important to remember that regulators may believe that the bank
received direct financial benefit from the unfair application of its
discretion in establishing the buy rate, and that such discretion
could be indicative of disparate treatment.
The second component of contract pricing is the dealer reserve
or “dealer markup.” 9 Banks and finance companies have historically
allowed the dealership to establish the contract rate at a level above
the buy rate. The difference determines the dealer reserve. While
most banks and finance companies place limits and other controls
on how much the contract rate may exceed the buy rate, the CFPB
has identified this as a lending policy that creates “significant risk.” 10
The CFPB has been clear in its bulletins and public statements—
a robust compliance monitoring program includes monitoring
and analysis of dealer reserve at the portfolio and dealership
levels. It is challenging to reach accurate conclusions regarding
potential pricing disparities based on dealership discretion. Much
of the challenge results from the structure of indirect automotive
transactions (see sidebar on page 13). In particular, attempting
to isolate a single pricing component from a transaction where
multiple prices are concurrently established raises difficult economic and analytical questions. These challenges are exacerbated
by limitations when using proxies.11 Additionally, the CFPB views
dealer reserve as a non-risk–based price. This has two important
implications. First, it mandates that dealer reserve be analyzed
separately from the buy rate. Second, it suggests the CFPB may
not find convincing an analysis of dealer reserve that includes
controls for the buyer’s credit worthiness. This view may persist
even when dealerships are exposed to charge backs of the dealer
reserve due to credit and prepayment risk for some number of
months after contract origination and, in more limited cases, for
the life of the contract. For the compliance officer, these issues
greatly increase the fair lending risk associated with dealer reserve.
Quantifying meaningful differences in dealer reserve across the
portfolio and at specific dealerships requires several analytical steps.
However, before any analyses can be undertaken, one must decide
how to measure dealer reserve. Basis points (bps) at origination and
dollars paid to the dealership are commonly used, but other options
may be used. This fundamental issue becomes complicated when
the bank allowed the dealership to choose among different payment
plans or when some dealerships are non-recourse, while others are
not. 12 At the portfolio level, the next step is to understand the average
difference in dealer reserves faced by each prohibited basis group
(or “raw” differences as they reflect no controls). Generally, two
raw metrics should be considered: the frequency of and the extent
(level) of the dealer reserve for each prohibited basis group. The
frequency metric identifies the percentage of contracts with a nonzero dealer reserve. The level metric measures the average amount
of dealer reserve. Just as in underwriting, the identification of raw
differences is not the end of the analysis, but rather the beginning.
The next step requires the analysis of portfolio level dealer
reserves while controlling for relevant factors. Most of the factors
will not be discernible from bank policies or practices. The dealer
reserve is a factor in arriving at the contract rate established by
the dealership and the vehicle buyer as part of the negotiation of
the overall transaction. Unlike the factors dictating underwriting
decisions or buy rates, the dealership’s supply function and the
buyer’s demand function are not to be found in the bank’s records.
Understanding these purchase-related factors that emerge from
the traditional economics of supply and demand is important
to the portfolio level analysis of dealer reserve. While there are
numerous unobservable dealership and buyer attributes that
contribute to the transactions (including the motivations of sup-
ply and demand), proxies for some attributes may be developed
from the bank’s transaction records. A regression analysis may
analyze the portfolio level dealer reserve and control for factors
that can be observed or inferred. To the extent that statistically
significant differences are identified, it is important to understand
that manual file reviews may be of limited value due to the general
lack of relevant information in the files.
Dealership-level analysis of dealer reserve is further compli-
cated by low contract volume at many dealerships from which
a bank purchases contracts. A common approach is to establish
a minimum volume level before including a dealership in the
analyses. For example, include only dealerships that assigned to
the bank 10 minority, 10 non-minority, and 30 total contracts
in a year. For each dealership that meets the 10/10/30 screen,
raw frequency and level metrics, as well as statistical tests, may
be calculated and compared across protected group contracts
at the dealership. More complex analyses, such as regressions,
are generally not possible at the dealership level due to the low
numbers of contracts assigned by a dealership.
Bank Participation in dealer Reserves
Finally, there is additional fair lending risk related to the dealer
reserve that has received limited attention recently—bank participation in the dealer reserve. Prior to extensive dealer reserve
litigation in the early 2000s, many indirect automotive finance
companies shared financially, or “participated,” in the dealer reserve. As these cases settled, most banks and finance companies
stopped participating in the dealer reserve. Paying the dealership
the full dealer reserve creates two potential fair lending risks—one
perceived and one actual. To understand these risks, one must
understand the mechanics of these payments to the dealership
and risks accepted by the dealership and bank.
When the dealership assigns the contract to a bank or finance
company and the contract rate exceeds the buy rate, the basis point
difference between the two rates can be calculated and converted
into dollars of dealer reserve. However, that dollar amount is based
on the assumption the contract will be repaid as agreed and will
not pay off early or default. Of course, many buyers refinance, sell
the vehicle, or experience a total loss in an accident prior to the
end of the contract. At the origination of the contract, neither
The CFPB has been clear in its bulletins
and public statements—a robust compliance
monitoring program includes monitoring
and analysis of dealer reserve at the
portfolio and dealership levels.
