Marina Arbetman Rabinowitz

1
Relative Political Capacity New Models Data
Updates

• Marina Arbetman Rabinowitz
• Kristin Johnson

2
Indicators of Political Development

• Indicators of Well being
• Electricity, calories per capita, education,
  health, etc. Deutsch (1966), Rokkan (1970), Gurr
  (1974), World Handbook of Social Indicators (WB),
  Human Development Index (UN), Millennium
  Development Goals.
• Democracy Research
• Participation, representation, electoral
  choice, institutions, bureaucracy (Campbell,
  Converse, Stokes and Miller, 1960 Verba, Nie and
  Petrocik, 1972 and Fiorina, 1981). Freedom House
  Index, Polity IV (democracy/ autocracy),
  Governance Indicators.

3
Theoretical issues

• These approaches each fall into the trap of
  measuring the consequences or reflecting
  phenomena other than capabilities but have the
  advantage of measuring the phenomena directly.
• Measures of well being reflect government and
  elite choice in resource allocation
• Measures of governance favor democratic systems
  of government over other forms of governance.

4
Relative Political Capacity

• Measures the discretionary power of the
  government to implement a desired policy.
• The Relative Political Capacity of a government
  is the ratio of revenues a government extracts,
  compared to predicted levels a society could
  attain based on economic endowment.
• RPC Actual Government Revenue
• Predicted Government Revenue

5
Relative Political Capacity
6
Examples RPC across time
Bolivia
High RPC
Thailand
Botswana
Angola
1960
1980
1960
2000
1970
1990
1980
2005
1990
1970
2000
2005
1990
1990
1970
2000
1960
1970
1980
2005
1980
2000
Low RPC
2005
2000
4000
6000
8000
0
GDP per Capita
7
RPC 2000
8
RPE General Model

• Yit a ß Xit Vr
• Yit Adjusted tax revenue for country i at time
  t
• X it Vector of variables that determine
  potential tax collection
• Vr White noise disturbance

9
Inter country comparisonsModels for Developing
Societies

• Model 1
• Tax/Gdp a ß1 time - ß2 min/gdp ß3 gdpCap
  ß4 exports/gdp ß5 Oil ?
• Tax (TaxRev-NonTax-SocSec)/Gdp
• RPC1 Real /Predicted
• Note Control for expenditures
• Model 2
• Tax/Gdp a ß1 time - ß2 min/Gdp ß3 agr/Gdp
  ß4 exports/Gdp ß5 Oil ?
• Tax (TaxRev-NonTax-SocSec)/Gdp
• RPC2 Real /Predicted

10
Inter country comparisonsModels for Developed
Societies

• Model 3
• Tax/Gdp a ß1 time - ß2 min/gdp ß3 gdpCap
  ß4 Exports/gdp ß5 health/gdp ?
• Tax (TaxRev-NonTax-SocSec)/Gdp
• RPC3 Real /Predicted

11
Empirical Inventory

• National Level model
• Unit of analysis country
• Number of countries in the sample 129
• Time span 1960-2005
• Notes Sample includes Poland 86-on, Hungary
  89-on, Romania 88 on. These countries use NMP in
  their National Accounts before those dates so the
  data is not comparable to GNP/GDP methodology).
  The sample includes China.

12
Variables Sources
13
Data Collection Issues

• Data relies on national government accounts
• Variances in tax reporting and accounting are
  noted by IMF or WB, in most cases.
• Currency changes and differences in reporting
  timing in high inflation situations need to be
  normalized.
• National sources may be necessary to supplement

14
Measurement Conceptualization

• Health expenditures before 1990s mostly
  government reported public expenditures, now
  private are part of the mix in both developed and
  developing but not always reported
• Social Security 1) some countries have
  privatized part of SS.
• 2) The argument to exclude social security has
  been that those transfers have been allocated
  previous to the collection and therefore there is
  little room for any further manipulation by the
  government. This is not the case, those funds
  are not untouchable, and the deficit can be
  manipulated by Central Banks or Congress
• Agriculture Still our best empirical fit for
  developing countries (not for developed) but
  1960s is the anchor so we need to control for
  time.
• Taxation Indirect vs. direct taxation

15
Sub National Models

• Yit a ß Xit Vr
• Yit Adjusted tax revenue for country i at time
  t
• X it Vector of variables that determine
  potential tax collection
• Vr White noise disturbance
• Conceptually, the same model as national level.
• Constraints to choosing the variables are country
  specific

16
Relative Provincial Political Capacity (RPPC)

• Model 1
• Transf/gdp a ß1 time ß2 min/gdp ß3
  agr/gdp ß4 rev/gdp ?
• Model 2
• Transf/gdp a ß1 time ß2 gdpCap ß3 rev/gdp
  ß4 min/gdp ?
• Model 3
• Rev/gdp a ß1 time ß2 transf/gdp ß3
  agr/gdp ß4 min/gdp ß5 Subsidies ?

17
Empirical Inventory

• Sub National Level Model
• Unit of analysis province/state/region
• Time span Ideally 1990 2005, dependant on
  individual country
• Bolivia, Brazil, China, EU, India, Indonesia,
  Iran, Mexico, Nigeria, Thailand, Sudan, USA,

18
Data Collection

• In many countries the transfers are decided
  according to a Congressional Allocation Funds
  formula. If the allocations are decided on
  annual negotiations (1), the variable captures
  the concept of RPPC. If the formula is
  inflexible (2), the flexibility comes from the
  allocation of central government expenditures to
  the provinces.
• It is also important to know that the General
  Appropriation Act does not show if or how much of
  the money is released, so instead of following to
  entry number (Transfers, Subsidies. Models 1 3)
  in some cases is better to look at the exit
  number (Expenditures. Model 2)

19
Variables RPPC
20
Sources Examples RPPC
21
Applications

• International Conflict
• Organski Kugler 1980 Lemke 1996 Tammen et al
  2002.
• Internal Conflict
• Kugler et al 1997 Benson Kugler 1998 Johnson
  2007.
• Demographic Transitions
• Arbetman, Kugler Organski 1980, 1994, 1999
  Feng Kugler Zak 2000, 2007.
• Economic Applications
• Exchange Rates Arbetman 1990 Growth Leblang
  1997 Inflation Alcazar 1997 Private
  Investment Feng Chen 1997, Feng 2004 Policy
  Implementation Snider 1997 Informal Markets
  Arbetman 1990, 1994 Arbetman Ghosh 1997.
• Sub National
• Rouyer 1997 Swaminanthan 2005 2007 Johnson
  2007.

22
Relative Political Capacity

• RPC at different levels of development