VARIABLE LIST FOR  
Dirk W. Early and Edgar O. Olsen (2002) 
"Subsidized Housing, Emergency Shelters, and Homelessness: 
An Empirical Investigation Using Data from the 1990 Census", 
Advances in Economic Analysis & Policy: Vol. 2: No. 1, Article 2.
http://www.bepress.com/bejeap/advances/vol2/iss1/art2 

DATA SET IS Early-Olsen.dta 

VARIABLE LIST (SEE ARTICLE FOR MORE DETAIL) 

shelter		Sheltered homeless 
street		Street homeless 
poor2 		Poverty population 
homepov2 	Total homelessness 
shltpov2 	Sheltered homelessness 
strtpov2 	Street homelessness 
fracshlt 	Percent of homeless in shelters 
fbedpov2	Shelter beds per poor person  
fitbdpov 	Shelter beds per poor person (fitted) 
fbedhome 	Shelter beds per homeless person 
fitbdhom 	Shelter beds per homeless person (fitted) 
rshltql 	Shelter quality 
hpi_r 		Price of rental housing service 
vr 		Rental vacancy rate 
rc 		Rent control 
rafdc 		AFDC 
rga 		GA 
rmcapov 	McKinney Act funding 
ur 		Unemployment rate 
halfpov2 	Percent very poor 
blackpov 	Percent of poor African-American 
fempov 		Percent of poor female 
avetemp 	Average March temp 
temp 		Temp during S-Night count 
averain 	Average March precipitation 
precip		Precipitation during S-Night count 
rmental 	Spending on mental hospitals 
unitpov2 	Availability of subsidized housing 
less5k 		Subsidized housing targeted toward very poor 
crime 		Crime rate 
neast 		Northeast 
south 		South 
west 		West 
ada 		Liberal political climate 
rpland 		Price of land 
rboeckh		Building cost index 
rmedinc 	Median income 
per1000 	Population of the metropolitan area 
ngov92 		Number of governments 
ngovcap		Number of governments per capita 

For some variables, a variable was created to control for missing observations. 
Original variable was coded 0 if missing and renamed with a “z” proceeding the variable name. 
The new variable is coded 1 if missing, 0 otherwise and named with “z01".  
For example, zless5k = 0 if less5k missing, = less5k otherwise, 
z01less5 = 1 if less5k missing, z01less5k = 0 otherwise