comes back down to where it was.
  3. Use the biweekly program to bring the term back down to where it was.

Congratulations, you can now finance more product with the same monthly payment.  I covered the concept for menu systems in Six Month Term Bump.  To do goal seeking, as I’ve shown here, you will need some Python (or a precocious teenager).

Lenders at Top of Funnel

Chase Auto recently rolled out a digital platform for car shopping … and financing.  I like it.  The link is here.  It seems that everyone today has a vehicle search page.  The original cast, Autotrader and Cars.com, with about a dozen TPC competitors, are now joined by OEM sites, public dealer groups, and marketplaces from Roadster and Carvana.

“More vehicle shoppers than ever have started to look for vehicle financing before ever setting foot in a dealership.”

Competition hinges on which information the customer will seek first.  In an era of reduced purchasing power, many customers will want to “secure financing before going to the dealer.”  That’s the prompt on the Chase website.  There’s a prequal button right there between the Lariat and the XLT.

Don’t take my word for it, though.  This J.D. Power study found that nearly half of all customers shop for financing before visiting a dealer – 62% among Gen Z – and they start more than 30 days out.

This is probably a negative development for captives, and indirect finance in general.  Banks have a lower cost of capital and better rates.  Chase, as you know, is also popular as an indirect lender.  They say there’s no conflict with their dealer channel, but what if they had to choose?

The reach hierarchy, by customer base, is:

  • Banks – eight digits (ongoing)
  • Car Makers – millions of cars per year
  • Dealer Groups – hundreds of thousands

Banks have more customers, by an order of magnitude, than even the largest car makers.  Ten years’ worth of loyal Toyota drivers doesn’t approach Bank of America’s 66 million customers.  The same ranking goes for website reach, with the banks getting 120 to 190 million visits per month, while Carvana, Ford, and Autotrader each get twenty something.

Capital One, by the way, also has a shopping platform.  Ally has a dealer locator.  Bank of America has a redirect to Dealertrack.  Capital One is pretty shrewd about encouraging buyers to bring the app with them into the dealership, so they can update the deal as needed.  Mobile-first responsive is good, but an app is better.  Bank customers will carry their bank’s app.

Captives have the home field advantage once the customer is in the dealership and, likewise, their position online is downstream from the OEM brand.  Captives are advised to be front and center on their manufacturer’s website.

Dealer groups, like AutoNation, must rely on their own brand to draw customer attention.  In terms of unit sales, even the largest dealer groups fall below tenth-ranked Subaru.  Note that Lithia chose to develop a new brand, Driveway, for their online business.

Of course, none of these is a direct measure of financing intent.  Only a fraction of online banking traffic is looking for an auto loan.  The point is that they’re looking for the loan first, and then the car.

Seasonal Adjustment Factors

I felt like doing something quantitative, so this week we look at seasonal adjustment factors.  Everybody always talks about SAAR, and you probably know that it stands for “seasonally adjusted annual rate,” but what does this really mean?

Well, suppose it’s March 2017, and you are wondering what total sales will be for the year.  The industry sold 1.55 million vehicles that month so, if you multiply by twelve months, you might estimate 18.6 million for the year.

You would be wrong, though, because March is always a strong month.  Here are the estimates produced by the simple “times twelve” method, relative to the actual total for 2017, which was 17.2 million.

Using data from Fred for the five years 2013 through 2017, and converting everything to a percentage, you can see how March always overestimates the year’s results.  Each year’s dots are a different color, though it doesn’t really matter which is which.

Some months are highly variable, like September.  Not a good gauge of anything.  Remember to distrust any SAAR figures published in September.  April, oddly, is a tight group and bang on the annual rate.  April 2018 sales were 1.4 million, so a good guess for the year is 16.8 million.

Taking an average across the five years, we find that March, May, August, and December each overshoot the annual rate by roughly 10%.  Finally, we convert these percentages into monthly adjustment factors.

Instead of multiplying last month’s sales by 12, multiply by the monthly factor to predict the year’s total.  Of course, we have more data than just a single month.  We can also look at cumulative sales since January.  For example, do a quarter of the year’s sales occur in the first quarter?

No.  It takes a while to make up for the weak January and February, and then the actual historical cumulative pace slowly comes into alignment with the idealized linear cumulative pace.  I made that chart, too, but it’s not pretty.  That’s enough quantitative stuff for this week.