Local Government Resource Allocation: The Approach In Wales

1
Local Government Resource AllocationThe
Approach In Wales

• Anthony Whiffen and Gareth Thomas
• Local Government Finance, Welsh Assembly
  Government

2
Presentation Outline

• Rationale for resource allocation formulae
• Development of Local Government Formula
• the Welsh context
• using regression
• using hierarchical modelling
• future issues

3
Main Local Government Functions

• Education
• Social Services
• Transport
• Planning

4
Three possible resource allocation mechanisms

• according to the size of bids from local areas
• according to historical precedent
• on the basis of some independent measure of need
  (formula)

5
Why use formulae? (1)

• reflect factors driving need to spend on
  particular services
• authorities unable to influence SSA
• authorities funded on equal footing
• minimises the need for judgement

6
Why use formulae? (2)

• transparent
• fair
• updated annually
• relatively economical to administer

7
Formulae development - the Welsh context

• 1996 Local Government in Wales reorganised
• 1998-2000 Independent Review of Formula
• 1999 (Welsh Assembly established)
• 1999 (Partnership Council established)
• 2001-02 First Settlement with New Formula
• 2001-06 Revisions continue following Review

8
Formulae development - the Welsh context
9
Resource Allocation using a formula

• Need to Spend - 3 main drivers
• Main client group (e.g. population, pupil
  numbers)
• Deprivation (e.g. Income Support claimants,
  people with long-term illness)
• Population dispersion

10
Aim discover a common pattern in the spending
responses of local government to the
characteristics of the population

• to what extent did higher spending go with larger
  populations?
• to what extent did higher spending go with
  particular types of deprivation?
• to what extent does higher spending go with
  population dispersion?

11
Formula development process

• Establish relevant client group / population

12
Establishing a client group
13
Formula development process

• Establish relevant client group / population

• Test other variables using SPSS
• Scatter graphs
• Simple Regression
• Stepwise procedure
• or use Hierarchical Regression

14
Regression Results
15
Issues/Problems with formulae

• Intuitive
• Collinearity
• Extreme values
• No model (local policies and efficiency)

16
Hierarchical modelling

• Independent review recommended using small area
  data and hierrarchical regression modelling
• Further study commissioned from York University
  assessing feasibility of data collection and
  application modelling technique

17
Advantages of using Hierarchical Modelling
(Hepple Rees (1998))

• identifies casual relationships between variables
• improves response between variables
• filters higher level effects
• truer reflection of relationship between need
  indicators and spend
• the ecological fallacy

18
The Ecological Fallacy
19
MLM

• Example of services used in
• Personal Social Services (PSS)
• Children and young persons
• Younger adults
• Older adults
• Education
• Pre and primary teaching
• Secondary teaching
• Form of relationship Yij ?o ?1xij uj eij

20
Example Older Adults PSS

• Ward based costs and client numbers basic unit of
  analysis. Data Unit PM forms)
• Unit costs on type of care domiciliary,
  residential and nursing
• Client data DoB, gender, home postcode,care
  costs per week, hours per week

21
Postcoding

• Over 30,000 clients in 868 wards
• Based on 2004 administrative boundaries
• 8.4 with no or incomplete postcode
• 3 authorities omitted due to poor recording of
  postcodes

22
Building The Model

• Over 30 variables tested (deprivation, population
  and sparsity)

23
Final Model

• Variables adding explanatory power of model
  included.
• 33.06 Persons aged 65 and over with a limiting
  long-term illness
• 31.15 Persons aged 65 and over living alone
• 20.75 Persons aged 85 and over
• 15.04 Income support recipients aged 65 and
  over

24
Transparency

• All formula changes agreed by DSG
• Commonly available datasets used
• Greater transparency in recent years
• Green book published since 2005-06 including
  internet
• Increased and wider scrutiny

25
Future issues

• Review of Deprivation and Sparsity
• Review of Education Funding
• Three Year Settlements

26
Further Reading

• Capitation funding in the public sector,
• Journal of the Royal Statistical Society Series
  A (Statistics in Society), 2001, vol. 164, no. 2,
  pp. 217-257(41)
• Hepple Rees (1998) Standard Spending
  Assessments and Multilevel Modelling. A Report
  for the Department of the Environment, Transport
  the Regions (Bristol Geoeconomics Research)