SPST_unbankedpopulationsurvey_Part_B_(Modified 12-3-08)

SUPPORTING STATEMENT

NATIONAL SURVEY OF UNBANKED AND UNDERBANKED HOUSEHOLDS

PART B

B. COLLECTION OF INFORMATION EMPLOYING STATISTICAL METHODS

1. Universe and Respondent Selection

The Census Bureau will conduct the January 2009 National Survey of Unbanked and Underbanked Households supplement (“Unbanked/Underbanked supplement”) in conjunction with the Current Population Survey (CPS), for which the universe is 116 million households. From this universe, the Census Bureau selects a sample of approximately 72,000 households each month, of which approximately 59,000 are eligible for interview. The Census Bureau actually interviews 54,000 households each month and will ask the items in the Unbanked/Underbanked supplement, as appropriate, for members of these households.

The Unbanked/Underbanked supplement represents a new collection. The enumeration unit will be the household, and the respondent will be a household member who participates in the financial decisions of the household.

The response rate for the basic CPS is about 91.5 percent. The Unbanked/Underbanked supplement will have a lower response rate because, at the time the basic CPS interviewing is conducted with a household, there may not be a qualified individual available to provide responses to the Unbanked/Underbanked supplement, and no follow-up attempts will be made. Although the FDIC does not know how much lower the response rate will be for the supplement, some data provided by Census staff for the CPS Food Supplement gives an idea of what type of response rate we can expect. For that supplement, the person who answered the labor force questions was the most knowledgeable/head of household 82 percent of the time. Also, the most knowledgeable person was available for about another 1.5 percent of the time. This suggests that a response rate of about 84% may be achieved for the Unbanked/Underbanked supplement, even though the criteria for identifying an eligible respondent for the household differs from that of the Food Supplement.

For an 84% response rate the potential for nonresponse is minimal. However, unless the response rate exceeds 85%, the FDIC plans to investigate the potential for nonresponse bias by making some basic comparisons of the characteristics of the households that provide responses to the FDIC supplement to those that provided responses to the basic CPS but not to the FDIC supplement. For example, a comparison will be made for renter- or owner-occupied households to see if the percent of renter-occupied households that responded to the FDIC supplement varied from the corresponding percent for those households that did not respond to the FDIC supplement. The same type of comparison may be made for other household characteristics that are available, such as type of structure (single or multiple-family) or household size (number of persons aged 15 or more).

If the FDIC finds that the percent of renter-owned units varies much between respondents and nonrespondents to the FDIC supplement, we can check to see if the answers to survey questions varies by renters versus owners, which would give us a measure of the potential for nonresponse bias. The same type of analysis will be conducted for one or two other household variables that are available.

2. Procedures for Collecting Information

The Unbanked/Underbanked supplement is a supplemental survey associated with the CPS in January 2009. Attachment D gives an overview of the CPS sample design and weighting methodology and response rates. The statistical properties of this supplemental survey will fall within those associated with the CPS itself.

3. Methods to Maximize Response

Response rates and data accuracy for the CPS are maintained at high levels through interviewer instruction, self-study training, and follow-up of refusal interviews with more experienced senior interviewers. Additionally, the Census Bureau closely monitors data output. (Refer to Item 5 of Attachment D for a discussion of the CPS nonresponse.)

4. Testing of Procedures

Questionnaire design was the result of a collaborative effort between the FDIC and a national consulting firm that specializes in public opinion research, with input from the Census Bureau’s Demographic Surveys Division and the Labor Department’s Bureau of Labor Statistics.

The FDIC’s contractor, under the FDIC’s supervision, conducted four rounds of cognitive pre-testing of the Unbanked/Underbanked supplemental questions. The tests began in February 2008, and concluded in December 2008. The goal of each round was to determine respondents’ comprehension of each question, test the flow of the questions, find major recall difficulties, ascertain the sensitivity or inappropriateness of any questions, and gauge the operational feasibility of the supplement.

The first round of cognitive pre-testing consisted of 25 interviews. Ten interviews were conducted in-person and 15 were conducted via telephone. Interviews were conducted with people representing a variety of ages, racial and ethnic backgrounds, levels of education and household size. Twenty-one participants qualified as “underbanked” and four qualified as “unbanked.” Most respondents in the first round gave the survey a positive review. Some respondents expressed that they had difficulty answering some questions due to terms or phrases that were unfamiliar. Some respondents considered banking done through a brokerage or with an online service (even if the online service was associated with a bank) as using a non-banking option to pay bills or borrow money. This would lead to over reporting the number of underbanked. More than three-quarters of those participating said the questions were appropriate for a Census Bureau Survey and that they would be willing to participate in such a study.

The survey questionnaire was revised to address feedback from the first round and subjected to a second round of testing.

The second round of testing consisted of 25 interviews. Ten interviews were conducted in-person and 15 interviews were completed via telephone. Nine of the respondents qualified as “unbanked” while 16 of the interview participants qualified as “underbanked”. Most of the survey questions tested very well and did not require changes. Interview participants were generally positive in their evaluation of the survey, considered the questions appropriate for a Census Bureau survey, and indicated that they would not decrease their willingness to participate in a Census Bureau survey due to the supplement questions. Minor modifications were made to the survey questionnaire following the second round including adding one response option, simplifying terminology, and changing a skip pattern.

A third and fourth round of testing were initiated as the result of revisions to the survey primarily to clarify that the focus of survey questions is on the household rather than on the individual respondent. In addition, minor modifications were made to some response options for purposes of clarity and consistency. The third and fourth rounds of testing consisted of 25 interviews and 24 interviews, respectively. Interview participants in both rounds were positive in their evaluation of the survey instrument, deemed the survey questions as appropriate for a Census Bureau supplement, and indicated the questions would not decrease their willingness to participate. After the third round of testing, minor modifications to the survey included adding a response option to several questions, changing a skip pattern, and adding clarifying terminology. No changes to the survey were recommended following the fourth round of testing.

5. Contacts Statistical Aspects and Data Collection

The following Census Bureau staff may be contacted on the statistical data collection and analysis operations:

Statistical Design

Ruth Ann Killion Chief, Current Population Surveys Branch

Demographic Statistical Methods Division

(301) 763-2048

Data Collection/Survey Design

Kathleen P. Creighton Assistant Division Chief, Continuing Surveys

Demographic Surveys Division

(301) 763-3814

6. Data Analysis Plan

The main survey estimates to be prepared from the data collected in the January 2009 National Survey of Unbanked and Underbanked Households will be the number and percentage of households that are “unbanked” (i.e., do not have a deposit account at a bank or credit union) or “underbanked” (i.e., have a deposit account but also rely on alternative, non-bank financial service providers). These estimates will be prepared for the nation, for states, and for large metropolitan areas (MAs).

For unbanked households, results from the FDIC CPS supplement will be used to quantify the reasons why households are unbanked. For both unbanked and underbanked households, estimates will be prepared of the usage levels of alternate, non-bank financial services, such as check cashing firms, payday lenders, pawn shops, and rent-to-own service providers.

Also, the economic and demographic information included in the CPS will be used to prepare tables that show, and allow comparisons of, the economic and demographic characteristics of unbanked, underbanked, and banked households. Furthermore, this basic CPS data will permit the FDIC to examine whether the usage of and reasons for choosing alternative non-bank financial services providers varies across economic or demographic segments of the population.

The FDIC will produce cross-tabulations of many of the basic CPS variables with responses to questions about: account ownership, reasons for non-account ownership (for unbanked respondents), the usage of specific alternative nonbank financial services, the reasons for use of these services, and the usage of prepaid cards. Following is a list of basic CPS variables that we plan to include in these tables:

1) Age of household reference person

2) Educational attainment of household reference person

3) Race/Ethnicity of household reference person

4) Marital Status of household reference person

5) Household composition (i.e., family or group household)

6) Citizenship of household reference person (US versus foreign born)

7) Household Income

8) Health Insurance Coverage of household reference person

9) Labor force status of household reference person

Cross-tabulations will be produced for the nation, for states, and for large MAs.

Example Table Shells for the FDIC Unbanked/Underbanked Population Survey

Table 1. Age of Household Reference Person by Banking Status: Estimated Total Household Counts (with Row Percents in Parentheses) for the Nation (or XYZ State or MA).

Age of Household Reference Person	Household Banking Status
	Banked	Underbanked	Unbanked	Total
16-24
25-34
35-49
50-64
65+
Total

Table 2. Education of Unbanked Household Reference Person by “Banks Have Inconvenient Hours” as a Reason for Not Having a Bank Account: Estimated Total Household Counts (with Row Percents in Parentheses) for the Nation (or XYZ State or MA).

Education of Unbanked Household Reference Person	“Banks Have Inconvenient Hours” Cited as a Reason for Not Having a Deposit Account
	Yes	No	Total
Less than HS
High School
Some College
College Degree
Post College Education
Total

Table 3. Citizenship of Household Reference Person by Use of Payday Lender: Estimated Total Household (with Row Percents in Parentheses) for the Nation (or XYZ State or MA).

Citizenship of Household Reference Person	Has Household Ever Used Payday Loan or Payday Advance Services?
	Yes	No	Total
U.S. Born
Foreign Born
Total

ATTACHMENTS

A1. FDIC National Unbanked and Underbanked Household Survey (English version)

A2. FDIC National Unbanked and Underbanked Household Survey (Spanish version)

B. Current Population Survey – Advance Letter and Frequently Asked Questions

C1. Current Population Survey Confidentiality Brochure

C2. Fact Sheet for the Current Population Survey

D. Current Population Survey Design and Methodology Overview, October 2006 (available at the following link: http://www.census.gov/prod/2006pubs/tp-66.pdf