SNC Program Modernization
Notice for Public Comment
1. To perform benchmark analysis and provide meaningful feedback to the
reporting banks, what data elements should the Agencies add, delete, or
change from the Expanded Reporter list?
Benchmark analyses could be augmented through the addition of company
related financial information, including but not limited to, net sales,
total assets, total liabilities, EBITDA and cash flow leverage calculations.
Information provided back to the banks should include the benchmark data and
statistical measures around the mean/ median benchmark data for PDs, LGDs
and ELs (e.g., standard deviation, etc.). However, contributors of data
should only be provided “like-data” in return (i.e., if expanded data is not
provided by an institution, expanded data should not be provided in return)
2. What are the effects on Expanded Reporters of providing data on credit
participations? The ease of data compilation for participated credits is
dependent on the bank’s internal system of reporting and in many cases will
require, at least, minimal manual preparation.
a. Are there data elements that reporting banks would not be able to
compile electronically without manual intervention?
The capture of derivative type credit exposure and unique identifiers
assigned by the regulating agencies may require some manual manipulation. To
minimize the manual nature of assignment, particularly for agency assigned
unique identifiers, the Interagencies may want to consider the creation of a
secure web-based relational application for agents that would allow the
electronic submission of basic deal information, such as customer name,
facility identified by type and amount and participating bank group.
The agency application would automatically assign a unique identifier
(auto-number) to the customer, facilities and participating banks. If
feasible, it could also provide a DUNS number for the customer and
participating banks.
The agent and participating banks would have access to the website which
would provide the unique identifiers as well as allow the download of SNC
data for import by the agent into its internal reporting systems. The agent
and participants could then import the data, at least for customer and bank,
into their internal systems. Unique identifiers assigned by the agencies for
facilities may have to be entered manually.
b. Are there equivalent data elements that would be easier to provide?
The agencies may want to consider the incorporation of the DUNS number as
a unique means to identify customers and participating banks in the SNC
pool. The DUNS number could be matched to the DUNS number in the submitting
banks database or the DUNS number could be used as the unique identifier,
thereby eliminating the need to create auto-number unique identifiers for
the customer and bank names.
3. For Basic Reporters, the Agencies anticipate that the effects of the
proposal will be minimal. What effects, if any, do reporting banks see from
the proposed changes?
The proposed changes for expanded reporters should, at least, expedite the
year-end data submission process for the SNC portfolio; that is, the
year-end process should be the last quarterly submission. The elimination of
the current process which requires the population of a standalone SNC
database with year-end data would further minimize the manual manipulation
of data and therefore be a great improvement.
a. The main change for Basic Reporters is improved data software. Are
there changes to the current software that would be particularly helpful?
An electronic file with SIC/NAICS conversion records would be helpful
especially if it allowed for the querying of records by SIC description.
Also, a file that contains the participating bank universe with its unique
Regulator ID would be very useful. Further, an open-architecture structure
would make creating passports between bank legacy systems and SNC data
warehousing easier.
b. Which, if any, additional data items would be useful for the Agencies
to collect, either to improve their understanding of the underlying
transactions or to provide better feedback to the reporting banks?
Non-performance data could be useful such as charge-off amount to-date as
well as agency outlook for the credit. In addition, it would be useful to
have an additional file delivered to the agent banks that includes the
ratings results. This file should include the unique identifiers for the
proper mapping of customer, bank and facility records to the submitting
bank’s internal reporting system.
c. What, if any, effects would the use of a common set of detailed
definitions have on Basic Reporters? Are there other alternatives that could
achieve the goals of reducing ambiguity and automating the data collection
process?
In an effort to reduce ambiguity, it would be helpful to have examples
accompanying each definition. It might also be useful to expand the above
proposed website with a Q&A bulletin board related to SNC data submissions.
The information collected through the site could be used to fine tune the
process and identify problems with it from the submitting bank’s
perspective. It could also be a source of information for submitting banks.
4. Are the criteria "agents 100 or more facilities and is a mandatory or
opt-in Basel II bank" reasonable to separate Expanded Reporters from Basic
Reporters? If not, please provide an alternative.
Yes.
5. Since more banks are using credit derivatives to manage their
exposures, should the Program begin to collect data on credit derivatives in
order to provide benchmarking feedback? Should the data files include credit
derivative positions used to manage portfolio risk, along with the same risk
metrics used for loans and other credit exposures?
The collection of data for credit derivatives seems reasonable; however,
if data is not readily accessible it will have to be compiled manually which
could be cumbersome for larger institutions. Maybe the submission of data
should be required only for those institutions that can easily achieve
submission with a roll-in future requirement for all other institutions. A
reasonable timeframe to comply would eliminate the need to compile
information manually, but would encourage larger banks to consider the
integration of such data in future submissions.
6. Are there obstacles to the Agencies' proposal to assign, distribute,
and maintain Regulatory IDs and, if so, what are they?
The only obstacle and time consuming process is in the initial assignment
of the existing portfolio. Some type of website, as proposed in response 2a
could potentially reduce the number of obstacles.
a. Should the Agencies distribute Regulatory IDs directly to participants
instead of relying on the Expanded Reporter Agent banks to do so?
Yes, this would be very helpful in electronic form or via a website.
b. Should Basic Reporters also distribute Regulatory IDs to their
participants? Are credit participations held by Basic Reporters' numerous
enough to provide useful, relevant feedback?
This could be cumbersome for basic reporters who are submitting annually. A
basic reporter may not have the capability to report meaningfully on their
portfolio given internal systems’ limitations; therefore, it may not be
advisable or cost effective for the agencies to do so.
c. Are there existing or planned commercial systems that might help
uniquely identify facilities and borrowers in place of the process proposed
here?
The DUNS number could serve as a unique identifier for customers and
banks although there are relational issues with DUNs numbers as well.
d. Would quarterly batch submission and Regulatory ID feedback for
Expanded Reporters be preferable or would those banks prefer to request the
Regulatory IDs throughout the year as deals are completed?
It would be beneficial to have access to the universe of potential customer
and bank participants via a website interface to an agency application; then
the expanded reporters could incorporate such information into a relational
data set.
7. Which technologies would best support the reporting banks in
requesting Regulatory IDs?
Any technology that allows for the electronic transmission of data is a
necessity as well as a reporting facility that provides easy to read canned
and customized reports.
8. For both Basic and Expanded Reporters, the Agencies propose to define
standard data requirements to support the secure file exchanges, and utilize
web-based data exchanges, such as XML and XBRL taxonomies and related secure
technologies, to exchange SNC examination data.
a. XML and XBRL taxonomies and related secure technologies to collect SNC
examination data appears a reasonable platform.
9. Do reporting banks store sufficient information in their databases to
electronically identify a SNC according to the current criteria - $20
million or more with three or more lenders regulated by the Agencies?
As long as the submitting bank captures participating banks and global
facility amounts, then logic should allow for a fairly accurate
identification of the SNC pool; however, data integrity is assumed in this
example.
a. Would reporting banks need a resource to determine if the Agencies
regulate a lender?
It would be helpful if the agencies developed a data table of all banks that
included a flag to identify lender regulation.
b. Would other criteria help reporting banks identify SNCs and submit
data electronically?
If the suggestion in “a” is not feasible, the agencies could provide a list
of all their regulated institutions; therefore, if a participant is not on
the list then it will be assumed that the lender is not regulated.
c. Would a larger data feed to the Agencies, which the Agencies would
then screen for SNC criteria and then extract SNC facilities, be easier for
reporting banks to administer?
If the submitting bank could not easily identify the universe of SNCs, then
the submission of a larger data feed would be helpful; however, this method
of extraction should not be a requirement for those banks that have the
ability to identify the pool, but would be a good cross check to assure data
integrity.
10. Assuming that the proposed list of data elements is adopted, how
could that data be best presented to provide value to Basic and Expanded
Reporters (i.e., what views would be most advantageous)? See response below.
a. Alternatively, should the Agencies simply provide raw data tables to
support bank-generated reports?
A combination of canned reports and raw data would be beneficial since it
will provide the submitting bank with the ability to stratify the portfolio
on an as needed basis using their internal reporting capabilities.
b. Will your reporting bank be able to receive the feedback data and
reports electronically by the proposed 2007 implementation date?
Yes and looking forward to it.
11. Are there any unintended consequences that might arise from the use
of this comparative information?
As long as the comparative data is presented to the submitting
institution and provides aggregate high level comparisons, then the sharing
of information should not result in unintended consequences. If however, the
data is accessible through a website, then the risk of hacking exists;
therefore, it will be imperative to develop a highly secure data house,
accessible by an identifiable user base.
With expanded access to other entities ratings and methodologies even in a
“masked” environment, the negative consequences of “group think” could
permeate the independent thinking of the individual institutions. The risk
here appears minimal though as the banking industry today tends to have very
robust, independent ratings systems that would provide a reasonable check
and balance.
12. The Agencies currently ask reporting banks to provide the name, city,
and state for SNC borrowers. This has often not been enough information to
clearly identify borrowers in the SNC database. The Agencies are looking for
additional data that reporting banks might provide to help identify their
borrowers more clearly (e.g., stock tickers, taxpayer identification
numbers, CUSIP numbers, MKMV's PIDs, etc).
a. Which of these additional data elements would be most useful for this
project?
It depends on data capture and integrity. A PID, TIN or ticker should be
easily accessible; however, these numbers are not always available from the
bank’s internal record source and would have to be assigned manually. The
DUNS number could also serve as a unique identifier. Further, items such as
MKMV Ids require subscription to a specific third-party software. The
interagencies should be sure not to provide a perception that they are
endorsing one vendor over another. Providing fields for many different
identifies would afford each institution the ability to populate as many
identifiers as is available within their architecture.
b. What are the minimum data required to clearly identify borrowers and
facilities?
From the submitting bank’s perspective, the internal customer number and
facility ID serve as the identifiers; however, to match each customer and
facility to a participant’s record is more difficult; therefore, the
regulating agencies may want to consider a web-based database that can be
accessed by the participating banks which provides the unique identifiers to
the submitting bank.
c. Which, if any, of these items do reporting banks store electronically?
The PID is available for publicly traded companies; however, not all
customers in a syndicated transaction are public. Therefore, the PID would
work for a small universe of customers. Each agent bank could request the
mapping of customers to a DUNS number using the customer name and address as
the primary means of identification. Then, all agents and participants would
use the DUNS number as a primary unique identifier. If the submitting bank
does not have a DUNS number for its customer and/or bank pools, it can be
requested from D&B. Going forward, the DUNS number could be obtained from
the company as a part of the loan-setup process.
d. Is the proposal to require submission of at least one of these items
reasonable?
Requesting at least one identifier is reasonable and most likely should
include the DUNS numbers since most banks could easily obtain the DUNS
number for the majority of customers in their database.
13. Over the past two decades, some of the industry's largest losses
involved credits extended to groups of related borrowers.
a. How are reporting banks identifying groups of related borrowers in
their own systems?
Related borrowers are tied together using a relationship number. The
advancement of Basel II initiatives will further enhance the robustness of
these relationships.
b. What data could participating reporting banks provide to help identify
related borrowers in SNC credits?
Common hierarchies for SNC credits available to participant banks would
minimize the duplication efforts performed institution to institution..
c. Could reporting banks electronically transmit data on guarantors for
credits, sponsors, or other related and relevant parties?
Currently, we do not have easy access to such data. Further, the
categorization of guarantor information other than names, addresses and SSN
are not standardized within the industry. Financial Institutions would need
extra assurance of compliance with privacy laws, et al. before they would
willingly disclose this type of information.
14. Could the reporting banks provide entries tracking the resolution of
credits over time, such as amounts charged off or sales of assets since the
last data submission?
For quarterly submission, charge-offs are captured for the month-end
date; therefore, it would require some programming to compute aggregate
amounts for each quarter. To facilitate the submission of such data, it
would be beneficial to allow the submission of a separate file containing a
unique identifier for solely those records in question, rather than require
the addition of several new columns of data most of which would be null.
15. The data submission software currently in use (OSCAR and SNC
Reporting Application) does not easily support aggregated reporting of SNC
information for all of a reporting bank's related entities. Should the
Agencies design software to permit aggregate, single-point, reporting of SNC
data for a reporting bank? Should electronic data file submission also allow
this type of reporting?
It may be easier to provide a file to the submitting bank for internal
reporting rather than design software for this purpose.
Submitted by
Margherita Stargard
VP - PCRM, Commercial Credit Services
617-994-7293
Mr. Paul Butler
SVP - Portfolio Credit Risk Management<
Citizens Bank
53 State Street
Boston, MA 02109
617-994-7341
