Transcript The Adoption of Constitutional Home Rule

The Adoption of
Constitutional Home Rule:
A test of endogenous policy decentralization
Jessica Hennessey
PhD Candidate
University of Maryland
Research Question
• Why did some states grant constitutional home
rule provisions to municipalities in the late 19th
and early 20th century?
Home rule: Municipality can independently
enact, adopt and amend charter
• Hypothesis:
States with more heterogeneous local
preferences are more likely to decentralize
control by granting home rule
Figure 1
Adoption of Constitutional Home Rule: 1875-1932
hrs
Home Rule States
Motivation
• Provide empirical evidence of policy
decentralization
• Understand origins of institution so we can better
evaluate its effects
• Role of heterogeneity within communities
• Historical importance in transition from state to
local government
Approach
• Step 1: Observe differences in municipalities
– In home rule states versus in non-home rule states
– Within home rule states: adopt home rule charter or
remain under general legislation
• Step 2: Estimate municipal-level latent taste for
home rule (use data from home rule states)
• Step 3: Compare predicted municipal
preferences in home rule states to predicted
municipal preferences in non-home rule states
Municipal Decision:
Home Rule or General Legislation
• Organization under General Legislation
– Set by state legislature, uniform for all municipalities
– Common features:
•
•
•
•
Structure (example: one mayor, 5 trustees)
Elections (frequency, number of wards)
Powers
Debt and tax limitations (shared by Home Rule municipalities)
• Home Rule
– Ability to draft and adopt own charter
– Formal recognition of local self-determination
– Features of charters
•
•
•
•
Governance structure
Facilitate ownership and operation of public utilities
Initiative, referendum, recall
Impose more stringent tax and debt limits
Who Adopts a Home Rule Charter
• Three primary factors: size, growth, and
infrastructure investment
• Degree of homogeneity/heterogeneity
– Alesina and La Ferrara (2000)
– Buchanan and Tullock (1962)
• Different from average municipality in state
(general legislation not appropriate)
Data Available
• Year of charter adoption
– 11 Home Rule States
– For first home rule charter
– Information up until 1935
• Municipal level
–
–
–
–
Wealth, Debt and Taxation from Census
Population, Gross Debt less Sinking Fund, Sinking Fund Assets
1890, 1902: all municipalities
1913, 1922, 1932: municipalities >2500
• County level
– ICPSR 2896: Native born population, urban population
– ICPSR 8611: Congressional election results
• % Democrat, % Other 3
2
• Political HHI: HHI   (partyshare)
i 1
Adoption of Constitutional Home Rule
Era of Home Rule
1875-1889
1933-1958
1896-1932
1959-1974
Home Rule versus Non Home Rule States
1890 Data for use in Probit Estimation
Home Rule
States
Non-Home Rule
States
Difference in
Means/Variance
Political HHI
0.49
(0.115)
0.50
(0.105)
**
(***)
% Democrat Votes
45.42
(20.21)
43.71
(21.93)
*
(***)
% Other Votes
9.93
(14.51)
11.00
(19.71)
(***)
5.09
(20.49)
3.98
(30.17)
(***)
0.26
(0.47)
0.29
(1.22)
(***)
7.04
(15.55)
6.71
(14.77)
(*)
Sinking Fund (per capita)
0.26
(1.96)
0.12
(0.48)
**
(***)
% Native born
82.27
(12.79)
84.41
(11.44)
***
(***)
925
1264
Variable
Population (in thousands)
Population Growth
Gross Debt Less Sinking Fund (per capita)
Number of Municipalities
Notes:
Standard deviations are in parentheses
***, **, * denotes significance at the 1 percent, 5 percent, and 10 percent level, respectively.
In HR States: Home Rule versus General Legislation
1890 Data for use in Probit Estimation
Home Rule
Municipality
General Legislation
Municipality
Difference in
Means/Variance
Political HHI
0.52
(0.15)
0.48
(0.1)
***
(***)
% Democrat Votes
51.62
(22.03)
43.32
(19.12)
***
(***)
% Other Votes
9.35
(13.99)
10.12
(14.68)
Population (in thousands)
14.24
(39.14)
1.99
(2.63)
***
(***)
Population Growth
0.39
(0.44)
0.21
(0.48)
***
Gross Debt Less Sinking Fund (per capita)
8.34
(12.28)
6.59
(16.5)
*
(***)
Sinking Fund (per capita)
0.64
(3.73)
0.13
(0.64)
***
(***)
% Native born
78.19
(14.12)
83.65
(12)
***
(***)
234
691
Variable
Number of Municipalities
Notes:
Standard deviations are in parentheses
***, **, * denotes significance at the 1 percent, 5 percent, and 10 percent level, respectively.
Econometric Specification
• Cross-section (1890) estimation
t  X ics  Zcs  (Vics,cs Vs )   Ss  uics
*
ics
2
• Panel data estimation
*
ticst
 X icst   Zcst   (Vicst ,cst Vst )2  Ss  Yt  uicst
Covariates
– Xics:
– Zcs:
– Vics,cs:
– Ss :
– Yt :
population, population growth, gross debt less
sinking fund per capita, sinking fund assets per capita
% native-born, political HHI, % Democrat, % Other
population growth, % native, gross debt less sinking fund
per capita, political party
state fixed effects
time effects
• Assumption: decisions are independent
Results on 1890 Data
Population (in thousands)
Population (in thousands) [squared]
Probit Model (Marginal Effects)
(1)
(2)
(3)
0.044
0.044
0.046
(0.007)***
(0.008)***
(0.007)***
Linear Probability Model
(4)
(5)
(6)
0.018
0.019
0.019
(0.002)***
(0.002)***
(0.002)***
-0.0001
(0.00002)***
-0.0001
(0.00002)***
-0.0001
(0.00002)***
-0.00006
(0.000008)***
-0.00006
(0.000008)***
-0.00006
(0.000008)***
Population growth
0.271
(0.048)***
0.170
(0.043)***
0.247
(0.048)***
0.212
(0.039)***
0.156
(0.039)***
0.199
(0.039)***
Different Political Party from State
-0.114
(0.036)***
-0.035
(0.032)
-0.052
(0.041)
-0.105
(0.029)***
-0.056
(0.029)*
-0.053
(0.030)*
0.019
(0.177)
-0.228
(0.185)
-0.346
(0.201)*
0.175
(0.147)
-0.004
(0.199)
-0.223
(0.195)
% Democrat Votes
0.007
(0.001)***
0.004
(0.002)**
0.004
(0.001)***
0.005
(0.001)***
0.002
(0.001)
0.003
(0.001)***
% Other Votes
0.006
(0.001)***
0.005
(0.002)***
0.004
(0.002)***
0.004
(0.001)***
0.002
(0.001)**
0.002
(0.001)**
% Native born
-0.886
(0.153)***
0.003
(0.166)
-0.953
(0.163)***
-0.644
(0.121)***
0.011
(0.181)
-0.689
(0.127)***
Gross Debt Less Sinking Fund (per capita)
0.002
(0.002)
0.004
(0.002)**
0.001
(0.002)
0.003
(0.001)**
0.003
(0.001)**
0.001
(0.001)
Sinking Fund (per capita)
0.016
(0.017)
0.004
(0.014)
0.006
(0.017)
0.006
(0.003)*
0.001
(0.003)
0.004
(0.003)
-0.140
(0.035)***
-0.070
(0.028)**
-0.126
(0.034)***
-0.086
(0.017)***
-0.058
(0.016)***
-0.086
(0.017)***
-0.689
(1.155)
0.194
(1.036)
-0.831
(1.223)
0.354
(0.899)
1.001
(1.038)
0.146
(0.918)
-0.00002
(0.00006)
-0.00006
(0.00004)
-0.00003
(0.00005)
-0.00001
(0.000004)***
-0.00001
(0.000004)***
-0.000006
(0.000004)
Political HHI
Population growth compared to state average [squared]
% Native compared to state average [squared]
Gross Debt Less Sinking Fund (per capita) compared to state average [squared]
State Fixed Effects
Y
Region Effects
Y
Y
Y
Time Effects
Observations
925
925
925
925
925
925
Results on Panel Data
Probit
(8)
0.002
(0.000)***
Linear Probability
(11)
0.003
(0.001)***
-0.000002
(0.0000008)**
-0.000004
(0.000002)**
Population growth
0.033
(0.011)***
-0.006
(0.011)
Gross Debt Less Sinking Fund (per capita)
0.001
(0.000)**
0.001
(0.001)*
Y
Y
1216
1216
Population (in thousands)
Population (in thousands) [squared]
State Fixed Effects
Observations
Duration Model
• Likelihood Function
n
n
i 1
i 1
ln L   d i ln f (t i , )   (1  d i ) ln S (t i , )
Where
f (t , )   (t , )S (t , )
 t , x,  , 0    ( x,  )0 (t )
• Specification
Weibull: t, X i   exp(X i  )t  1
• Same X as used in probit model, include state
effects
Duration Results
Hazard Ratio
1890
Change in estimated
transition rate w/ 1σ
increase
Panel
Change in estimated
Hazard Ratio
transition rate w/ 1σ
increase
Population (in thousands)
1.027
(0.015)*
89%
1.011
(0.007)
47%
Population (in thousands) [squared]
1.000
(0.000)
-39%
1.000
(0.000)
-23%
Population growth
2.620
(0.625)***
56%
1.524
(0.127)***
80%
Gross Debt Less Sinking Fund (per capita)
1.038
(0.011)***
72%
1.006
(0.003)**
19%
State Fixed Effects
Y
Y
1102
987
Time Effects
Observations
Notes:
Robust standard errors in parentheses
***, **, * denotes significance at the 1 percent, 5 percent, and 10 percent level, respectively.
All specifications assume the Weibull distirbution.
Sample is all municipalities in home rule states.
Home rule states in 1890 analysis are California, Colorado, Michigan, Minnesota, Nebraska, Ohio and Texas.
Home rule states in panel data analysis are Arizona, California, Colorado, Michigan, Minnesota, Missouri, Nebraska, Ohio,
Comparing Municipalities
• Use estimates to generate predictions for
municipalities in Home Rule states and NonHome Rule states
– Probit: probability of adopting HR charter
– Duration: probability of survival (not adopt HR charter
by 1935)
• How to quantify heterogeneity of preferences
– Strength of preference
– Variance
Summary of Results
Table 5: Comparison of Home Rule and Non-Home Rule Municipalities
Strength of Preference
Home Rule States
Non-Home Rule
States
Variance of Preference
Home Rule States
Non-Home Rule
States
Probit
1890
Panel
0.12
0.04
0.06
0.01
*
*
0.08
0.03
0.04
0.01
*
*
Linear Probability
1890
Panel
0.18
0.18
0.12
0.09
*
*
0.06
0.02
0.02
0.01
*
*
Duration
1890
Panel
0.12
0.06
0.09
0.04
*
*
0.03
0.03
0.01
0.01
*
*
Note:
* Denotes significant difference at 1% level
Panel data predictions are constructed as the average for each municipality.
Conclusion
• States that adopted home rule have greater
heterogeneity in estimated municipal preference for
home rule than states that did not adopt home rule
• Contribution
– Empirical evidence of policy decentralization
– An explanation for institution of home rule
– Unique municipal-level dataset
• Future work
– Laboratory federalism: empirical exploration with a
spatial/geographic component
– Political economy: effects of HR
– How does HR relate to other facets of the state-local
relationship, such as the evolution of tax structures