Distributional Impacts of Congestion Pricing

1
Distributional Impacts of Congestion Pricing

• Douglass B. Lee, Jr.
• International Symposium on Road Pricing
• Key Biscayne, FL
• November 2003

2
Purpose

• Efficiency versus Equity
• Equity
• horizontal
• vertical
• Objective Measurement (vs. eye of beholder)
• direct impact (as a tax)
• relative to current financing methods (recycling)
• after behavioral response (incidence)

3
Two Major Dimensions of Equity

• Horizontal
• treat equals equally
• Vertical
• impact on the distribution of income reduce
  disparity (progressive), increase disparity
  (regressive), or neutral
• impact on income class proportional to income is
  neutral
• not everything is regressive

4
Conclusions

• Vertical equity impacts of peak pricing are not
  bad, not enough to be an obstacle to efficiency
• true before revenue recycling revenues can
  easily correct equity impacts (and also kill
  efficiency gains)
• Alternative instruments are also not that bad,
  e.g., gasoline excise tax, property tax
• Earmarking may be politically useful but is
  economically dubious if revenues are fungible
  (e.g., gambling revenues for education)

5
1. Direct Impact

• Definition of income
• household, family, consumer unit, per capita
• money, earned income, before tax, after
  transfers, permanent income, wealth
• Method of Measurement
• tax returns
• expenditures
• self-reported (c. 15 no response for NPTS)
• categories, quantiles, density function

6
Example Raw Income Data
7
1999 Household Income
source 2000 Current Population Survey, Census
Bureau
8
NPTS 1995
9
Shares of Income by Decile
10
Peak Congested Trips

• Selected NPTS trips
• mode auto, SUV, van, or pickup
• start time 630 to 930 AM or 330 to 700 PM
• travel day weekday
• place urban, suburban, or second city (not town
  or rural or undetermined)
• n 52,000 out of 409,000 trips
• Ignores
• peak direction
• trip length
• degree of congestion (demand vs. capacity)
• occupancy

11
Tax on Peak Trips

• Each trip charged a fixed amount
• same amount (toll) for each trip
• Tax is borne entirely by any person currently
  driving in the peak
• whether they stay or change is not considered
  (yet)

12
Distribution of Peak Trips
13
Peak Trips versus Income
14
Distributional Analysis

• Restate impact of policy as a percent of income
  for each income class
• Horizontal line is neutral falling (tax) is
  regressive

15
Peak Trips as Percent of Income
16
Observations

• Tax impact generally rises with income, but not
  as rapidly
• A tax per trip on existing peak travelers is
  mildly regressive, according to these data
• Data are noisy, and noise tends to flatten the
  distribution (i.e., toward random)
• Permanent income reduces regressiveness

17
Shares of SR-91 Users by Income Group
18
Histogram of SR-91 Users by Income
19
Histogram with Fitted Density Function
20
Distribution Fitted to 1999 Household Income
21
(No Transcript)
22
Relative Burden Function
23
Direct Burden for SR-91
24
2. Recycling of Revenues

• Fuel excise tax, general sales tax, local
  property tax reduction
• Consumption of gasoline and oil as percent of
  income/expenditures
• excise tax is proportional to gallons, not to
  dollar amount paid (i.e., not a sales tax)
• Compare revenue-neutral policies

25
Distribution of Gas/Oil Expense
26
Gas and Oil Expenditures by Decile
27
Gasoline Expenditures as of Income
28
Comparison of Two Taxes
29
3. Shifting and Incidence

• Individuals who change behavior reduce the burden
  of the tax/toll on themselves those who stay and
  pay bear the full brunt
• More careful analysis has been done for national
  tax instruments, but not for congestion tolls or
  value pricing
• Behavioral responses can be estimated, but data
  are very few

30
Shifting the Burden

• If employers pay commuting, burden is shifted to
  owners and customers
• If costs are tax deductible, burden is diffused
  among taxpayers
• Behavior changes to avoid the tax/toll cause some
  redistribution of the burden
• Not zero-sum delay is deadweight loss
• e.g., tax on luxury (gt100 feet) yachts

31
Supply-Demand Elasticities
32
Behavior Changes

• No change
• high time-value travelers may prefer time savings
  with toll
• Shift route
• increases time for traveler, impacts travelers on
  parallel routes
• Shift to carpool or transit
• may leave traveler better off (cost plus time),
  impacts other modes
• Shift time
• traveler worse off than before but less than
  paying the toll
• Shift locations
• workplace or residence or other destination

33
Correcting Vertical Equity Impacts

• Use of revenues from congestion pricing (Small
  rule)
• reduce general taxes or other user fees
• improve attractiveness of alternative
  transportation modes
• redistribute revenues according to income (low
  income receives largest compensation)