Log in / Register
Home arrow Business & Finance arrow Reengineering community development for the 21st century
< Prev   CONTENTS   Next >

Publicly Available Performance Data

Financial Performance

Two bank financial performance measures are commonly accepted as comparable indicators of profitability: return on equity (ROE) and/or return on assets (ROA). Of the two, ROA may be preferable for CDFI banks because they often have lower "gearing ratios” due to the nature of their business and hence are unable to earn higher ROEs. These measures may vary year to year, so if one is interested in long- run performance, averages must be calculated over time. For depository institutions, data necessary to construct such measures are available through publicly accessible databases such as FDIC's Statistics of Depository Institutions (SDI).

Dobbs and Roller (2005) go further, suggesting in addition to ROE, compound revenue growth over time is also an important measure of the long-run financial performance of a business. Using the sixty-two accredited CDFI Fund institutions in 2006, we experimented with possible composite measures of financial performance that blend a number of underlying dimensions of financial performance, including ROE, ROA, efficiency ratio, tier 1 leverage ratio, charge offs to loan book, and loan loss to reserves.

We ranked the CDFIs in each category for each year, then calculated the average of the six rank scores. However, perhaps not surprisingly, there is a strong correlation (0.76) between ROE and the composite financial performance measure, since most of the measures feed into an ROE or ROA result. Furthermore, it requires an a priori judgment to assign scores in certain categories: for example, should higher charge offs necessarily lead to lower rankings if this is the result of a decision to enter riskier markets, and if the risk-adjusted return is still adequate? Consequently, we have reverted to the simple, consistent option of using ROA averaged over a preceding five-year period as a long-run financial performance measure for CDFI banks.

Social Performance

Table 7.2 summarizes publicly available databases relevant to social performance. FDIC's SDI database, which provides only financial information, is not used in our analysis. Each database has some limitations:

HMDA: While most (but not all) of the CDFI Banks report HMDA data, this captures only home lending, which is a declining proportion of most CDFI banks' real estate portfolios: single- and multifamily lending has declined from 43 percent to 36 percent of total loans outstanding between 1996 to 2004; cor-

Table 7.2. Summary of Public Data Bases


Maintained By

Who Reports

What the Data Tell Us

1. Community Reinvestment Act (CRA)


All supervised banks with assets above $250 m pre-2005 required to report loan data; annually; threshold now $1bn in assets

Originations of loans in developmental categories (to small business; agriculture) by county/census tract; CRA exam ratings, undertaken periodically by supervisors provide an indication of official rating of development performance

2. Home Mortgage Disclosure Act (HMDA)


All depository lenders and other qualifying mortgage lenders in general with home purchase loan originations exceeding $25 m annually

Applications for and originations and purchases of home loans by type and by characteristics of borrower by census tract as opposed to loans outstanding

3. Summary of Deposits (SOD)


All FDIC-insured institutions; annually

Location of branches and deposits booked per branch

4. Statistics on Depository Institutions (SDI)


All FDIC-insured institutions; quarterly

Only financial information

respondingly, commercial real estate, which may only be picked up in CRA reports, has become increasingly important (32 percent in 2004, up from 20 percent over the same period). This means that HMDA is picking up less and less over time of the real estate-related development lending of CDFIs.

CRA: Up until 2004, only entities larger than $250 million in assets were required to complete the detailed CRA reports, which enable new lending to small businesses to be identified separately by area. The rise in the reporting threshold to $1 billion from 2005 and the change in the methodologies mean that most CDFI banks, and indeed most small banks in general, will no longer be required to report origination at a tract level, so whatever the historic value of this information, its value as a source of detailed data on small banks has already diminished. (In 2005, there were only 1,103 CRA reporting banks, only 59 of which had assets less than $249 million.)

SOD: Differential practices on booking deposits distort the information value of data on the value of deposits by branch. Therefore, a measure of

Table 7.3. Number of Observations Found in Each Database in Each Year





1. CRA








3. SOD




4. SDI




percentage of branch offices in qualifying areas was used as a proxy for the focus of banking presence.

To construct a database for CDFI banks, considerable effort was necessary to extract data from the different databases and combine them. The SOD and SDI data for each year and for all institutions in the CDFI bank dataset are available at the FDIC Web site. The underlying HMDA and CRA datasets used in this analysis were available at the Federal Financial Institutions Examination Council (FFIEC) Web site for years after 1999 and could be ordered on CD-ROM for the earlier year in our sample. The transactional databases for CRA and HMDA raw data are large, requiring considerable manipulation to obtain usable data that combines the information by institution. While there are some commercial services that provide links across the databases by institution to facilitate such comparison, we found these to be very expensive; hence, we decided to test accessing these data directly from source.

Not all CDFI banks existed in 1996; and many of those that did, because of their size, were not required to report in each year (see Table 7.3). So few CDFI banks qualified to undertake detailed CRA reporting during this period that we could not use these data.

Given these limitations, what use can be made of these data related to social performance in particular? One of the key differences between CDFI banks and others is, in theory at least, that they do a sizable proportion of their business, whether lending or deposit taking, in qualifying areas. Business undertaken in such qualified areas is considered development-related business, which qualifies for consideration under the CRA lending test. Using databases mentioned earlier, differences should be measurable over time and can be calibrated to enable better comparison among entities that do relatively more, or relatively less, business in these areas. Specifically, we proposed two social performance measures (SPMs), which allowed us to distinguish among CDFIs and within broader groups (see Table 7.4):

• Development Lending Intensity (DLI), which should apply to all categories of lending deemed to have positive social impact (such as housing, SME, rural, commercial real estate) measured at a qualifying area level.

Table 7.4. Social Performance Distinguishing Features




1. Development Lending Intensity (DLI)

The total value of development- related loans originated and purchased by a bank in qualifying areas in a year as a percentage of total loans originated by bank / at time t. DLI may also be calculated relative to some measure of size, such as tier 1 equity or total assets

A higher DLI value means relatively more of a bank's lending takes place in qualifying areas; a DLI value increasing over time indicates an intensifying focus on lending in qualifying areas

2. Development Deposit Intensity (DDI)

(iii) DDI(t)

(i) The value of deposits booked in qualifying areas as a percentage of total deposits; or

(i) The extent to which the bank draws its resources from qualifying areas

(ii) DDI(b)

(ii) Percentage of total offices located in qualifying areas

(ii) The extent to which the bank provides deposit and retail financial services in qualifying areas

• Development Deposit Intensity (DDI), which considers deposits taken or branches located in qualifying areas, similar to the service test considered as part of a CRA examination.

In many ways, these measures are at the heart of the CRA examination. The CRA examination report records the originations of home loans and small business loans by number and value, differentiated by geography of the borrower and even income of the borrower for home loans. This profile is compared against averages for a regional comparison grouping of banks to form an opinion on the reasonableness of the distribution of bank lending patterns. Similarly, for the location of retail service points, the current spread of branches and the trend in branch openings and closings in different income areas is considered in coming to an assessment of the perfomance of the institution.

The outcome of a CRA examination, conducted every three years or so, is a rating on a four-point scale. Examination methodology has been standardized across the bank supervisory agencies to attempt to make CRA ratings more comparable. Nonetheless, ratings are highly clustered in the upper two categories (outstanding, satisfactory) and are used only for regulatory purposes. They have not, to our knowledge, been used for investment purposes, because the CRA rating does not, nor does it seek to, systematize the underlying information in a way that can be used by investors to understand social perfomance. Instead, the derivation of DLI and DDI measures on a consistent basis is relevant to categorizing and comparing the social performance of banks with one another and over time.

The DDI measure is accessible annually through the FDIC's SOD database. However, the calculation of DLI from publicly available data presents a challenge: very few of the CDFI banks (live, from Table 7.3), or small banks under $1 billion in general, are required to produce annual lending reports for CRA. Because of lower exemption thresholds, many more are required to report HMDA data: forty- three in 2004 (Table 7.3). However, to calculate DLI on one asset class alone may create a misleading picture of the development lending of a bank: the bank may undertake little housing lending in investment areas because the bulk of its business is SME or commercial real estate investment in these areas. Since at present we can calculate only the DLI measure for home lending, this measure has descriptive power only for those entities that focus on home lending. We therefore distinguish DLI on the basis of HMDA alone as DLI (HMDA) and apply it for lenders for which home lending is more than 50 percent of their loan portfolios in 2004. For those lenders that are not housing focused, DLI (HMDA) provides a guide only to their housing activity, which is relatively small, and says nothing about their social performance.

Found a mistake? Please highlight the word and press Shift + Enter  
< Prev   CONTENTS   Next >
Business & Finance
Computer Science
Language & Literature
Political science