Annex A
PUBLIC PERFORMANCE TARGET: REMOVING MORE
FAILED ASYLUM SEEKERS THAN NEW ANTICIPATED UNFOUNDED APPLICATIONS
— During Q2 2006 (April to June), 5,070
failed asylum seekers, including dependants, were removed[8]
compared with an estimated 4,185 applicants in Q2, including dependants,
who it is predicted will fail to be granted refugee status or
other leave. This represents an improvement on the performance
for previous quarters, reflecting increased removals and falling
application figures.
| Period | Applications
(including
dependants)
| Percentage
predicted to
prove to be
unfounded
| Number
(including
dependants)
predicted to be
unfounded
| Removals (including dependants)
TotalNumber (out of
total) who left the
UK without informing the Immigration Authorities
| Performance (removals/predicted unfounded applications)
|
| Q1 2004 | 10,730 | 74%
| 7,910 | 4,090 | —
| 52% |
| Q2 2004 | 9,305 | 77%
| 7,135 | 3,790 | —
| 53% |
| Q3 2004 | 10,485 | 76%
| 7,980 | 3,595 | —
| 45% |
| Q4 2004 | 10,100 | 76%
| 7,660 | 3,435 | —
| 45% |
| Q1 2005 | 8,370 | 75%
| 6,265 | 3,485 | 15
| 56% |
| Q2 2005 | 7,270 | 73%
| 5,305 | 3,765 | 110
| 71% |
| Q3 2005 | 7,806 | 71%
| 5,545 | 4,225 | 195
| 76% |
| Q4 2005 | 7,395 | 68%
| 5,010 | 4,210 | 145
| 84% |
| Q1 2006[9]
| 7,530 | 63% | 4,730
| 4,930 | 275 | 104%
|
| Q2 2006 | 6,380 | 66%
| 4,185 | 5,070 | 235
| 121% |
| Of which: | |
| | | |
|
| April 06 | 2,050 | 65%
| 1,335 | 1,735 | 85
| 130% |
| May 06 | 2,210 | 66%
| 1,455 | 1,670 | 100
| 115% |
| June 06 | 2,120 | 66%
| 1,400 | 1,665 | 55
| 119% |
| | |
| | | |
BASIS OF
CALCULATION OF
TARGET
The target calculation is based on a comparison, for a given
period, of:
(a) the number of unsuccessful asylum seekers (including
dependants) removed with
(b) the number of new applicants (including dependants)
who it is predicted will fail to be granted refugee status or
other leave, after the cases have gone through the initial decision
and appeal process (first two tiers), based on recent rates for
grants at initial decision, the proportion of those refused that
appeal, and the proportion allowed at appeal.
Further details are given in the attached technical note.
TECHNICAL NOTE: CALCULATING PERFORMANCE AGAINST THE TIPPING
THE BALANCE TARGET
INTRODUCTION
1. The Tipping the Balance target states that the number
of failed asylum applicants removed each month exceeds the number
of new asylum applicants who, it is predicted, will not be granted
leave to remain in the UK, as a result of their asylum application.
2. In order to measure progress, a methodology has been
developed to predict the proportion of new asylum applicants that
will fail to be granted any form of leave (refugee or other status),
up to and including the 2nd appeal tier. The method uses historical
data to calculate the percentage of arriving asylum applicants
that are predicted to go on to fail to be granted leave. The methodology
is being kept under review with further work being undertaken,
including on calculating the proportion of withdrawals, assessing
whether a bias adjustment is needed for appeal determinations
and assessing whether improved predictions would arise from using
cohort-specific data[10].
3. This technical note describes:
— what constitutes an unfounded application;
— the calculations used to estimate the proportion
of applications that are predicted to be unfounded;
— what constitutes a removal of a failed asylum
seeker;
— assumptions used in the calculations;
— ongoing reporting against the target;
— reliability of the asylum statistics;
— comparisons of predicted number of applications
estimated to be unfounded with actual figures; and
DEFINITION OF
AN UNFOUNDED
ASYLUM APPLICATION
4. For the purpose of measuring this target, an unfounded
asylum application[11]
is one that subsequently results in an applicant having no status
in the United Kingdom, eg refugee status, Humanitarian Protection
(HP) or Discretionary Leave (DL). It includes any dependants of
that person. The definition includes applicants (and their dependants)
who:
— have withdrawn their application for asylum[12]
or
— are refused asylum and do not appeal against their
refusal or
— are refused asylum and who have no in country
right of appeal against their refusal (NSA[13]
and Third Country cases[14])
or
— are refused asylum and submit an appeal and that
appeal is not allowed and is not withdrawn by the Home Office[15].
This does not include applicants (and their dependants) who
are granted any form of leave in whatever category regardless
of whether their asylum application was refused at initial decision
(eg those who are unaccompanied minors granted discretionary leave).
CALCULATION USED
TO ESTIMATE
THE PROPORTION
OF APPLICATIONS
THAT ARE
PREDICTED TO
BE UNFOUNDED
5. The predicted number of unfounded applications in,
say, June 2006 is based on:
(a) the published number of applications in June 2006,
including dependants;
(b) the predicted proportion that will fail to be granted
refugee status or other leave, after the cases have gone through
the initial decision and appeal process (first two tiers), based
on historical data[16]
(an example calculation is provided in Annex A).
DEFINITION OF
A REMOVAL
OF A
FAILED ASYLUM
SEEKER
6. All removals, including dependants, where the person
being removed, or at least one member of the family being removed,
has at some point during their time in the United Kingdom submitted
an application for asylum. This includes:
— Third Country cases where the application for
asylum is not accepted;
— voluntary returns and voluntary assisted returns;
— those who it is established have left the United
Kingdom without informing the immigration authorities (eg detected
at UK embarkation controls or identified as having left the UK
when they apply for entry clearance at a British Mission overseas).
7. IND has developed operational systems for identifying
persons in the last category.
8. To aid with interpretation of trends, the numbers
where it is established the person has left the United Kingdom
without informing the immigration authorities are presented separately.
ASSUMPTIONS USED
IN THE
CALCULATIONS, AND
THEIR JUSTIFICATIONS
9. The predicted proportion that will fail to be granted
refugee status varies over time depending on trends in the proportions,
by nationality, that are successful at initial decision or at
1st or 2nd appeal tier. Annex B illustrates the stages at which
an asylum application can become unfounded.
10. For initial decisions on applications in June 2006
(say) the calculations make a number of assumptions, which are
outlined below.
(a) Cross-sectional data for the period October 2005 to
March 2006 can be used as the basis for predicting the proportion
of applications in June 2006 that will prove to be unfounded.[17]
The period of October 2005 to March 2006 has been selected as
recent data should be most representative of June 2006 performance.
This six-month period will permit more accurate estimates than
January to June 2006 as it allows for late-reported cases to be
submitted. A six-month period was chosen in order to reflect averages
over a sufficiently long period so that the assumptions are sufficiently
robust and stable and the impact of seasonal (or other) differences
is reduced. Historical calculations for earlier application periods
use the most recent six-month period. Further details are in Annex
C.
(b) The calculations have been performed separately for
the "top 20"[18]
nationalities and for all other nationalities grouped together
so that subsequent changes in the distribution of applications
by nationality can be accommodated. In particular, unfounded rates
vary considerably by country, as does the volume intake, and so
the mix of nationalities can have an impact on the unfounded rate.
(c) The proportion of applications withdrawn prior to
initial decision is 3%. This is based on analysis of outcomes,
where available, for applications made in the first half of 2005
(January to June). Withdrawal data tend to get revised and, in
addition, recent months are thought to have been affected by the
Family ILR exercise[19].
Therefore, in preference to the October 2005 to March 2006 period,
alternative more stable and robust data, more relevant to new
applications being made in June 2006, have been used for this
assumption.[20]
(d) To be consistent with the published asylum application
figures, where an applicant makes more than one application for
asylum, only the first will be included: all subsequent applications
will be excluded.
11. In the cross-sectional data on decisions for October
2005 to March 2006, used to estimate the proportion and number
unfounded, there are an atypically high proportion of applications
that have taken a long time to decide. The latter have a low proportion
that are unfounded, which results in the per cent unfounded understating
the probability of any particular application being unfounded.
Hence, to address this bias, one needs to compare the unfounded
per cent among the determinations in October 2005 to March 2006
with a version of that percentage which removes the possible over-representation
of decisions that have taken a long time to reach.
12. To accommodate these effects, the proportion of initial
decisions in October 2005 to March 2006 has been re-weighted for
each nationality group to reflect the distribution of the time
taken from application to an initial decision in the January to
June 2003 cohort. Similar calculations have been made for earlier
periods.
13. For cases that subsequently go through the appeal
process, the calculations are based on data for a cohort of refusals
to estimate the proportion of refusals for which an appeal is
received, and a cohort of first tier appeal determinations to
estimate the proportions of first tier appeals that had the appeal
allowed at either the first or second tier. As it can take up
to nine months for an appeal to be received and determined, the
calculations for June 2006 use the cohort of refusals in July
to December 2005 and the cohort of first tier appeal determinations
in July to December 2005. Again the calculations have been performed
separately for the "top 20" nationalities and for all
other nationalities grouped together. Details of the cohorts used
for earlier periods are given in Annex C.
14. The number of cases subject to Judicial Review is
proportionally small, and omitting these does not affect the robustness
of the calculation.
SUCCESS CRITERIA
15. The target will have been achieved if:
— the number of failed asylum seekers (including
dependants) removed in a particular month
exceeds
— the number of applicants (including dependants
) in that month who it is predicted will fail to be granted refugee
status or other leave, after the cases have gone through the initial
decision and appeal process (first two tiers).
ONGOING REPORTING
OF PROGRESS
AGAINST THE
TARGET
16. It is currently planned that progress against the
target will continue to be published on a quarterly basis at the
same time as the asylum statistics. This will include any improvements
to the prediction methodology and revisions to the removal figures
for those it is established have left the UK without informing
the immigration authorities.
RELIABILITY OF
THE ASYLUM
STATISTICS
17. The asylum statistics were assessed by the National
Audit Office (May 2004) as "generally reliable".[21]
COMPARISONS OF
PREDICTED NUMBER
OF APPLICATIONS
ESTIMATED TO
BE UNFOUNDED
WITH ACTUAL
FIGURES
18. Cross-sectional analyses of initial decisions have
formed the basis of the calculations of the predicted number and
proportion that subsequently prove to be unfounded. Any cohort-based
approach is likely to add further complexity to the approach adopted
and, in addition, two specific issues were identified that led
to the continued adoption of the cross-sectional approach.
(a) The size of the backlogs, where a decision is awaited.
For initial decisions this grows from 250 for applications in
Q1 2004 to 500 in Q1 2005 and 1,300 in Q4 2005. The corresponding
figures for those awaiting a 1st tier appeal determination are
100, 195 and 1,675.
(b) The distribution of decision outcomes varies with
the length of time from initial application. This suggests that
the modelling required for recent cohorts, particularly for the
three monthly 2006 cohorts will be very complex and needs to take
account of the length of time since initial application. Given
the size of the backlogs, and the fact that the distribution of
backlogs will differ (eg for the 2006 cohorts the backlog will
be mainly among initial decisions) the resultant estimates of
the number/proportion unfounded will potentially be imprecise
and/or subject to biases.
However, despite these difficulties we will continue to work
on this approach to understand the sensitivities of the outputs
to these changes in data used for the calculations.
19. This is less of a problem for the 2004 cohorts, where
the backlogs awaiting a decision are proportionally very small.
A comparison of the predicted and actual numbers that were unfounded
for each of the quarters in 2004 is given below. It suggests that
for this period the prediction methodology described in this note
over estimates the number and proportion of unfounded applications.
| | Number unfounded
| | | |
|
| | Predicted
| Cohort analysis
|
| Period | Applications | No.
| % | No. | % |
Backlog |
| Q1 2004 | 10,730 | 7,910
| 74% | 7,635 | 71%
| 295 |
| Q2 2004 | 9,305 | 7,135
| 77% | 6,925 | 74%
| 305 |
| Q3 2004 | 10,485 | 7,980
| 76% | 7,520 | 71%
| 425 |
| Q4 2004 | 10,100 | 7,660
| 76% | 7,010 | 69%
| 480 |
| | |
| | | |
REVISION OF
FIGURES
20. The Q4 2005 and Q1 2006 appeal figures were updated
using data for Jul-Dec 2005 (as opposed to April-Sept 2005) as
enough time has lapsed to allow late reported cases to be submitted.
For the same reason Q1 2006 initial decisions were updated using
data for Oct 2005-Mar 2006 (as opposed to Jul-Dec 2005).
21. Other figures revised for the 2005 period are due
to late entered data and removing duplicate cases. Revised removals
data for 2005 has also been used following data cleansing exercises.
Annex A
METHOD FOR PREDICTING THE PERCENTAGE OF APPLICANTS THAT
WILL FAIL TO BE GRANTED REFUGEE STATUS OR OTHER LEAVE, AFTER THE
CASES HAVE GONE THROUGH THE INITIAL DECISION AND APPEAL PROCESS
(FIRST TWO TIERS)
1. The prediction method is illustrated below, for applications
in June 2006, from nationals of Eritrea.
2. Total initial decisions in October 2005-March 2006,
including dependants and withdrawals prior to initial decision
in January-June 2005.
| Initial decisions1 | % of total2
|
| Granted asylum | 495 | 37
|
| Granted HP/DL or ELR | 65 |
5 |
| Refused | 695 | 55
|
| Withdrawn | 5 | 3
|
| Total | 1,260
| 100 |
1 Provisional figures. Source: CID database as at 17 July 2006.
2 Percentages have been re-weighted, see paragraph 11 on page 4.
| | |
3. Based on these data, 58% of applicants are
not granted status up to and including initial decision, and become
potentially removable.
4. However it is necessary to take account of those who
appeal and have appeals allowed, as follows.
— Of applications refused in July-December 2005,
73% appealed, by 17 July 2006.
— Out of 350 first tier appeal determinations in
July-December 2005, 61% had appeals allowed, at either the first
or second tier.
| Appeal Stage | Appeal determinations in
Jul-Dec 2005 (350)
|
(a) First tier—allowed or withdrawn by Home Office
| 155 |
(b) Second tier—allowed for appellant
| 65 |
(c) Second tier—allowed for Home Office
| 10 |
Allowed = a + b - c
Allowed as a percentage
(210/350)
| 210
61% |
| |
— Combining these figures gives:
55% x 73% x 61% = 24% of all applications that are
refused and are then successful at appeal (or conceded by Home
Office).
— The overall percentage of applicants predicted
to be unfounded is then 58% - 24% = 33%.
NOTES
All data above produced by Immigration Research and Statistics
Service, part of Research, Development and Statistics in the Home
Office.
Data are provisional and rounded to nearest 5. Data may not
sum to totals shown because of rounding.
Annex B
THE STAGES AT WHICH AN ASYLUM APPLICATION CAN BECOME UNFOUNDED[22]
Annex C
SUMMARY OF PERIODS USED IN CALCULATIONS
| Reporting date | Period of data for
initial decisions
| Period of data for % that appeal
and appeal determinations
1
|
Q2 2006
Q1 2006 | Oct 2005-Mar 2006
Oct 2005-Mar 2006
2
| Jul-Dec 2005
Jul-Dec 2005
2 |
| Q4 2005 | Jul-Dec 2005 |
Jul-Dec 2005
2 |
| Q3 2005 | Apr-Sep 2005 |
Apr-Sep 2005 |
| Q2 2005 | Jan-Jun 2005 |
Jan-Jun 2005 |
| Q1 2005 | Oct 2004-Mar 2005
| Oct 2004-Mar 2005 |
| Q4 2004 | Jul-Dec 2004 |
Jul-Dec 2004 |
| Q3 2004 | Apr-Sep 2004 |
Apr-Sep 2004 |
| Q2 2004 | Jan-Jun 2004 |
Jan-Jun 2004 |
| Q1 2004 | Oct 2003-Mar 2004
| Oct 2003-Mar 2004 |
| | |
1 Based on data for a cohort of refusals to estimate
the proportion of refusals for which an appeal is received, and
a cohort of first tier appeal determinations to estimate the proportions
of first tier appeals that had the appeal allowed at either the
first or second tier.
2 Q1 2006 initial decisions and Q4 2005 and Q1 2006 appeal
figures were updated using data for Oct 2005-Mar 2006 and Jul-Dec
2005 respectively as enough time has lapsed to allow late reported
cases to be submitted.
Annex B
(Not Printed)
8
Includes persons departing voluntarily after enforcement action
had been initiated against them, persons leaving under Assisted
Voluntary Return Programmes run by the International Organisation
for Migration, and those who it is established have left the UK
without informing the immigration authorities. Back
9
Q1 2006 has been revised using updated data which results in
the percentage predicted to be unfounded reducing from 69% to
63%. Hence the number of anticipated unfounded applications is
recalculated to be 4,730 (from 5,165 previously) which are less
than the number of failed asylum seekers removed (4,930). See
paragraph 20 of the technical note for further details. Back
10
See Home Office Statistical Bulletin 14/06, Asylum Statistics,
United Kingdom, 2005, for an example of cohort-specific analyses. Back
11
Annex B illustrates the stages at which asylum application can
become unfounded. Back
12
Not all withdrawals necessarily lead to individuals becoming
potentially removable. However, since some withdrawals are known
to result in removals and are included in the removals totals,
all such cases are included in the prediction of numbers becoming
unfounded, as well as (where appropriate) in removals totals. Back
13
Non-suspensive appeal cases: ie removal is not suspended pending
appeal. Back
14
A Third Country case is an application for asylum from a person
who can be returned to a safe third country. A safe third country
is not the applicant's country of origin. It is one where we are
satisfied a person will be neither persecuted nor subjected to
torture or inhuman or degrading treatment or punishment and one
which would not remove a person in breach of the principles of
the UN Refugee Convention or the European Convention on Human
Rights (ECHR). Back
15
Where the appeal determination is recorded as "withdrawn
by Home Office" it is assumed that the case has been conceded
by the Home Office. Back
16
Predicted proportions for unfounded applications are used because
of the time taken from initial application to final decision.
A majority of the final decisions are made within nine months
but a proportion can take longer, and hence an estimate is required
for recent months' intake. Back
17
For cross-sectional data analysis, ratios are based on the stage
reached (decisions, for example) in a fixed period. This differs
from cohort analysis whereby ratios are based on the path of a
fixed group of applicants from their initial application to the
end of the asylum process. Back
18
Outside these top 20 intake generating nationalities, numbers
of applications per nationality are insufficient to make separate
calculations. Back
19
The Family ILR (Indefinite Leave to Remain) Exercise, announced
by the Home Secretary on 24 October 2003, allows certain asylum-seeking
families who have been in the UK for four or more years to stay.
To qualify, the main applicant of the family unit must have applied
for asylum before 2 October 2000 and must have had at least one
dependant aged under 18 (other than a spouse) in the UK on 2 October
2000 or 24 October 2003. Back
20
In due course the approach to calculating the "Withdrawal"
data will be updated to the most appropriate methodology-see paragraph
2 in the Introduction. Back
21
"Asylum and migration: a review of Home Office statistics"
http://www.nao.org.uk/pn/03-04/0304625.htm Back
22
This diagram relates to the calculation method used in this note.
Hence, Judicial Reviews and appeals to the European Court of Human
Rights are not shown. Back
|
