Provision of Public Goods in a Ratepayer Democracy

1
Provision of Public Goods in a Ratepayer
Democracy

• Toke S Aidt
• University of Cambridge
• Jayasri Dutta
• University of Birmingham
• Martin Daunton
• University of Cambridge

Supported by ESRC grant L138251006
2
The questions?

• Does local democracy help or hinder the solution
  of collective action problems?
• How does the allocation of voting rights in local
  elections matter for provision of local public
  goods?
• Can the extension of the voting franchise be a
  source of retrenchment?

3
The context

• Public health crisis in urban Mid-Victorian
  England and Wales.
• Local government (the Corporations)
• Spending and taxation were almost fully
  decentralized.
• Fiscal decisions were made through ratepayer
  democracy.
• Two facts
• Uneven spending on public health.
• Uneven extension of the municipal franchise.
• The retrenchment hypothesis (Hennock, 1973,
  Szreter, 1997)
• Uneven spending explained by uneven extension of
  voting franchise.

4
The Retrenchment hypothesis.
g/cap
franchise
5
Overview

• Institutional background
• Sketching a model of ratepayer democracy
• An empirical test of the retrenchment hypothesis

6
Municipal Corporations (1835-1888)

• Responsibility Main providing of public health
  and other local public goods (1865-1888).
• Self-governance Corporations could decide how
  much and when to spent.
• Self-financing Local property taxes (the rates)
  were the main source of income.

7
(No Transcript)
8
Ratepayer democracy

• Who could vote?
• male inhabitant householders (after 1869 also
  females).
• 2.5 years of residence (reduced to 1 in 1869).
• Enrolled on the Burgess Roll (registration).
• No poor relief.
• Payment of the rates.
• Municipal franchise based on the principle of no
  taxation without representation.
• ratepayers were enfranchised.
• disenfranchised residents could not be taxed.

9
Extension of the franchise registered
voters/adult male population
10
.24
Fraction
0
0
100
Extension of the franchise in 1852
11
.19
Fraction
0
0
100
Extension of the franchise in 1865
12
Municipal Franchise Act of 1869 Assessed Rates
Act of 1869
.16
Fraction
0
0
100
Extension of the franchise in 1871
13
Municipal Corporation Act of 1884
.2
Fraction
0
0
100
Extension of the franchise in 1884
14
The uneven extension of the franchise

• Temporal pattern explained by national
  legislation.
• Spatial pattern explained by local differences in
• The boundaries of rated property.
• The rights of indirect taxpayers (compounding).
• The rights of occupiers of shared accommodation.
• Administrative inefficiency and the cost of
  registration.

15
The Retrenchment hypothesis.
g/cap
franchise
working
upper
middle
Social class
16
Empirical evidenceData material

• Sources BPP, in particular the Local Tax Returns
  and the Censuses
• issues of reliability.
• issues of missing data.
• Three cross sections of spending and franchise
  data (1868, 1871 and 1886).
• Unbalanced panel with 141 boroughs.
• Sanitation
• Expenditure side mixture of capital and current
  expenditures.
• Revenue side loans used for capital
  expenditure, but often related to other things
  than sanitation.

17
Empirical Specification
-

Retrenchment hypo
Strategy I Conditional independence.

• Class composition (industrial employment)
• Wealth distribution (rateable value)
• Population density
• Population size
• Fixed regional effects or fixed borough effects
• Fixed time effects.
• Borough-specific trends

18
Strategy II Instrument variables.

• Cross section variation endogenous (voluntary
  enfranchisement).
• Time series variation is exogenous (forced
  enfranchisement).
• Use past values of the franchise as instruments
  (1852, 1865, 1871).
• Reform dummy variables (reform in 1869 and 1882).

19
Table 3 Results for spending on sanitation per
(1000) capita.
20
Table 4 Results for borrowed funds per (1000)
capita.
21
(No Transcript)
22
Summary and robustness

• U-shaped relationship with a turning point around
  40 per cent.
• Robust to
• addition of other control variables.
• longer fiscal lags
• clustered standard errors and autocorrelation
• general-to-specific
• random effects estimation.
• The instrument variable model
• Passes Hansens J test for validity of
  instruments and the instruments are highly
  relevant in the first stage

23
Conclusions

• We provide evidence consistent with the notion
  that ratepayer retrenchment was an important
  factor behind the uneven response to the public
  health crisis in mid-Victorian British cities.
• We analyse a unique example of the principle of
  no-taxation-without-representation and explore
  some of the theoretical implications that flow
  from it.

24
Links to the literature

• Empirical research on consequences of franchise
  extension (Lott Kenny, JPE 1999 Husted and
  Kenny, JPE 1997 Aidt et al. EER 2005).
• Economic theories of the franchise extension
  (Acemoglu Robinson, QJE 2001 Persico
  Lizzeri, QJE 2004 Justman Gradstein, EEH
  1999).
• Economic development today (World Bank, 2000,
  Bardhan AER 2002).

25
(No Transcript)
26
2
Tu
rning point

35.6

39.9

41.5

40.9

43.5

Method of estimation

PCSE

Fixed

Median

IV
-
GMM

IV
-
fixed

Regional fixed effects

Yes

No

Yes

Yes

No


27
(No Transcript)
28
Links to the literature

• Economic history the standard of living debate
  (Hennock, 1973, Szreter, 1997, Wolh, 1983, Offer,
  1981, Fraser, 1976 etc.)
• Empirical research on consequences of franchise
  extension (Lott Kenny, JPE 1999 Husted and
  Kenny, JPE 1997 Aidt et al. EER 2006).
• Economic theories of the franchise extension
  (Acemoglu Robinson, QJE 2001 Persico
  Lizzeri, QJE 2004 Justman Gradstein, EEH
  1999).
• Economic development today (World Bank, 2000,
  Bardhan AER 2002).

29
A model of Ratepayer democracy

• Local public good (g) gt higher factor
  productivity gt higher profits and wages.
• g financed by local property taxes (the rates).
• Wealthy capitalists (H)
• Higher profits (large effect).
• Regressive rates.
• Middle class capitalists (M)
• Higher profits (modest effect).
• Regressive rates.
• Working class men (L)
• Higher wages.
• Live in rented accommodation, so with fixed
  supply, the incident of the rate falls on the
  landlords (the capitalists).

30
Specification of the city economy
Incomes
Policy preferences
Fiscal policy
Group sizes
31
We ask three questions

• Under which conditions do the retrenchment
  hypothesis hold?
• When is the franchise extended endogenously?
• What are the implications of a forced franchise
  extended?

32
Three Franchise Regimes

• H-franchise only wealthy capitalists can vote,
  stand for election and have their property taxed.
• HM-franchise all capitalists can vote, stand for
  election and have their property taxed.
• HML-franchise all citizens can vote, stand for
  election, and all property, including rented
  accommodation, can be taxed.

33
PoliticsCitizen Candidate Model(Besley Coate,
1997 Osborne Slivinsky 1996)

• Any enfranchised citizen can stand for election
  at a cost
• Election among the candidates using plurality
  rule.
• Citizens vote sincerely.
• Uncontested candidate wins for sure.
• Ties are broken by an unbiased lottery.
• Winning candidate decides fiscal policy subject
  to council budget. Default policy g0 if nobody
  stands for election.

34
Equilibrium outcomes

• H-franchise wealthy capitalist runs and wins.
• HM-franchise middle class capitalist runs and
  wins.
• HML- franchise wealthy capitalist runs and wins
  with support of workers.
• Two effects determine impact on spending
• Tax base effect ( of enfranchised citizens).
• Preference effect (wealth distribution).

35
PROPOSITION 1 (The retrenchment
hypothesis) Spending on local public goods is
higher under H-franchise than under
HM-franchise if and only if Spending is
higher under HML-franchise than under H- or
HM-franchise.
Intuition The preference effect must
dominate the tax base effect.
Implication Middle class retrenchment more
likely in boroughs with large wealth inequality
and a small middle class.
36
H to HM increase in g
HM to HML increase in g.
H to HM decrease in g
HM to HML increase in g.
1
1
37
Extension of the franchise

• Are the elite and/or the middle class willing to
  give away the franchise?
• Are those who receive the franchise (the middle
  and/or working class) willing to accept it?
• Voluntary enfranchisement Pareto improvement
  and is likely to happen endogenously.
• Forced enfranchisement Pareto non-improvement
  and is likely to be caused by exogenous factors,
  e.g., national legislation.

38
Observation Voluntary extension from HM to HML
is impossible because the middle class will
object - HM-franchise gM but tax payment. -
HML-franchise gH and effectively no enlargement
of tax base.
PROPOSITION 2 (Voluntary enfranchisement) Enfranc
hisement of the middle class is voluntary if
and only if the middle class is willing to
accept it and is always associated with an
INCREASE in spending on local public goods.

• Middle class only willing to accept the vote if
  spending goes up
• H-franchise gH and no tax payment
• HM-franchise gM but tax payment.

39
Middle
V
1
1
40
PROPOSITION 3 (Forced enfranchisement) Enfranchis
ement of the middle class is forced if and only
if the elite is unwilling to give it away and is
always associated with a DECREASE in spending
on local public goods.

• Intuition
• H-franchise gH but have to pay themselves.
• HM-franchise gM but can share cost with middle
  class,
• so if g increase, the elite will certainly want
  to give the vote to
• the middle class.

41
Middle
V
Elite
PF
F
1
1
42
Implications

• Retrenchment CANNOT be associated with voluntary
  enfranchisement. It happens when the franchise is
  forced upon the middle class.
• Spatial variation is mostly a result of voluntary
  enfranchisement.
• Temporal variation is mostly a result of forced
  enfranchisement.

43
(No Transcript)
44
(No Transcript)
45
Specification of the city economy
Policy preferences
Incomes
Fiscal policy
Group sizes
46
Why is this of interest?

• Economic history perspective
• 90 per cent of all investments made locally.
• Statistical test of the retrenchment hypothesis.
• Empirical research on consequences of franchise
  extension
• Complement to US and European-wide studies. (Lott
  Kenny, 1999 Husted and Kenny, 1997 Aidt et al
  2005).
• Large inter-borough variation in extension of
  franchise.
• Large sample of fiscally and politically
  independent units.

47
Why is this of interest?

• Economic theories of franchise extension
• The puzzle why did the elites give away voting
  rights? (Acemoglu Robinson, 2001 Persico
  Lizzeri, 2004 Justman Gradstein 1999).
• No-taxation-without-representation gives rise
  to a different logic of enfranchisement.
• We provide a simple model of franchise extension
  in a ratepayer democracy.
• Itenary

48
Why is this of interest?

• Economic development
• Many LDCs face similar public health problems in
  urban areas as mid-Victorian cities in Britain.
• Decentralization of public services in LDCs as a
  strategy for improving provision and reaching the
  poor (World Bank, 2000, Bardhan 2002).
• We stress the positive role of local elites and
  the danger of middle class retrenchment.

49
Specification of the city economy
Demography Wealthy capitalists (H) nH
Middle class capitalists (M) nMgtnH Workers
(L) nLgt nHnM
Endowments Capital k, labour and housing
h. Wealthy capitalists kH two units of
h. Middle class capitalists kM two units of
h. Workers one unit of labour.
Technology
50
Specification of the city economy
Preferences
Perfect competition in labour market and in
market for rented accommodation no market for
capital.
Equilibrium incomes
Taxation The tax base is rated property.
House owners pay the rate, i.e., tenants
(workers) pay indirectly through the
capitalist owner (compounding). Rate is
regressive levied at a uniform rate.
51
Specification of the city economy
Housing Capitalists consume one unit privately
and supply the other unit to a competitive
market for rental accommodation where workers
rent. Incident of any rate on rented
accommodation falls 100 on the capitalists.
Policy preference functions
nf is of enfranchised citizens.
52
PoliticsCitizen Candidate Model(Besley Coate,
1997 Osborne Slivinsky 1996)

• Any enfranchised citizen can stand for election
  at a cost
• Election among the candidates using plurality
  rule.
• Citizens vote sincerely.
• Uncontested candidate wins for sure.
• Ties are broken by an unbiased lottery.
• Winning candidate decides fiscal policy subject
  to council budget. Default policy g0 if nobody
  stands for election.

53
PROPOSITION 1 (The retrenchment
hypothesis) Spending on local public goods is
higher under H-franchise than under
HM-franchise if and only if Spending is
higher under HML-franchise than under H- or
HM-franchise.
Intuition The preference effect must
dominate the tax base effect.
Implication Middle class retrenchment more
likely in boroughs with large wealth inequality
and a small middle class.
54
Summary

• Voluntary enfranchisement of the middle (and
  working) class likely in cities with low wealth
  inequality and a large middle class.
• Forced enfranchisement of the middle class likely
  in cities with high wealth inequality and a small
  middle class.
• Partially forced enfranchisement may or may not
  be associated with an increase in spending on
  sanitation.

55
Definitions

• Voluntary enfranchisement those who give the
  vote away and those who get it are both better
  off.
• Forced enfranchisement those who give the vote
  away and those who get it are both worse off.
• gt

56
PROPOSITION 2 (Voluntary enfranchisement) Enfranc
hisement of the middle class is voluntary if
and only if the middle class is willing to
accept it and is always associated with an
INCREASE in spending on local public goods.
PROPOSITION 3 (Forced enfranchisement) Enfranchis
ement of the middle class is forced if and only
if the elite is unwilling to give it away and is
always associated with a DECREASE in spending
on local public goods.
57

• Who could stand for election?
• enfranchised residents, but with minimum wealth
  requirement (till 1882).
• Councilors were unpaid, with meetings during
  working hours.
• Enfranchised working class men were effectively
  prevented from running.

58
H to HM increase in g
HM to HML increase in g.
H to HM decrease in g
HM to HML increase in g.
1
1
59
Extension of the franchise

• Are those giving away the franchise (the elite
  and/or the middle class) willing to do so?
• Are those who receive the franchise (the middle
  and/or working class) willing to accept it?
• Voluntary enfranchisement represents a Pareto
  improvement and is likely to happen endogenously.
• Forced enfranchisement is Pareto deteriorating
  and must to caused by exogenous factors, e.g.,
  national legislation.

60
Middle
V
Elite
PF
F
1
1
61

• Retrenchment
• Tax base effect small (small middle class)
• Preference effect large (large wealth
  differences)
• Voluntary or forced enfranchisement
• Enfranchised agents may extend the franchise in
  exchange for a larger tax base.
• Disenfranchised agents may not want to be
  enfranchised because they then have to pay taxes.
• Voluntary enfranchisement represents a Pareto
  improvement and is likely to happen endogenously.
• Forced enfranchisement is Pareto deteriorating
  and must to caused by exogenous factors, e.g.,
  national legislation.