Supplementary memorandum from HM Revenue
and Customs
At the PAC meeting on 24 January 2005 in respect
of the C&AG's 2003-04 Inland Revenue Standard Reporting featuring
Tax Credits I agreed to provide the Committee with a note this
month on the demographics of the Tax Credit claimants with overpayments
and errors. I would also like to inform you of some indicative
findings about the level of Tax Credit claimant compliance.
DEMOGRAPHICS OF
OVERPAYMENTS
The Committee asked for a note outlining the
groups of people who have been most affected by overpayments and
errors and the demographic analysis of these groups.
The analysis below shows families with overpaid
awards. The first column covers all overpayments, whether caused
by official error or by changes in income or other circumstances.
In respect of errors, there are limits to the types of error that
we can clearly identify. In particular, no demographic analysis
is available of the 455,000 families affected by the computer
system fault described at paragraph 2.10 of the C&AG's 2003-04
Inland Revenue Standard Report. We have, however, been able to
identify families affected by certain other known computer system
errors. The figures cannot be added together as some families
can appear in more than one column.
A Families for which the computer based the award
calculation on less than their full income due to a system error,
whether or not they disputed the overpayment later;
B Families who reported their annual incomes
when renewing their awards but this income was incorrectly disregarded
for subsequent years, leading to under or overpayments. Includes
all such families, whether or not they became aware of the error
or contacted us;
C Families where there has been a system
error identifying the amount of overpayment recovered from the
next year's payments. Includes all such families, whether or not
they became aware of the error or contacted us.
| * | Strictly speaking, these are counts of overpaid awards, not families.
|
| ** | lncludes awards for families that had qualifying children earlier in the year but, for example, the last child had left full time non-advanced education during the year.
|
NB some figures have been rounded, therefore some columns
do not add up.
CLAIMANT COMPLIANCE
You will recall that to measure the general level of compliance
in tax credits we are carrying out an annual programme of random
enquiries on a statistically representative sample of finalised
claims across the tax credit population. As with other "random
enquiry" exercises, its main purpose is to:
— measure the proportion of claimants that
are non-compliant;
— measure the financial consequences of non-compliance;
— measure the effectiveness of our automated
risk assessing processes ie factors, or combinations of factors,
in a claim that denote a high risk of noncompliance; and
— use the results to help to refine our risk
assessment processes for the future by developing new ways to
identify high risk claims.
We randomly selected a sample of about 4,700 awards. Because
of the size and diversity of the claimant population, and the
possible variations of compliance risk, we stratified the sample
by type of claim so that we could measure the level of compliance
for various claimant groups, as well as for claimants as a whole.
The sample was designed to achieve an overall precision of ±
1.4% at worst (with 95% confidence) in the resulting estimate
of the proportion of compliant cases. This was based on a set
of assumptions about the proportions of non-compliance that would
be found in each stratum.
We are carrying out a full check on each claim in the selected
sample to establish whether there was any non-compliance, and
where we identify some, we are measuring the financial consequences
of that non-compliance. The random enquiry programme could not
start until recipients had provided the Department with details
of their final 2003-04 incomes, which meant that we were unable
to start work on some cases until after 31 January 2005.
Due to the time needed to complete these investigations,
particularly when the claimant is self employed (we need to tie
in this investigation with a check of their income reported on
their Income Tax Self Assessment), we do not expect to publish
final results for 2003-04 until Spring 2006. Whilst some investigations
can be dealt with relatively quickly, ie claimants on PAYE where
income can easily be cross checked, some will take longer to resolve,
especially when we need to obtain information from third parties
eg childcare providers. But we now have indicative findings. These
come with a strong warning that they are subject to a wide margin
of error. This is because they are only based on the subset of
the sample cases where a full enquiry had been completed by mid-May
2005. It is likely that these cases were the more straightforward
(and probably more compliant, or at least less likely to be deliberately
non-compliant) ones. This means the full results are likely to
see an increase in the proportion of non-compliant cases and the
financial consequences estimates. The deliberate error proportion
and associated financial cost estimate are also likely to rise.
INDICATIVE RESULTS
The 1,385 cases where the checks had been completed by 15
May 2005 provide the basis of the indicative results. This represents
only about 29.6% of the total sample. After grossing the sample
up to the tax credits population, we estimate that 13.1% of the
overall population of some 5.3 million awards (excluding out of
work cases) are non-compliant. The associated financial consequence
of this noncompliance is £460 million per annum, some 3.4%
by value.
Where it was established that the claim was wrong, tax credit
compliance officers were asked to indicate whether the non-compliance
was due to a deliberate attempt by the claimant to receive tax
credits to which they were not entitled or due to errors in the
claim that that were not deliberate. In only 0.3% of awards examined
and closed to date was the error deemed to be deliberate. The
financial cost associated with this is estimated to be £30
million. The remainder is due to other errors on the part of the
claimant. We would expect the relative proportions to change by
the time the exercise is completed as these results are only indicative
and because the cases that are taking longer to complete may fall
into different categories. It is reasonable to assume that cases
settling quicker tend to be more straightforward so the proportion
of cases that are not compliant is likely to rise when the full
data is available.
As this is the first random enquiry exercise on tax credits,
we cannot use the results from previous exercises to forecast
what these results might be once the full sample has been worked.
FINAL REPORT
We are aiming to complete the investigation of all cases
within the sample as soon as possible and we expect the 2003-04
random enquiry program to be completed in January 2006. Quality
checking and full analysis of the final data will take time to
complete so we expect to produce a final report by Spring 2006.
Once we have established a baseline figure for the level of non-compliance
in tax credits, our intention is to target year on year a reduction
in this figure beginning in 2006-07.
OTHER COMPLIANCE
RISK MONITORING
WORK
As well as putting in place the random enquiry programme,
there was a clear business need to obtain evidence about the types
of non-compliance and the effectiveness of our risk assessment
system before July 2005. To that end, we carried out an exercise
designed to inform our risk assessment process by:
— looking at the effectiveness of the risk
rules which have been developed to support tax credits and protect
the system from abuse;
— provide an indication of the level and type
of non-compliance around claimant circumstances; and
— support and inform the compliance regime
for tax credits.
The research involved reviewing a small statistically valid
sample of 1,000 provisional (not final) awards for 2003-04. As
these claims were provisional, we could not use our random investigation
powers, so where we were unable to identify a risk from the original
claims they were not investigated. We then carried out examinations
on those claims where we had identified a risk. But as the sample
cases were provisional awards, and overpayments are a normal part
of the tax credits system (eg arising from changes in income and
circumstances that claimants are not required to report until
the end of the year), the results could not, nor were intended
to, provide a robust, comprehensive measure of the levels of claimant
error or the financial impact of those errors. The random enquiry
programme was designed to do this.
The main risk areas confirmed by that research related to
incorrect childcare costs, undeclared partners and income discrepancies.
The work confirmed that our risk scoring system is effective in
identifying these risk areas and differentiating between high
and low risk claims. This confirmed our view that the risk scoring
system is working as planned and proving valuable in targeting
investigative activity by our compliance teams on the most appropriate
cases.
The research found that the proportion of non-compliant claimants
lay in the range 5%-11%. Where it was established that the claim
was wrong and that the error arose from non-compliance, the reason
for the error was recorded as either deliberate error or negligence.
Only 2.2% of the cases deemed non compliant was due to deliberate
error.
COMPARISONS OF
LEVELS OF
NON COMPLIANCE
IN WFTC/DPTC AND
CTC/WTC
Child and Working Tax Credits (CTC/WTC) are very different
in scale and scope compared to their predecessors, Working Families'
Tax Credit (WFTC) and Disabled Persons Tax Credit (DPTC) because
of the annual nature of awards and because they are designed to
react to any changes in personal circumstances during the period
of an award. Also, significantly more families benefit from them,
including a large proportion that are only receiving the family
element. These fundamental differences do not lend themselves
to direct comparison of levels of non-compliance between the two
schemes.
David Varney
Chairman
18 July 2005
| 
