Corrections to the Imputed Variables Related to Housing Equities in AHEAD 1995, and HRS 1996, 1998, and 2000 Imputation Data
September 9, 2003
Problems have been found with the imputations for primary home variables in HRS 1996, 1998, and 2000 data, and for the primary and second home variables in AHEAD 1995 data. The basic problem is that we have imputed home values for the households whose answers to the home ownership questions (including questions about farm/ranch, mobile home, and single-family house, apartment, and townhouse) were neither "own" nor "rent", but "other". We now believe that those households should be treated as "not-owning home", and, thus, should have a value of "0" (zero) on all the variables related to the home.
The variables affected by this data alert include the imputed values of the primary home, mortgage balance and payment, and real estate tax in all the waves identified, as well as the imputed values of the second home, mortgage balance and payment,and real estate tax in AHEAD 1995 (see Table 1 below). The constructed net worth variables in the HRS 1998 and 2000 wealth sections are also affected.
Table 1. Housing Variables Affected by the Data Alert | ||||
Concept of Variable | Names of the Variables Affected | |||
AHEAD95 | HRS96 | HRS98 | HRS00 | |
Primary Home Variables | ||||
Primary Home Equity | DHOME1 | EHOME1 | FHOME1 | GHOME1 |
Primary Home Value | D2246X | E2246X | F2753X/F2760X | G3071X/G3078X |
1st Mortgage Owned | D2256X | E2256X | F2773X | G3091X |
Payment on 1st Mortgage | D2252X | E2252X | F2769X | G3087X |
2nd Mortgage Owned | D2263X | E2263X | F2780X | G3098X |
Payment on 2nd Mortgage | D2260X | E2260X | F2777X | G3095X |
Other Loans Owned | D2270X | E2270X | F2787X | G3105X |
Payment on Other Loans | D2267X | E2267X | F2784X | G3102X |
Real Estate Tax | D2292X | E2292X | F2809X | G3127X |
Second Home Variables | ||||
Second Home Equity | DHOME2 | |||
Second Home Value | D2435X | |||
Mortgage Owned | D2445X | |||
Payment on the Mortgage | D2441X | |||
Real Estate Tax | D2480X |
The problem can be corrected fairly simply by modifying the imputation data according to the answers to the home ownership questions (including questions about farm/ranch, mobile home, and single-family house, apartment, and townhouse) in the core.
The detailed correction statements follow.
(1) Corrections to the Variables Related to Housing Equities in AHEAD 1995
if D2226=7 or D2229=7 or D2234=7 then do; DHOME1=0; D2246X=0; D2256X=0; D2252X=0; D2263X=0; D2260X=0; D2270X=0; D2267X=0; D2292X=0; end; if D2409=7 or D2412=7 or D2417=7 then do; DHOME2=0; D2435X=0; D2445X=0; D2441X=0; D2480X=0; end;
(2) Corrections to the Variables Related to Housing Equities in HRS 1996
if E2226=7 or E2229=7 or E2234=7 then do; EHOME1=0; E2246X=0; E2256X=0; E2252X=0; E2263X=0; E2260X=0; E2270X=0; E2267X=0; E2292X=0; end;
(3) Corrections to the Variables Related to Housing Equities in HRS 1998
if (F2743=7 and F2746=. and F2751=.) or (F2743=. and F2746=7 and F2751=.) or (F2743=. and F2746=. and F2751=7) then do; FHOME1=0; F2753X=0; F2760X=0; F2773X=0; F2769X=0; F2780X=0; F2777X=0; F2787X=0; F2784X=0; F2809X=0; end;
(4) Corrections to the Variables Related to Housing Equities in HRS 2000
if (G3061=7 and G3064=. and G3069=.) or (G3061=. and G3064=7 and G3069=.) or (G3061=. and G3064=. and G3069=7) then do; GHOME1=0; G3071X=0; G3078X=0; G3091X=0; G3087X=0; G3098X=0; G3095X=0; G3105X=0; G3102X=0; G3127X=0; end;
To give users a sense of the scale of the changes related to this data alert, Table 2 compares the mean values and standard deviations for the relevant housing equity and networth variables before and after the corrections.
Table 2. Housing Equity and Networth before and After Corrections | |||||
AHEAD95 | HRS96 | HRS98 | HRS00 | ||
Mean | Mean | Mean | Mean | ||
PRIMARY HOME EQUITY |
Before Corrections |
69,523 (115,880) |
70,952 (95,840) |
77,310 (217,451) |
87,737 (216,927) |
After Corrections |
64,247 (115,191) |
69,331 (95,737) |
73,562 (217,177) |
83,160 (216,374) |
|
SECOND HOME EQUITY |
Before Corrections |
3,749 (30,688) |
|||
After Corrections |
2,885 (27,479) |
||||
NET WORTH |
Before Corrections |
281,025 (1,039,409) |
326,587 (973,878) |
||
After Corrections |
277,276 (1,039,764) |
322,010 (974,439) |
|||
Note: Standard deviations in parentheses. |