Guidance

Housing Benefit: caseload management information overview

Published 27 January 2016

What is caseload management information?

Caseload management information is an estimate of how well each Local Authority is managing their Housing Benefit (HB) caseload. By identifying the changes of circumstances that lead to a reduction to entitlement, Local Authorities can prevent or minimise overpayments from occurring, thereby minimising Fraud and Error.

The measure compares the value of the reductions to entitlement identified by each Local Authority against the value of reductions we expect each Local Authority could potentially identify, given the size and composition of its HB caseload. This estimate is presented as a percentage.

How is caseload management information derived?

The two components (actual reductions achieved and expected reductions) which make up the Caseload Management Information are derived as follows:

Actual reductions achieved

Local Authorities send DWP customer-level data electronically each month in the Single Housing Benefit Extract (SHBE). SHBE provides information on current live HB claims, including entitlement details. By comparing the weekly HB entitlement of individual claims in consecutive monthly SHBE extracts we can identify where the weekly HB entitlement has reduced. The actual reduction achieved by a Local Authority is the sum of the value of these reductions in entitlement observed each month for that Local Authority.

Expected reductions

The total value of reductions that we expect a Local Authority could find is dependent on the size of its individual caseload, the types of claimants in that caseload and the value of the reductions that tend to occur in that caseload, which will be partially a consequence of the rent levels found in that Local Authority.

The method we use to estimate the level of reductions that we expect each Local Authority should be able to find involves four stages:

1. Breaking down every Local Authority’s caseload into different claimant types, based on their propensity for undergoing a reduction to entitlement.

2. Applying an adjustment factor to each Local Authority to reflect the different values of reductions that tend to occur in their caseloads, to bring them into line with each other and enable comparison.

3. For each claimant type, setting the expected level of reductions per claimant by finding the 90th percentile of reductions per claimant identified across all Local Authorities in the past1, using the adjustment factors to enable comparison across all Local Authorities.

4. Translating those expected levels of reductions per claimant, for each claimant type, into a total expected level for each Local Authority.

The detailed steps we take are as follows:

Stage 1):

We first segment each Local Authority’s HB caseload, using a risk model which groups claimants by the likelihood of having an overpayment. Since the vast majority of overpayments are a result of changes of circumstances, this is a way of grouping claimants according to their propensity to undergo a change in circumstance leading to a reduction to entitlement.

Stage 2):

Using the historical data^{[footnote 1]}), we then divide the total value of the reductions found in each risk group by the number of claimants in that risk group, to calculate the average value of a reduction per claimant in each Local Authority. Diagram 1, below, outlines this calculation:

Diagram 1: Calculating the average reduction per claim in the historic data

Calculation: Total value of reductions divided by Risk group 1 caseload = Reduction per claim

LA	Risk group 1 caseload	Total value of reductions	Reduction per claim
1001	500	£4,000	£8.00
1002	1,300	£18,000	£13.85
1003	2,800	£15,000	£5.36
1004	1,600	£11,500	£7.19
1005	2,600	£16,000	£6.15
1006	600	£4,500	£7.50
1007	900	£6,100	£6.78
1008	1,700	£6,600	£3.88
1009	3,800	£18,000	£4.74
1010	2,400	£22,500	£9.38

Before finding the 90th percentile, we first need to account for variations in the value of reductions that tend to occur across Local Authorities, which will be partially a consequence of different rent levels. In practice, this will affect reductions to entitlement which take the claimant off HB altogether; whereas the value of reductions which decrease entitlement to a lower level of HB will not usually be a consequence of the level of HB entitlement.

We therefore split each Local Authority’s reductions to entitlement into those which lead to a termination of benefit and those which just lead to a decrease in benefit. For those which lead to terminations, we adjust the reduction per claim found by each local authority in accordance with the relative average value of HB entitlement in that local authority. To do that, we calculate the average Housing Benefit award^{[footnote 2]} in each risk group. The proportional difference between the Local Authority’s average HB award and the national average HB award gives us the adjustment factor for that Local Authority. We then adjust the reduction per claim figure for terminations by dividing it by the adjustment factor. For those reductions which do not lead to terminations, we do not apply any adjustment factor the reduction per claim figures.

Stage 3):

We then set the expected level of reductions by finding the 90th percentile of reductions to entitlement per claim identified across all Local Authorities in the historic data. For the terminations, we use the adjusted reduction per claim figures. This is illustrated in Diagram 2, below.

Diagram 2: Finding the 90th percentile

LA	No. of reductions	Average reductions	Adjustment factor	Reduction per claim	Reduction per claim (adjusted)
1001	100	£30.00	1.5	£6.00	£4.00
1002	400	£35.00	1.8	£10.77	£6.15
1003	600	£20.00	1.0	£4.29	£4.29
1004	500	£18.00	0.9	£5.63	£6.25
1005	900	£14.00	0.7	£4.85	£6.92
1006	200	£17.00	0.9	£5.67	£6.67
1007	300	£16.00	0.8	£5.33	£6.67
1008	250	£20.00	1.0	£2.94	£2.94
1009	1200	£12.00	0.6	£3.79	£6.32
1010	1000	£18.00	0.9	£7.50	£8.33
RG1		£20.00

Value of reductions achieved over number of reductions

Average in each LA divided by the average reduction across all LAs (£20)

Reduction per claim

Reduction per claim divided by adjustment factor

Result: 90th percentile (adjusted) = £7.06

Stage 4):

To apply the expected level of reduction per claim to each Local Authority, we multiply it by the most recent caseload estimates to produce a total expected reduction figure. For the terminations, we then multiply this by the most recent adjustment factors^{[footnote 3]}, in order to align the expected value back to each Local Authority’s original level. Diagram 3, below, illustrates this calculation.

Diagram 3: Calculating expected reductions for each risk group

LA	Risk group 1 caseload	Expected reduction per claim (adjusted)	Adjustment factor	Expected (re-adjusted back)
1001	500	£7.06	1.46	£4,649
1002	1,300	£7.06	1.7	£14,429
1003	2,800	£7.06	1.03	£18,835
1004	1,600	£7.06	0.96	£11,508
1005	2,600	£7.06	0.68	£11,591
1006	600	£7.06	0.81	£4,275
1007	900	£7.06	0.84	£5,344
1008	1,700	£7.06	1.05	£12,228
1009	3,800	£7.06	0.59	£16,211
1010	2,400	£7.06	0.88	£14,964

90th percentile (adjusted) = £7.06

Calculation: Number of claims in risk group x expected reduction per claim (£7.06) x adjustment factor

We are then able to sum up all the expected values across the risk groups to produce a total expected value for each Local Authority. Diagram 4, below, demonstrates how we then compare the total expected value of reductions to the actual value of reductions achieved, to arrive at the percentage figure.

Diagram 4: Summing across all the risk groups (RG)

LA	RG1 expected	RG2 expected	RG3 expected	RG4 expected	etc	Total expected	Total actual	Actual / expected
1001	£5,759	£7,506	£8,677	£8,504	-	£164,489	£85,534	52%
1002	£17,824	£22,770	£26,322	£25,795	-	£498,966	£399,173	80%
1003	£20,848	£28,023	£32,394	£31,746	-	£614,086	£343,888	56%
1004	£13,631	£14,412	£16,660	£16,327	-	£315,817	£255,812	81%
1005	£14,287	£18,216	£21,057	£20,636	-	£399,173	£359,256	90%
1006	£5,885	£5,104	£5,900	£5,782	-	£111,844	£97,304	87%
1007	£7,056	£7,206	£8,330	£8,163	-	£157,909	£135,801	86%
1008	£15,124	£17,014	£19,669	£19,275	-	£372,850	£141,683	38%
1009	£18,622	£22,819	£26,379	£25,851	-	£500,049	£410,040	82%
1010	£17,888	£21,618	£24,990	£24,490	-	£473,726	£213,177	45%

1. The past period we have used is the 12months immediately preceding the introduction of the Fraud and Error Reduction Incentive Scheme (FERIS), to ensure that the levels of reductions found were not temporarily inflated as a consequence of changes in activity prompted by the introduction of FERIS. ↩

2. In adjusting the historical reduction per claim figure we use the annual average HB award in the twelve months of that period (i.e. the average HB award from October 2013 to September 2014). ↩

3. The adjustment factor used to scale the expected value of changes back up for each local authority is derived using the latest financial year’s average HB entitlement in that local authority. For example, in order to scale up the 2015/16 expected values we use the 2014/15 average HB entitlement. ↩