Free DP Download 19 December 2019 - THE TRUE COST OF AUSTERITY FOR THE HOMELESS AND LOW-INCOME HOUSEHOLDS: UK evidence
HOUSING INSECURITY, HOMELESSNESS AND POPULISM: Evidence from the UK
Thiemo Fetzer, Srinjoy Sen, Pedro CL Souza
CEPR DP No. 14184 | 09 December 2019
UK public spending cuts that substantially lowered housing benefits aimed at helping low income households pay rent amounted to an annual reduction equivalent to a £470, rising to as much as nearly £2,400 in many parts of London. The implications of this policy had high social and human costs, as it substantially increased evictions, individual bankruptcies, property crimes, temporary accommodation, statutory homelessness and rough sleeping.
These are the findings of a new CEPR study by Thiemo Fetzer, Srinjoy Sen and Pedro Souza, which investigates the consequences of a UK austerity policy that attempted to reduce the fiscal cost of providing housing assistance to low income households. Before April 2011, local housing allowance covered up to the median level of market rents; from April 2011 onwards, only rents lower than the 30th percentile were covered.
On top of this, the actual fiscal savings of the cut are much smaller than anticipated: for every pound saved by the central government, council spending to meet statutory obligations for homelessness prevention increases by 53 pence.
The study also finds that the policy had significant effects on representative democratic participation in UK elections: in the most affected districts, electoral registration rates dropped markedly. This finding is reproduced during the 2016 EU referendum vote, where the authors find evidence that the turnout was substantially lower. The study also finds that the support for Leave was higher in those places, which is possibly driven by a composition effect on the electorate since the proclivity to vote Remain was substantially higher among those who did not turn out to vote on that occasion.
Figure 8: Impact of change in reference rent on measures of statutory homelessness
Notes: Figure plots from the regression studying the impact of the cut to local housing allowance to cover the median rent to only cover the 30th percentile of rents from April 2011 onwards. All dependent variables are measured as rates relative to the number of resident households in a district. The dependent variable in Panel A measures the number of statutory homeless individuals. Panel B is the street count of rough sleepers. All regressions control for local authority district fixed effects and year fixed effects. 90% confidence bands obtained from clustering standard errors at the district level are indicated.