for technical support and answers to user and technical data collection issues (e.g., evaluate against the 2000 census counts.
questions. The overall objectives are to get sampling, weighting, etc.). The group pro- The housing-based estimates are based on
users into the system and make productive posed forming a task force to focus on the the U.S. Census Bureau’s current housing
use of the data. measurement of remittances and migration unit estimates (those used as ACS housing
Asked who can have access to the data in household surveys. The United States controls) based on building permits, mobile
enclave, Mulcahy explained that research- will chair the task force and host the first home data, and estimates of housing loss.
ers must submit proposals that are reviewed meeting with members, including UNECE HUBERT work includes a closer look at the
by NORC with producer input. Asked member and nonmember countries, inter- permit data and housing loss estimates, as
how much they charge for access, Mulcahy national organizations, and other experts. well as research into the potential contribu-
reported that it is $100 per week per user, The objectives are to advance methodologi- tion of counts from the Master Address File
which is about the marginal cost of hav- cal work on the measurement of remittanc- (MAF). The MAF is complicated, however,
ing an additional user on the system. The es through household surveys and develop and HUBERT seeks to identify which of sev-
objective is not to make money. international recommendations for how to eral alternative extracts provides the best basis
best collect such data. The task force also for housing-based estimates.
Measuring National and
will work to develop a list of best practices; The results suggest that (for 2000) county
International Remittances
a research agenda; and standards for data population estimates produced with the cur-
Elizabeth Grieco of the U.S. Census tabulation, publication, and dissemination. rent AdRec method perform better than those
Bureau started by describing a Conference Grieco also reported that the U.S. based on housing data. The mean absolute
of European Statisticians (CES) work plan Census Bureau is hoping to add a question percent error (MAPE) for the AdRec meth-
to improve migration statistics and an on remittances to the Current Population od was 3.2 and the MAPE for the housing-
expert group meeting on measuring remit- Survey (CPS), but still needs to get it based estimate was 5.9. Estimates based on
tances and migration. The topic was how approved. Asked if such a remittance ques- the average of the two methods had a MAPE
best to measure remittances, a term not tion would be better asked on the Consumer of 3.8. Although housing-based estimates
formally defined by the presentation, but Expenditure Survey or the American for large counties were about as accurate as
which clearly relates to transfers of money Community Survey, Grieco acknowledged AdRec estimates, the AdRec advantage holds
from immigrants back to their countries the logic of those suggestions, but noted it across many types of counties, and its esti-
of origin. is much easier to get a question added to the mates were more accurate in 70% of counties
The Conference of European Statisticians CPS. When asked what exactly constitutes (containing 61% of U.S. population).
is responsible for guiding the work of the a remittance, she indicated it includes per- The housing-unit method starts with hous-
United Nations Economic Commission for sonal transfers from migrants back to their ing totals, but uses estimates of occupancy,
Europe (UNECE), which holds work ses- country of origin, where the recipient could persons per household, and persons in group
sions about every two years and addresses be a person, church, business, or other enti- quarters to build the estimate of total popu-
issues related to the consistency and collec- ty. Other questions concerned money sent lation. So, there is concern that errors in the
tion of migration statistics. A specific CES to the origin country in exchange for some- estimation of these components may inflate
objective is to harmonize concepts and defi- thing sent in return, or cash brought (and the errors for the housing-based estimates.
nitions of remittances between balance of left) by migrants on a visit to their country Persons per household (PPH) is a particular
payments and household surveys. The most of origin. Clearly, there are definition and concern, and HUBERT has used “uncon-
recent UNECE work session on this topic measurement challenges. trolled” PPH estimates from ACS because
was in November 2006 in Edinburgh, UK. the PPH estimates reported on ACS products
Grieco noted the increased interest in
Housing Unit–Based Estimates
can be volatile and reflect the relationship
remittances and migration data, with spe-
Research Team
between sometimes inconsistent population
cific questions concerning the differences Jason Devine of the U.S. Census Bureau and housing controls (rather than the PPH of
between estimates of remittances based on explained the bureau’s Housing Unit–Based sample households). Devine described work
balance of payments data (based primar- Estimates Research Team (HUBERT) was designed to improve PPH estimates (includ-
ily on bank data) versus those based on established in response to the 2006 conference ing contract work with external research-
household surveys. and congressional hearing devoted to the U.S. ers), and it is hoped PPH estimates can be
Grieco then described a January 2008 Census Bureau’s estimates program. Among improved. However, evaluations suggest that
expert group meeting on the measurement the recommendations from these events was even a 50% improvement to PPH accuracy
of remittances using household surveys. The that the bureau consider alternative methods (combined with improvements to the hous-
meeting in Suitland, Maryland, was hosted based on housing unit data. This presenta- ing and occupancy estimates) will only bring
by the U.S. Census Bureau, UNECE, and tion was a review of HUBERT’s work to date the accuracy of the housing-based estimates
the World Bank and included participants comparing housing-based estimates to those to about the same level already achieved by
from universities and international orga- produced with the current administrative the AdRec method.
nizations. The topics covered included the records (AdRec) component method, which Looking ahead, Devine described a
types of remittance data needed by users, the tracks migration with matched IRS returns. HUBERT/estimates conference scheduled for
use of household surveys to measure remit- The basic approach is to use the two June, where further results will be presented and
tances in sending and receiving countries, methods to produce estimates for 2000 and scrutinized. Attendance will be by invitation. ■
18 AMSTAT NEWS MAY 2008
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64 |
Page 65 |
Page 66 |
Page 67 |
Page 68 |
Page 69 |
Page 70 |
Page 71 |
Page 72 |
Page 73 |
Page 74 |
Page 75 |
Page 76 |
Page 77 |
Page 78 |
Page 79 |
Page 80 |
Page 81 |
Page 82 |
Page 83 |
Page 84 |
Page 85 |
Page 86 |
Page 87 |
Page 88 |
Page 89 |
Page 90 |
Page 91 |
Page 92