Analytics

ARS utilizes a quantitative approach to risk management, and in the automotive finance space this centers around a thorough understanding of vehicle depreciation. Accelerated depreciation is the key driver of loss for the GAP and EWT products; lower vehicle values increase the frequency of claim in both cases, and for GAP vehicle values also impact claim severity.

ARS analytics takes a two tiered approach to understand and forecast vehicle depreciation:

Macroeconomic factors (supply, demand, employment, fiscal policy, etc.) are used to develop neural net models of the Used Vehicle Consumer Price Index (UV CPI), a key component of consumer spending. These models then form the basis for thousands of forecasts across a variety of simulated economic scenarios resulting in a distribution of probability weighted outcomes at the “market level”.
Additionally, third party market value data is used to derive base depreciation curves for all units marketed in the US and Canada since 2000. These base curves allow for the projection of unit specific wholesale values in a “neutral” macroeconomic environment, and are then adjusted by the projected macroeconomic factors, resulting in a distribution of probable unit values in the probable economic environment.

Beyond depreciation, consumer/driver behavior is the second key factor to understanding automotive finance risk. For example, drivers are not more or less likely to be involved in a collision based on the value of their vehicle. But whether that collision results in a total loss (GAP) is a direct result of the current market value of that unit vs. the outstanding loan balance. The ARS GAP model incorporates this data in order to understand and estimate the probable GAP loss for each car/month exposure in our portfolio based on the actual (or projected) market value and loan balance of each specific unit. Similarly, a lessee is not more or less likely to expose their vehicle to excess wear (EWT) based on the relationship between residual value and market value of the unit. But lessees are significantly more likely to return a unit at lease end if the market value is significantly below residual. Here again, and the ARS EWT model utilizes these ratios at the unit specific level to better estimate lease return rates and subsequent EWT losses.

These GAP and EWT risk models have been developed from a robust warehouse of historical data which aggregates enrollments and claims for more than 12MM contracts originated by ARS, and it’s precursor since 2000. This historical data provides a robust sample for complex building modeling and scenario testing. Access to this historical and forecasted data is facilitated through a user-friendly interface which enables quick and repeatable report building and insight development for both internal use and client guidance.