Social Performance Measurement for CDFI Banks

David Porteous and Saurabh Narain

How can one assess the social performance of community development banking institutions (CDBIs)? The National Community Investment Fund has developed the term social performance to mean those depository institutions that attain positive social impact while remaining financially strong. CDBIs include banks, thrifts, and credit unions but are not restricted to those accredited by the U.S. Department of Treasury's Community Development Financial Institutions Fund (CDFI Fund). In this chapter, our analysis focuses on banks and thrifts (hereafter collectively called CDFI banks) but not credit unions.

Two problems stand in the way of better performance assessment. First, while there are methods to assess financial performance of banks, CDBIs are, by definition, “double bottom line" institutions that balance profit making with positive social impact. CDBI shareholders and management consciously make trade-offs between the two bottom lines – or the three, if one includes environmental impact, as is increasingly common (e.g., the Global Reporting Initiative, global- As a result, measurement of the financial bottom line and comparison with “single bottom line" banks based on financial performance alone are inadequate. Second, even within the universe of depository entities concerned with social impact, there is no standard system of performance measurement. Instead, a multitude of approaches is used, often requiring collection of considerable additional information.

Efforts have been made to collect better data about the CDFI sector; however, some of these efforts – such as the CDFI Data Project ( asp) – have focused on compiling aggregate data rather than comparing individual institutions. The CDFI Assessment Rating System (CARS) ( provides standardized means of rating CDFI loan funds only and does not cover CDBIs. Also, the CARS relies on self-selected methods of determining impact and, like most rating services, requires collection of specialized information.

If CDBIs are to be a standard class, able to tap wider sources of retail investment, then there is a need to create standardized means of performance assessment so that CDBIs can be included in the broader universe of banks. This could enable investors to make investments on the basis of their own balance of objectives between financial and social return. Entities offering an attractive combination of both may benefit from greater access to capital and even a lower cost of funds than is possible in the absence of calibrating the investment trade-off. However, any approach to measuring social performance across banks must rely on publicly available data, since costs of collection would limit the universe of entities to a small self-defined set. Fortunately, there are rich sources of public data relating to the activities of home lenders – the Home Mortgage Disclosure Act (HMDA) – and depository institutions – Community Reinvestment Act (CRA) or the Federal Deposit Insurance Corporation's (FDIC) Summary of Deposits (SOD).

What then are the limits to which such publicly available data can be pushed in the cause of meaningful performance management for CDBIs? Can data be used to assist in performance measurement of CDBIs? To do this, we first survey performance assessment approaches, then assess the scope and usefulness of public data sources. Then, we access publicly available databases to create a profile over time of the sixty-one depository institutions accredited as CDFIs in July 2006. Our sample focuses on the period since the creation of the CDFI Fund in 1996, when the CDFI designation was started. We test and demonstrate the value of two measures derived from these databases. In the presence of sufficient data, these measures may be used as social performance metrics in themselves; certainly, they can function as lenses that enable better categorization and comparison of like with like. This categorization is an essential step in better performance comparison across entities, often missed.

The CDFI Fund Web site ( reported the names of sixty-one certified banks and thrifts in July 2006. While these entities were certified at various dates during the past ten years (and it has been impossible to secure a date of first certification in all cases from the CDFI Fund), we use the July 2006 cutoff to define the CDFI bank dataset for which we collect and report retrospective performance data.

To make the volume of data more tractable, three years were selected to represent the start, middle, and end points of this period: 1996, 2000, 2004. Table 7.1 compares the size distribution of our CDFI bank sample with all U.S. banks in 2004. In addition, it is relevant to compare lending activity undertaken by CDFI banks. Figure 7.1 reports the percentage of outstanding loans at year end in each of the subcategories related to real estate lending for CDFI banks on the left and small banks (less than $1 billion in assets) on the right.

Figure 7.1 shows real estate lending as the largest category for both CDFI and small banks as a whole – close to 70 percent by 2004. The difference between the two groups of banks is in composition of real estate lending: for CDFI banks, share of single- and multifamily loans has fallen from 43 percent of all loans to 36 percent (with a much higher share for multifamily); in their place, share of commercial real estate grew dramatically to a third in 2004 from a fifth in 1996. Small banks have increased single- and multifamily lending (50 percent to 52 percent), and home equity lending (although this remains a small proportion).

Table 7.1. Relative Size Comparison of CDFI Banks (Number of Banks Unless Specified Otherwise)

Gross Assets

End 2004

CDFI Banks

All U.S. Banks

< $250 m



> $250 m; < $1 bn



> $1 bn



Average total assets

$173 m

$168 m

Figure 7.1 Composition ol Real Estate Portfolio of CDFI Banks Versus Small Banks

Composition ol Real Estate Portfolio of CDFI Banks Versus Small Banks

Because of the focus on CDFIs, we used the CDFI Fund definition of qualifying Investment Area.[1] This status is conferred by the CDFI Fund on the basis of census results that show unemployment and poverty rates against national benchmarks and median income lower than area median. The latest list of investment areas is based on 2000 Census results, published in 2001. Unfortunately, we were unable to obtain the list of qualifying areas for prior years. Hence, we re-created the prior list using similar underlying measures of poverty, unemployment, and relative income based on 1990 Census data. This gave us some 20,000 qualifying tracts for the earlier two years and the official number of 24,795 in 2004. Area definitions also were adjusted to allow for changes in the 2000 Census data, using a specific translation algorithm.

  • [1] Other screens would be low and moderate income (LMI), published annually by HUD, which considers tract income relative to area median, and more recent categories of financially distressed and financially underserved areas declared by FFIEC in terms of recent regulation. While this measure will be updated more regularly than CDFI investment areas, it has become available only recently and covers only nonmetropolitan counties.
< Prev   CONTENTS   Next >