Managing the Bathtub of Tax Debt UK Chapter of the System Dynamics - - PowerPoint PPT Presentation

Managing the Bathtub of Tax Debt

UK Chapter of the System Dynamics Society Annual Gathering February 2011 Nick Whitehouse & Nicola Fletcher

KAI Enforcement & Compliance | | 2

Contents

• 1. Background – what is debt?
• 2. Analysis issues
• 3. Interim modelling & need for enhanced debt model
• 4. Complexity of the problem – “the bathtub of debt”
• 5. System dynamics modelling approach
• Development of process maps
• Partial view of simulation model
• Example of outputs
• 6. Next Steps & Future uses of the model

KAI Enforcement & Compliance | | 3

Background – What is debt?

• What is a tax debt? – It‟s where a taxpayer owes us money by a certain date, and

has not paid it

• What is not a tax debt? – Where a taxpayer should be paying us some / more

money, but we don‟t know about it (avoidance – if it‟s legal, evasion – if it‟s not)

• What different types of tax debt are there? – Main types are:
• Known debt – have told us how much they owe us (usually in a return), but

have not paid it

• Estimated debt – we know that they owe us something, but we have to

guess how much (based on what they historically owed)

• Spurious debt – we think they owe us something, but they do not (e.g. due to
• ur systems allocating payments to the wrong accounts)
• Penalties – extra debts for not paying the previous debts (either at all or on

time)

• How much debt is there? – Headline debt balance is around £20bn, but this can

be misleading – especially due to large volumes and values of churn over time (annual clearances are over £50bn)

KAI Enforcement & Compliance | | 4

Analysis issues (1)

• Following the recession, the level of interest in debt has risen, and so has the

analytical resource devoted to it. This has uncovered a few issues…:

• Historically, what happened to debts of different values?
• High value debt – letters, phone calls, distraint, court action, bankruptcy –

active pursuit

• Low value debt – letters, often not much else – passive pursuit
• In theory, staff would carry out active pursuit on the highest debt values, and keep

moving down the values as far as they can

• In practice, the value reached differs heavily by tax type. The rules used for

entering information on the database differ so much by tax type that staff can not move from one tax to another – e.g.:

• Amendment = £100 Does this mean amend the debt up by £100, down by

£100, or amend the debt value to £100. Could be any of the above!

• Debts across tax types do not get linked up!

KAI Enforcement & Compliance | | 5

Analysis issues (2)

• What‟s the counter-factual to what Debt Management & Banking (DMB) do?
• If we stopped doing anything, what would happen to debts?
• „Self-clearance‟ rate – what percentage of debts pay without us needing to

do anything?

• Impact of debt value on this rate
• Need to factor this into benefit / cost calculations
• How much debt (& benefit) is up for grabs? Take out:
• Self-clearances
• Bankruptcies and unavoidable „write-offs‟
• Spurious debts
• …and all that‟s left are debts where we could get payment on what would
• therwise have become a write-off
• HMRC have been doing more recently to make sure we bring those debts in…

KAI Enforcement & Compliance | | 6

Analysis issues (3)

• What is the value of an FTE working in DMB?
• Historic answer = Total DMB receipts / DMB staff
• Marginal impact – what difference would one person more / fewer make?
• Self-clearance rate – especially for high value debts
• When is a debt not a debt?
• PAYE CHAPS payments – same-day bank transfers
• Payment is made on deadline day, but not posted on our systems until the

next day…

• …but the transfer to the debt system is done overnight after the deadline –

meaning that a spurious debt is created, and then cleared the next day

• This was inflating both annual new debt and debt payment figures by…
• £36bn!
• Also of relevance to first issue…

KAI Enforcement & Compliance | | 7

Interim modelling

• Model required for transformation programme (DTP) benefits estimates -
• DMB have changed the way they pursue debts – moved to a „Campaigns‟

approach (i.e. hit as many debts as possible while they are new), taking advantage of the variation in inflow rates for each tax type across the year

• System change combined with expansion of the call centre – if more debts

are resolved at an early stage (by letters and phone calls), then there is less demand for later, more costly, stages of pursuit (e.g. court action)

• Quick Excel model produced, based on the assumption that the change

will allow us to actively work more lower value debts

• Benefits estimates flowchart (intentionally cautious estimates):

100 additional debts worked 60 debts paid 40 debts cleared through non-payment or not cleared 48 would have paid anyway 12 would not have paid

12 = Additional Receipts 60 = Debt Receipts

KAI Enforcement & Compliance | | 8

Improved debt model – possible uses & results

Replacing the interim model with an enhanced model - to improve understanding of inflows & outflows, and provide reliable & consistent results

• Possible uses for the model include:
• Producing improved and more robust estimates of the benefits / costs of

changes, such as change programmes or staff reductions

• Producing enhanced debt forecasts
• Simulating „what-if‟ scenarios of changes to:
• Debt inflow rates
• Outsourcing of debts to Debt Collection Agencies (DCAs)
• DMB staff deployment, processes and priorities
• Optimising the deployment of DMB staff and the use of DCAs
• Model results will need to be shown in terms of:
• Debt balance
• Payments and / or other clearances
• Roll-rate (debts cleared within 30 or 90 days)

KAI Enforcement & Compliance | | 9

Approach to modelling

What type of model to make?

• Considered spreadsheet modelling, discrete event simulation & system

dynamics

• Used SD approach because:
• Stocks and flows - debts modelled as stocks that flow between different parts of

the system, with flow rates and routes decided by data inputs

• Feedback loops – debt inflows next year are affected by what happened this

year

• Delays – not as simple as a queue building up – e.g. waiting for a debtor to

respond to a letter

KAI Enforcement & Compliance | | 10

Complexity of the problem - The bathtub of debt

• Bath analogy:
• Debt is the water, DMB systems are the bathtub, inflow of new debts from

the tap, outflow through the plughole

• Complications…
• Turning the taps – inflow rate can change due to:
• Economy
• Policy changes (e.g. new penalties) – both ways
• Preventing spurious debts
• Deterrent effect of successful pursuit (feedback loop from plughole)
• Changing the size of the plughole – outflow rate can change due to:
• Economy
• Pursuit by DMB

KAI Enforcement & Compliance | | 11

The bathtub of debt

• More complications…
• Need more than one plughole,

with different flow rates for each:

1. Payments 2. Write-offs & remissions 3. Sent to DCAs (which may come back to the bath later…)

• Need more than one bath – one for each kind of debt
• Need more than one kind of liquid – debts of different values will behave,

and be pursued, very differently

• The liquid changes over time – the longer it stays in the bath, the lower the

chance we‟ll receive payment, therefore more likely to end up going down the write-off plughole… but, in the meantime, it might self-clear

• Some of the liquid is not real – e.g. spurious debts. Does this go down a

plughole, or just evaporate?!

KAI Enforcement & Compliance | | 12

System dynamics modelling of debts

• Model framework using system dynamics
• Link behaviour of three basic stocks and flows – staff resources, debt

items & debt values

• Each type of debt modelled separately
• New debts flow in to model (e.g. outstanding returns, incorrect payments)
• Debts can be transferred between Debt Management functions
• Cleared debts flow out of model (e.g. payments, write offs etc.)
• Capture the factors affecting the behaviour of the system
• Workshops with DMB colleagues to produce process maps for each type
• f debt
• Simulation model of system
• Vensim software – user friendly tool with Excel interface (for both inputs

and outputs)

KAI Enforcement & Compliance | | 13

Example of process maps

For each IDMS 99 issued - working assumption of 33% response rate Of respondees - working assumption

• f 95% phone call, 5% letter

promise to pay remains at BEU TTP remains at BEU Hopper Campaign DMTC for

• utbound calls

EIS IDMS 99 issue DMTC inbound calls Field Force Debt Technical Office BEU TTP monitoring

• ffice

Write Off Office debts in

• therwise

no response phone call in letter in refusal or no contact ttp agreed recommend write off promise to pay or dispute ttp agreed recommend write off promise to pay or dispute refusal ttp agreed recommend write off no suitable response recommend insolvency action forward to BEU forward to BEU recommend write off no suitable response recommend insolvency action court action required (if cannot seize goods) ttp agreed recommend write off promise to pay recommend insolvency action recommend insolvency action recommend write off case with write off, cessation or TTP flag if high value (£100k+) recommend insolvency action recommend insolvency action issue IDMS 99

KAI Enforcement & Compliance | | 14

Partial view of model

• Made use of subscripts to subdivide the stocks and flows for CT debts by value bands

and age bands

DTO CT Items CT Items in to DTO CT Items out from DTO DTO CT Items Cleared DTO CT Items Aging DTO CT Items Outcome Rate DTO CT Items Actioned DTO CT Items Available Average time in cohort <DTO Contact Availability>

KAI Enforcement & Compliance | | 15

Partial view of model – determining debt age

• DTO CT Items Aging[Debt type]=(DTO CT Items[Debt type]-CT Items out from DTO[Debt

type]-DTO CT Items Cleared[Debt type])/Average time in cohort[Debt type]

• DTO CT Items[New]=+In[New]-Out[New]-Cleared[New]-Aging[New]
• DTO CT Items[All but new]=+Aging[Previous cohort]+In[All but new]-Out[All but new]-

Cleared[All but new]-Aging[All but new]

DTO CT Items CT Items in to DTO CT Items out from DTO DTO CT Items Cleared DTO CT Items Aging DTO CT Items Outcome Rate DTO CT Items Actioned DTO CT Items Available Average time in cohort <DTO Contact Availability>

KAI Enforcement & Compliance | | 16

Partial view of model – prioritising debts

• DTO CT Items Actioned[Debt type]=ALLOCATE BY PRIORITY(DTO CT Items

Available[Debt type],Debt priority[Debt type],ELMCOUNT[Debt type],0.5,DTO Contact Availablity[CT])

• DTO Contact Availability[Stock type]=ALLOCATE BY PRIORITY(DTO Total Stock[Stock

type],Stock priority[Stock type],ELMCOUNT[Stock type],0.5,DTO Staff*Average items worked per DTO Staff per day)

DTO CT Items CT Items in to DTO CT Items out from DTO DTO CT Items Cleared DTO CT Items Aging DTO CT Items Outcome Rate DTO CT Items Actioned DTO CT Items Available Average time in cohort <DTO Contact Availability>

KAI Enforcement & Compliance | | 17

Example of output from model

Total Debt Balance £m - Scenario for outsourcing debts to DCAs

£16,000 £17,000 £18,000 £19,000 £20,000 £21,000 £22,000 £23,000 Nov-10 Jan-11 Mar-11 May-11 Jul-11 Sep-11 Nov-11 Jan-12 Mar-12 May-12 Jul-12 Sep-12 Nov-12 Jan-13 Mar-13 May-13 Jul-13 Sep-13 Nov-13 Jan-14 Mar-14 May-14 Jul-14 Sep-14 Nov-14 Jan-15 Mar-15 Base Model Outsourcing to DCAs

KAI Enforcement & Compliance | | 18

Next Steps & Future uses of model

• Core structure of model is in place but we need to test model with real

(historical) data & make revisions to model & assumptions

• Make use of model to answer key questions e.g.
• What would be the impact of outsourcing particular debts to DCAs?
• How can we optimise the use of our resources (staff & their actions)?
• What would be the impact of delaying/advancing activity on particular

types of debts?

• Are roll rate targets achievable? How do we need to deploy our staff to

reach the roll rate targets?

• What is the value of an FTE? IAO recommendation – “change focus to

see what each operation brings in per member of staff”

KAI Enforcement & Compliance | | 19

Any Questions?