This paper is released to inform interested parties of ongoing research and to encourage discussion of work in progress. Any views expressed on statistical, methodological, technical, or operational issues are those of the authors and not necessarily those of the U.S. Census Bureau.

In August 2008, the U.S. Census Bureau sponsored a supplement to the monthly Current Population Survey (CPS) questionnaire. The Immigration/Emigration Supplement, also referred to as the Migration Supplement, included questions on monetary transfers and four additional topics – citizenship, year of entry, residence one year ago, and residents and emigrants abroad. The Monetary Transfers section is noteworthy because it represents the first time questions on both the giving and receiving of transfers were included on a large, federally sponsored, nationally representative survey in the United States. It is also significant because the questions were administered to both native and foreign-born households. The purpose of this paper is twofold. First, it describes the CPS Migration Supplement, a new source of migration-related data. Second, it discusses the quality of the monetary transfer data and presents some initial results on the occurrence, frequency, and amount of transfers sent and received. 

In August 2008, the U.S. Census Bureau sponsored a supplement to the monthly Current Population Survey (CPS) questionnaire. The Immigration/Emigration Supplement, also referred to as the Migration Supplement, focused on five topics – citizenship, year of entry, residence one year ago, residents and emigrants abroad, and monetary transfers – and included questions not commonly found on other surveys. The Monetary Transfers section is especially noteworthy because it represents the first time questions on both the giving and receiving of transfers were included on a large, federally sponsored, nationally representative survey in the United States. This section is also noteworthy because, unlike other polls and surveys that target immigrants, the transfers questions were administered to both native and foreign-born households. In total, six questions on monetary transfers were included on the Supplement, asking households: whether or not they sent money to or received money from friends and relatives living outside the United States; how often they sent and received; and how much they sent and received. These questions represent a first step toward providing data for an area of research where little previously existed.

The purpose of this paper is twofold. First, it describes the 2008 CPS Migration Supplement, a new source of migration-related data. Second, it discusses the quality of the monetary transfer data and presents some initial results on the occurrence, frequency, and amount of transfers sent and received. The data are tabulated and discussed for both native and foreign-born households. Typically, analysts define the nativity status of a household based on the nativity of either the householder or the primary respondent. However, for this paper, the nativity status of a household is based on the nativity status of its adult members. Specifically, foreign-born households include any household with at least one foreign-born member 18 years of age or older. Conversely, native households are those households with no foreign-born members age 18 or older.¹ As will be shown, the monetary transfers behavior of native and foreign-born households exhibit substantial differences.

This paper begins by defining “monetary transfers,” explaining why this term – rather than the more common term “remittances” – was used in the Migration Supplement. Next, the Current Population Survey is briefly discussed, providing some basic information necessary to understand the structure of the Migration Supplement questionnaire and the advantages and limitations of the data collected by this instrument. Details about the goals of the Supplement are then reviewed and its structure discussed. The paper then turns to a detailed description of the Monetary Transfers section, reviews the number of households in universe for each question, and discusses item nonresponse. Data tables are then provided that summarize the information needed to answer four questions: 1) What proportion of households in the United States sends or receives monetary transfers? 2) How often do households send or receive monetary transfers?   3) How much money do households typically send or receive? 4) How much money do households send or receive in total? All of the tables present data for both native and foreign-born households, and differences in the monetary transfers behavior by nativity are examined. This paper concludes with a discussion on the quality of the monetary transfer data collected by the Migration Supplement and a comparison with results provided by other research.

In this paper, the term “monetary transfers” is used rather than the more common term “remittances.” There are two reasons for this. First, there is no single, standard definition of remittances, and authors and organizations have used the term to discuss a variety of resources: from cash exchanges only; to cash and “in-kind” transfers (e.g., goods and services); to cash and in-kind transfers plus compensation to employees (e.g., wages, salaries); to cash and in-kind transfers, compensation to employees, and capital transfers of financial assets; to even more complex measures that include social benefits and transfers to non-profits (Bach, 2005; International Monetary Fund (IMF), 2010). The term monetary transfers is used in this paper and in the Migration Supplement to clearly define what the data include.

Second, in order to collect data that were as accurate as possible, it was decided to focus on a specific behavior (i.e., the exchange of money between members of resident and nonresident households) that could be defined and measured by a small number of questions and collected by a household survey. Monetary transfers, while only part of how many researchers define remittances, represent a significant part of “personal transfers” – all cash or in-kind transfers between members of resident and nonresident households – a concept used by the Bureau of Economic Analysis in their annual remittance estimates for the United States (Bach, 2005; Bai and Hoang, 2010) and other researchers and organizations (Haug and Teotino, 2008; IMF, 2010). While limited, the concept of monetary transfers was used in the Migration Supplement to provide analysts studying remittances with the most accurate, useful data possible.

The Current Population Survey (CPS), sponsored jointly by the U.S. Census Bureau and the U.S. Bureau of Labor Statistics, is the primary source of labor force statistics for the population of the United States. The CPS is the source of numerous high-profile economic statistics, including the national unemployment rate, and provides data on a wide range of issues relating to employment and earnings. Although labor market information is central to the CPS, the survey provides a wealth of other demographic, social, and economic data that are widely used in both the public and private sectors.

The CPS is administered by the Census Bureau using a probability selected sample. Approximately 72,000 housing units are selected from the sampling frame, resulting in a sample of about 60,000 occupied housing units eligible for interview.² Approximately 55,000 households are interviewed each month.³ The universe is the civilian noninstitutionalized population living in housing units in the 50 states and the District of Columbia and members of the Armed Forces living in civilian housing units on a military base or in a household not on a military base. People in institutions, such as prisons, long-term care hospitals, and nursing homes are ineligible to be interviewed in the CPS. One person generally responds for all eligible members of the household. The CPS is administered via telephone by means of computer assisted telephone interview (CATI) or in person by means of computer assisted personal interview (CAPI) only. There is no mail out or paper version of the questionnaire. Housing units are in the survey for 4 consecutive months, out for 8 months, and then return for another 4 months before leaving the sample permanently.

In addition to the regular labor force questions, the CPS often includes supplemental questions on a wide range of topics, such as tobacco use, computer use, and voting patterns. The CPS interview is divided into three basic parts: 1) household and demographic information, 2) labor force information, and 3) supplement information. Supplemental questions are added to the CPS in most months. Perhaps the most well known supplement is the Annual Social and Economic (ASEC) supplement, fielded in February, March, and April of each year.

For additional information on the Current Population Survey, see Current Population Survey Design and Methodology (Technical Paper 66) available on the U.S. Census Bureau’s Web site at //www.census.gov/prod/2006pubs/tp-66.pdf.

In August 2008, the U.S. Census Bureau sponsored and fielded the Migration Supplement to the CPS monthly questionnaire. Broadly speaking, the purpose of the Supplement was to provide the Census Bureau with additional data to aid its mission of improving knowledge of the size and characteristics of the foreign-born population in the United States. The Supplement had four specific purposes: 1) to inform and possibly improve estimates of net international migration used in the postcensal population estimates produced annually by the Census Bureau’s Population Division; 2) to provide additional information on current migration patterns and the accuracy of various measurements of years “lived” in the United States by the foreign born; 3) to determine if accurate and reliable data on monetary transfers could be collected by a nationally representative, government-sponsored survey; and 4) to evaluate if rates of naturalization could be reliably estimated using household survey data.

The Migration Supplement included questions on: changes in citizenship status; year of first arrival, subsequent exits and entries, and time spent outside the United States; residence one year ago, either inside or outside the United States; movement by household members outside the United States within the last year and basic demographic information on those who emigrated; and money sent to or received from family and friends living outside the United States. The content of the questionnaire was developed and cognitively tested by the Census Bureau in the fall of 2007. The Supplement was administered August 17-26, 2008. Both CATI and CAPI were used to administer the Supplement. CATI interviews were fielded August 17-20 and CAPI interviews were fielded August 17-26. After data collection was completed, focus group debriefings of telephone interviewers and field representatives were conducted to evaluate the Migration Supplement (de la Cruz and Logan, forthcoming).

The Migration Supplement universe consists of every CPS household member, of all ages, including those in the Armed Forces. Additionally, the Migration Supplement collected information on people who lived in the housing unit during the previous 12 months but no longer did so. Since typically only one person per household was interviewed, most of those included in the sample had their information provided by proxy.

A question on citizenship status is included on the monthly CPS questionnaire to determine the citizenship status of all persons in a household. This question is asked only during the first month in sample and is not asked again in subsequent interviews. The Supplement citizenship question was asked only if a household was in its second or subsequent interview and if one or more members on the roster indicated they were noncitizens during the first interview. Questions from the Year of Entry section were asked only of household members who were not born in the United States. As in all sections of the Migration Supplement, the definition of the “United States” included only the 50 states and Washington, D.C.; Puerto Rico and the U.S. Island Areas, such as Guam and American Samoa, were considered to be outside the United States. Questions from the last three sections – Residence One Year Ago, Residents and Emigrants Abroad, and Monetary Transfers – were asked of all members of all households. Most households were only asked the questions from the last three sections.

The purpose of the Monetary Transfers section was to collect data about the occurrence, frequency, and amount of money sent and received by households in the United States to and from friends and relatives living outside the United States. The Migration Supplement represents the first time questions on monetary transfers were included on a federally sponsored, nationally representative survey in the United States.

In the Migration Supplement, “monetary transfers” refer to personal transfers of money to and from relatives or friends, regularly or occasionally, for support or as a gift. They do not include transfers of money to or from organizations, such as donations to charities; the transfer of goods or services (i.e., in-kind transfers); or loans. Monetary transfers include all personal transfers between someone staying or living in a household at the time of the transfer and someone staying or living anywhere outside the United States at the time of the transfer.

The Migration Supplement included six questions on monetary transfers, three questions on the money sent and three questions on the money received by the household during the previous 12 months. More specifically, the questions were:

1. In the last 12 months, did anyone in this household give or send money to relatives or friends living outside the United States? Please include all gifts of money. Do not include loans.
2. About how many times in total over the last 12 months was money given or sent? (Your best guess is fine.)
3. In the last 12 months, about how much in total dollars was given or sent? (Your best guess is fine.)
4. In the last 12 months, did anyone in this household receive money from relatives or friends living outside the United States? Please include all gifts of money. Do not include loans.
5. About how many times in total over the last 12 months was money received? (Your best guess is fine.)
6. In the last 12 months, about how much in total dollars was received? (Your best guess is fine.)

All questions were asked at the household level, so single transfers by anyone in the household to or from anyone outside the United States during the previous 12 months were aggregated, creating a single total for the entire household. The Monetary Transfers section was asked of all households participating in the Supplement, including those with no foreign-born members. If the household answered “yes” to the question on sending money to relatives or friends outside the United States, they would be asked questions on the frequency and amount of money sent, then skip to the question on receiving money from relatives and friends outside the United States. If they answered “no” to sending money abroad, they would skip to the question about receiving money. If they answered “yes” to receiving money, they would be asked questions on the frequency and amount of money received, and then the Supplement concluded. If they answered “no” to receiving money, the Supplement concluded.

The August 2008 CPS sample included 71,573 housing units, including 58,821 eligible and 12,752 ineligible housing units.⁴ Of the 58,821 eligible housing units, 54,282 households were interviewed. The response rate was 92.3 percent, which was comparable to the response rates for the three previous monthly surveys (May, June, and July 2008) and the survey in August 2007.

Of the 54,282 interviewed households, 7,560 households had at least one foreign-born member age 18 years or older, while the remaining 46,722 households had no foreign-born adult members. If households did not complete the last section of the monthly CPS questionnaire, declined to participate in the Migration Supplement, refused to complete prior sections of the Supplement, or refused to answer the Monetary Transfers section, they would not have been in universe for the questions on monetary transfers. A total of 50,872 households were in universe for the monetary transfers questions, including 44,388 native and 6,484 foreign-born households (Table 1). If households responded “yes” to sending money, they were in universe for the questions on the number of times and amount sent. Additionally, if they responded “yes” to receiving money, they were in universe for the questions on the number of times and amount received. A total of 2,185 households were in universe for questions on the number of times and amount of money sent, including 451 native and 1,734 foreign-born households. A total of 458 households were in universe for questions on the number of times and amount of money received, including 171 native households and 287 foreign-born households.

In the Migration Supplement, if a household failed to provide complete and usable information for a data item, an answer was not allocated or assigned. Missing data, or item nonresponse, occurred when households refused to answer a question, responded “don’t know” or provided some other invalid response, or were in universe but were not administered the question.⁵ For each question in the Monetary Transfers section, the item nonresponse rate was calculated by dividing the number of households with missing data by the total number of households in universe for that item (Table 1). Overall, the item nonresponse rates for the Monetary Transfers section questions were low, especially for the questions on whether or not the households sent or received money (both 1.4 percent) and how many times money was sent (5.1 percent) or received (6.3 percent). Not unexpectedly, the item nonresponse rates for the amount sent (12.4 percent) or received (15.1 percent) were higher, as respondents often view income-related questions as sensitive topics and are more reluctant to answer (Marquis et al., 1986). Of note is the difference in response rates between native and foreign-born households. For all of the Monetary Transfers section items, the nonresponse rates for native households were lower than those for total households, while the rates for foreign-born households were higher. These higher rates, however, are still comparable to the rates found for other items included in the CPS (U.S. Census Bureau, 2006).

A total of 50,148 households provided valid responses to the question on whether or not they sent money to relatives or friends outside the United States during the previous 12 months; 2,158 households responded “yes” and were asked questions on the frequency and amount sent. In addition, a total of 50,152 households provided valid responses to the question on whether or not they received money from relatives and friends abroad; 458 households responded “yes” and were asked questions on the frequency and amount received. When the household survey weight is applied, the number of households is estimated to be 115.5 million for the questions on whether or not money was sent and received, 5.9 million for the questions on the frequency and amount sent, and 1.1 million for the questions on frequency and amount received.

While the size of the total unweighted household sample is considerable, caution must be exercised when producing weighted estimates based on subdivisions of a population, as the number of unweighted cases that form the basis of those estimates can be significantly reduced. For example, the estimate of 1.1 million households that responded “yes” to receiving monetary transfers includes 778,000 foreign-born households and 371,000 native households; the estimate of the total number of households is based on 458 unweighted cases, including 287 unweighted cases used to estimate the number of foreign-born households and 171 unweighted cases used to estimate the number of native households. Division of the total number of households into too many subgroups could result in estimates based on too few unweighted cases and large standard errors. In this paper, the estimates presented are based on household groups with at least 50 unweighted cases. The number of categories used in the percent distributions also was kept low to ensure an adequate number of cases in each cell. In addition, the data for the frequency and amount of monetary transfers sent and received were accepted as reported.

What proportion of households in the United States sends or receives monetary transfers?

Between August 2007 and August 2008, 5 percent of all households reported sending monetary transfers to relatives and friends outside the United States (Table 2). Foreign-born households were more likely to send money than native households: 27 percent of foreign-born households reported sending monetary transfers, compared with 1 percent of native households.⁶ Of the 5.9 million households that reported sending monetary transfers abroad, 84 percent were foreign born.

Approximately 1 percent of all households reported receiving monetary transfers from friends and relatives outside the United States during the 12 months before being surveyed (Table 3). There are differences when the households are divided by nativity. Less than 1 percent of all native households compared with 4 percent of all foreign-born households reported receiving monetary transfers. Of the 1.1 million households that received money from abroad, 68 percent were foreign born.

A total of 364,000 households, representing less than 1 percent of all households, both sent and received monetary transfers, of which 75 percent were foreign born.

How often do households send or receive monetary transfers?

Of all households that reported sending monetary transfers to relatives and friends outside the United States and the number of times money was sent, over half (54 percent) remitted between one to four times and about 30 percent remitted 10 or more times (Table 4). On average, households sent monetary transfers 6 to 7 times during the previous 12 months. Considerable differences exist between the remitting patterns of native and foreign-born households. Among native households that sent monetary transfers abroad, about half (52 percent) remitted between one and two times, compared with 30 percent of foreign-born households. Only 14 percent of native households reported remitting 10 or more times, compared with 33 percent of foreign-born households.⁷ Foreign-born households, on average, sent money abroad about 7 times during the previous 12 months, compared with 4 to 5 times for native households.

Differences also exist between native and foreign-born households in the patterns of monetary transfers received. For all households that reported receiving monetary transfers from relatives and friends outside the United States and the number of times received, 60 percent received money one to two times while 40 percent received money three or more times, averaging 3 to 4 times during the previous 12 months (Table 5). Native households were less likely to receive monetary transfers more frequently: 31 percent of native households received monetary transfers 3 or more times compared with 44 percent of foreign-born households. On average, foreign-born households reported receiving monetary transfers about 3.8 times during the previous 12 months, while native households received money about 3.6 times.⁸

How much money do households typically send or receive?

Of those households that reported sending monetary transfers to relatives and friends outside the United States and the amount sent, over half (53 percent) sent less than $1,000, including 34 percent that sent less than $500 (Table 6). When the percent distributions of the amount sent by native and foreign-born households are compared, foreign-born households are more likely than native households to remit annual amounts greater than $500 – while 43 percent of native households reported sending more than $500, 71 percent of foreign-born households did so. Foreign-born households also are more likely to remit larger amounts: 11 percent of foreign-born households remitted $5,000 or more, compared with 7 percent of native households. The median amount sent also reflects this difference. Half of native households sent $384 or more, while half of foreign-born households sent $1,007 or more.

While foreign-born households are more likely than native households to remit large amounts, they are also more likely to receive large amounts. For all households that reported receiving monetary transfers from relatives and friends outside the United States and the amount received, 39 percent reported receiving less than $500 while 24 percent reported receiving $5,000 or more (Table 7). Over half (54 percent) of all native households received less than $500, with about 7 percent receiving $5,000 or more. By comparison, about 31 percent of foreign-born households received less than $500, while close to one-third (32 percent) received $5,000 or more.⁹ Again, the median amount received also suggests foreign-born households are more likely to receive larger amounts: half of native households received $383 or more, while half of foreign-born households received $1,088 or more.

How much money do households send and receive in total?

Data from the CPS Migration Supplement suggest that households sent an estimated $11.7 billion to relatives and friends outside the United States during the 12 months before the survey (Table 8). Due to sampling variability, a 90 percent confidence interval around the amount has a lower bound of $10.6 billion and an upper bound of $12.8 billion.¹⁰ In addition, households received an estimated $5.6 billion. Combined, a net estimated amount of $6.1 billion was sent abroad. Of the $11.7 billion sent, $10.5 billion, or 90 percent, was sent by foreign-born households; of the $5.6 billion received, $4.6 billion, or 82 percent, was received by foreign-born households. Foreign-born households sent abroad a net estimated amount of $5.8 billion.

The August 2008 CPS Immigration/Emigration Supplement is a new source for data on several migration-related topics not commonly covered by surveys, including changes in citizenship status, year of first entry and subsequent exits and reentries, residence one year ago, residents and emigrants abroad, and monetary transfers. The Monetary Transfers section is especially noteworthy because it represents the first time questions on both the giving and receiving of transfers were included on a large, federally sponsored, nationally representative survey and because these questions were asked of both native and foreign-born households.

One of the main reasons the monetary transfers questions were included on the Migration Supplement was to determine if accurate data about this topic could be collected on a government survey. We believe the data are of good quality for three reasons. First, the item nonresponse rates were relatively low for each of the questions, even when the households were divided by nativity. While the item nonresponse rates for the questions asking the amount sent (12 percent) and received (15 percent) were higher than for the other questions, they were still relatively low; that is, other income-related items on the CPS have been shown to have item nonresponse rates ranging from less than 1 percent to 36 percent, with an average of 17.7 percent (U.S. Census Bureau, 2006). All of the nonresponse rates for the monetary transfers items fall below this average.

Second, the patterns presented by the monetary transfer data seem reasonable and follow expectations, especially when divided by nativity. Both native and foreign-born households send and receive transfers. However, when compared to native households, foreign-born households were more likely to send or receive transfers more often and in higher amounts. They also represent a higher proportion of those households that reported sending or receiving monetary transfers. These patterns mirror the results found by other studies comparing native and foreign-born remittance behavior (Lopez et al., 2009).

Third, the quality of the monetary transfer data benefited from being part of the CPS, a well-established survey following industry standards on all aspects of data collection and publication. Especially important for the Monetary Transfers section are: the use of a sampling frame that represents all housing units in the United States, based on the results of Census 2000 and updated regularly; the sample design, a multistage stratified probability sample tailored to provide reliable estimates representative of the national population by various demographic and labor force characteristics; the large sample size, with approximately 72,000 assigned housing units from 824 sample areas; the use of well-trained, permanent Census Bureau telephone interviewers and field representatives to collect the data from respondents; and the use of statistical weights, including population controls, to improve representation and adjust for nonresponse bias.

However, while the quality of the monetary transfer data in general appears to be good, there is some indication that there may have been misreporting of the amount of monetary transfers sent and received by both native and foreign-born households and, possibly, of the number of times sent and received. During post-interview debriefings, several field representatives reported hesitation by some respondents to report the amount sent and received (de la Cruz and Logan, forthcoming). Also, as respondents moved through the Monetary Transfers section – from the questions on whether or not the household sent and received money to the questions on the frequency and then the amount sent and received – the item nonresponse rate increased. This increase in the relative rates may reflect an increasing unwillingness by respondents to answer questions, which could indicate an increase in misreporting. As with other personal income data (Bollinger, 1998; Peterson and Kerin, 1980), misreporting usually means underreporting, so the average and total amount of monetary transfers sent and received may be underestimated.

Unfortunately, only a limited amount of research on the remittance behavior of immigrant households in the United States exists, none of which is directly comparable to the results of the Migration Supplement. For example, the U.S. Bureau of Economic Analysis (BEA) is the government agency responsible for publishing the Annual Revision of the U.S. International Accounts (Bai and Hoang, 2010) which includes an estimate of personal transfers of the foreign-born population resident in the United States to households abroad.¹¹ The BEA defines “personal transfers” as all transfers in cash or in-kind made or received by resident households to or from nonresident households. Using a model-based approach and data from the American Community Survey, the BEA estimated personal transfers by the foreign-born population resident in the United States to households abroad to be $36.9 billion in 2007, $38.5 billion in 2008, and $37.6 billion in 2009.¹²

A second study by the Inter-American Development Bank’s (IDB) Multilateral Investment Fund (MIF) focused on remittances sent by immigrants in the United States to Latin America and the Caribbean. The MIF commissioned Bendixen and Associates, a polling and research firm based in Miami, Florida, to interview 2,511 adult immigrants from Latin America by telephone in Spanish from May 3-25, 2006 (IDB, 2006).¹³ Based on these data, IDB estimated that Latin American migrants working in the United States sent $45 billion to their home countries, up from approximately $30 billion estimated by a similar study done in 2004. It is unclear if their use of the term “remittances” includes only monetary transfers (making their results more comparable with those from the Migration Supplement) or both monetary and in-kind transfers (making their results more comparable to the BEA estimates). Also, little additional information about the methodology used to develop the estimates was published by either IDB or Bendixen and Associates, making an assessment of the accuracy and representativeness of the estimates difficult.

Both the BEA and IDB estimates are considerably higher than the approximately $12 billion in monetary transfers estimated by the Migration Supplement to have been sent by both native and foreign-born households outside the United States. There are several possible reasons for these differences, including: the definitions used to define the type of resource exchange (e.g., monetary transfers, personal transfers, remittances) included in the estimates; differences in the methodologies used and the accuracy and representativeness of the estimates produced; under- or overestimation of the amount reportedly sent; among others. A more detailed comparative analysis – beyond the scope of this paper – would help shed light on the reasons for the differences among these and other estimates of remittances.