180 likes | 298 Views
Estimating area variations in the take up of means-tested benefits. FRS Users meeting Royal Statistical Society 23 rd April 2007 Dominic Richardson Social Policy Research Unit. Project outline. Part of the Indices of Deprivation Income domain HMRC and DWP data on benefit receipt
E N D
Estimating area variations in the take up of means-tested benefits FRS Users meeting Royal Statistical Society 23rd April 2007 Dominic Richardson Social Policy Research Unit
Project outline • Part of the Indices of Deprivation • Income domain • HMRC and DWP data on benefit receipt • But take-up varies by… • Benefit type • Socio-demographic characteristics • Pension credit • Tax credit • Income support and income based JSA • Acorn codes / Region
Methods • CPAG benefit outlines • Benefit unit level • Checking estimates for robustness • Mismodelling statistics • Predicted versus reported • Case take-up • Bivariate case and cash • Log reg predicting likelihood of being ENR • Spatial / personal and family / entitled amount
Case and cash take up before adjustment Case take-up (percentages) Cash take-up (percentages)
Pension credit • Replaced MIG in October 2003 • Guarantee credit • minimum income • Over 60s • Savings credit • rewarding savers • Over 65s • Some pensioner receive both
Basic Requirements • Age • Marital Status • Applicable Income • Net BU income • Disregards • Capital income • Applicable amount • Minimum guarantee • Premiums • Housing costs • Entitled Amount • GC plus SC
Bivariate Results • Pension Credit case take-up varies by: • Gender • Tenure • Marital status • Disability • Equivalent income • Age • Region • Acorn type • It does not vary by: • Ethnicity • Highest educational level.
Bivariate Results • Pension Credit cash take-up varies by: • Gender* • Ethnicity • Marital status* • Age • Tenure* • Highest education qualification* • Age group • Equivalised income group* • Region • Acorn type
Likelihood of being an ENR (FRS data) • Significantly higher for: • Couples • Wealthy achiever areas • Educated Urbanity • East midlands • London • Significantly lower for: • Those with higher entitlements • Asian communities • Struggling families • Burdened singles • South west
Under-reporting • Evidence of under-reporting of benefit receipt in the FRS • May impact on the take-up estimates if bias between groups exists • So if we are to appropriate inform the ID… • Test for bias in reporting of pension credit • Pensioner type • Region / acorn code
Data matching • DWP matches administrative data to FRS data • Type of match is recorded • Name and surname • Age and sex • House number and postcode • Date (distance) matching • Entitlement groups recalculated with new receipt data • Provides adjusted estimates in cash and case take -up
Results after datamatching Case take-up estimates FRS 2004/5 Non take-up figures are inflated by 33%
Who is more likely to under-report • Females • Tenants • Non-white groups • Separated/divorced • Single pensioners • Benefit units with disabled individuals • Below degree level
What are the implications for the estimates for PC 0405? • Significantly higher: • Couples (-) • Wealthy achiever areas (+) • Educated Urbanity (+) • East midlands (-) • Secure families • Significantly lower: • Higher entitlements (-) • Asian communities (-) • Struggling families (+) • Age (older) • Settled suburbia • Prudent pensioners • Yorks and Humberside • No longer significant • Burdened singles / London / South West
Conclusions • Area variations are evident… • …however • Overall take-up is higher (under-reporting occurs) • Bias in under reporting rates by pensioner type • Bias effects the estimates of take-up • Suggestion that Income Domain of ID is not adjusted using take-up estimates • Models only explain 17-19 percent of the variation • Only bias in PC so far explored • Further analysis on bias in reporting take-up using Tax Credit receipt may be possible