Transcript Getting smart about subsidies

Getting smart about subsidies
Markus Goldstein
The World Bank
Subsidies are all around us



Obvious ones such as agriculture, urban
transport, fuel
But think about the air we breathe (e.g.
subsidies for pollution control), the water we
drink (subsidies for water utilities)
Not all subsidies are bad, but in order to tell
the good from the bad we need to think
smart.
How do we think smart about
subsidies?
Think systematically
 What are subsidies and why are they
important?
 Why do we subsidize?
 Who gets the subsidies?
 How do we implement subsidies?
Why are subsidies important?

Subsidies are political
Why are subsidies important?

A lot of money is spent on subsidies



In 2005, the EU was spending 2 euros per cow,
per day (puts them above the poverty line)
Egypt spent an estimated 6.9% of GDP on fuel
subsidies and 1.6% on food subsidies in 2005/6
Subsidies can have big policy effects

Malawi removed all primary school fees,
enrollment went up by 1 million children over 6
years, and the enrollment gap between rich and
poor closed
We often view subsidies too
narrowly



The tendency is to think in terms of
producer subsidies (e.g. agriculture) but
they apply to both supply & demand
A subsidy is when the price is lower
than the cost, where the government
provides the funds
And a tax is a negative subsidy
Why subsidize?
Do we know why we
subsidize?

Take the case for subsidizing primary health care:






Some would argue it’s a right, so we do it for equity reasons
Others would argue market failure, e.g. incomplete information
Still others would argue that it is an externality – healthier people
are better able to participate in society
So the answer is both yes and no – we can see a set of
justifications, maybe even more than one for a given subsidy
To understand whether a subsidy is working or not, we need to
know why -- the main reason why -- we did it in the first place
Understanding why a subsidy is in place, also provides grounds
for taking the debate into the political realm and dealing with
fiscal constraints
So why do we actually
subsidize?




Markets aren’t working right – efficiency
We care about the poor – equity
There may (or may not) be tradeoffs
across these
These are tempered by the hard facts
of real constraints, in particular:


Fiscal constraints
Politics
Making markets work right

Externalities


e.g. government subsidies to identify and treat
tuberculosis
Special case of externalities is public goods


e.g. government funds (100% subsidy) national defense
Market failures due to information issues

e.g. the rationale for subsidizing microcredit
Promoting equity

Best option: give the poor the cash.
But:




Administration costs are too high
Possible political constraints
So we subsidize…
But we need to consider what we mean
by equity
An equity/efficiency tradeoff?
Often, but not always

Reasons there might not be an
equity/efficiency tradeoff:

When correcting market failure is not possible or
too costly, some redistribution can improve
efficiency

e.g. think about subsidizing education, Psacharopoulos
(1994) estimates returns to primary education of 24
percent in Africa and 20 percent in Asia…we could try to
create a credit market to pay for primary education but
this is really not feasible.
Not always a tradeoff -2

Getting rid of some kind of subsidies can
improve both equity and efficiency


e.g. In Japan, the average farm hh income is
110% of average hh income, yet over 2000-2
producers received $47.5 bn in subsidies
Between 1996-2000, the US government
mailed farm subsidy checks to 31 addresses in
Beverly Hills and 803 in Washington D.C.
For better or worse, equity and
efficiency run into politics and
real budget constraints…
3 questions to assess the
political constraints
Who are the winners/losers and who
will they vote for (if they vote)?
 Who are the interest groups around this
issue and how much power do they
have?
 What is the time horizon of policy
makers?
 Outcome need not be fair nor efficient

Politics in action: food
subsidies in Sri Lanka

Food subsidy, almost universal from
independence



Maintained by gov’ts of different ideology
in the name of political stability and equity
Broad based political support, particularly
from organized labor
Subsidies to producers as well to keep
them happy
And making things better
could make them worse…

1978-9 large devaluation, new gov’t





Implemented means test  ½ the pop.
Change from rations to food stamps and froze new issues
and nominal budget allocation
Removed price subsidies
Protected wage earners who were no longer getting
subsidies through inc. wages (interest groups)
But then, benefits erode with inflation…political
constituency? (time horizon of policy makers,
beneficiaries did not have enough voting power)
Three ways to pay the bill
1. Tax and redistribute

e.g. Health in Finland



taxes from the state (progressive income and
indirect) and municipalities
Patients pay only 24%, the remainder is
financed by local (43%) and national gov’t
(18%) and national health insurance (15%)
Over 80% of Finns are satisfied with their
health care, against EU average of 41%
Paying the Bill (2 )
2. Cross-subsidize



Can be within or across sector
Gabon: urban electricity subsidizes water and electricity in
rural areas and small towns.
Ecuador: surcharge on telecom subsidizes water
3. Grants

e.g. Tajikistan Pamir Power Project



Poor people, low capacity, irregular supply
Concession put in place, but move to market prices for power
immediately not feasible (500% increase required)
Mix of IDA, Swiss and IFC funding for 10 year gradual phase
out of subsidized prices
Some things you need to
know to pay the bill:
1. Tax and
redistribute
2. Cross
subsidize
3. Grants
Tax incidence, tax costs
Incidence of benefits, now
and post change
Elasticity of demand – both
for suppliers of funds and
beneficiaries
Size of financing gap
Funding source
Exit strategy, since not
sustainable

Do we know who we
subsidize?
Primary education is progressive, but higher education is not
Country
Early
Childhood,
Preschool
Primary
Secondary
Tertiary
Bolivia
(2002)
Progressive
Regressive
Regressive
Cambodia
(1997)
Neutral
Regressive
Regressive
Cambodia
(2002)
Progressive
Regressive
Regressive
Ecuador
(1999)
Progressive
Bell Curve
Regressive
Progressive
Bell Curve
Regressive
Neutral
Regressive
Regressive
Jamaica
(2000)
Progressive
Madagascar
(2001)
Overall
Neutral
Mexico
(2002)
Progressive
Progressive
Regressive
Regressive
Neutral
El Salvador
(2002)
Progressive
Progressive
Regressive
Regressive
Progressive
Vietnam
(1998)
Progressive
Regressive
Vietnam
(2002)
Progressive
Regressive
Health varies…
Country
Type
Distribution
Bolivia (1999)
Social Insurance
Regressive
Public Expenditure
Progressive
Health Care IESS
Regressive
Health Care MSP
Relatively Neutral
Health Care SSC
Progressive
Mexico (2002)
All Public Health Expenditure
Regressive*
El Salvador
(2002)
Hospital Care
Relatively Neutral
Primary Health Care
Progressive
Ecuador (1999)
Jamaica (19912002)
Average Ratio of Poorest to Richest
Quintile for Public Hospital Use
(Q1/Q5)
1.83
Jamaica (19912002)
Average Ratio of Poorest to Richest
Quintile for Primary Care Use
(Q1/Q5)
0.82
Jamaica (19912002)
Average Ratio of Poorest to Richest
Quintile for Outpatient Care (Q1/Q5)
1.76
Infrastructure subsidies are clearly regressive…
progressive
regressive
So who do we subsidize?


First step is to look at incidence data –
what share of the benefits go to an
income group?
But both the benefits and the costs
(taxes) to the individual matter
At the national level, adding them up is key…
Mexico: Benefits and Taxes
20
10
0
1
2
3
4
5
6
Percent
-10
-20
-30
Benefits
Taxes
-40
-50
-60
Decile
7
8
9
10
To estimate the net effect of public subsidies - tax
Mexico: Net Effect
100
80
% of hh income
60
40
20
0
1
2
3
4
5
6
-20
-40
decile
7
8
9
10
How do we implement
subsidies?
The rationale for the subsidy
tells us where to start…

If we are doing
subsidies for equity
reasons we need to
target both:


Access
Affordability

If subsidies are
justified for
efficiency concerns,
we need to target:


Overcoming/reducing
a market distortion
Changing behavior
And what we need to know

Know your poor in order to achieve
access and affordability

Know your markets (including public
goods) when trying to remove
distortions

Know your elasticity when trying to
change behavior
We have some (potentially)
powerful tools

Know your poor (access and affordability)


Know your markets (remove distortions)


Market assessments, CEM, PSIA, WTP
studies
Know your elasticities (change behavior)


poverty assessments, poverty maps
Market assessments, ESW, WTP studies
Data is key – household, investment climate,
market SURVEYS
But what about the
constraints? (1)

Political economy

Know your polity



PSIA
Keep in mind the political economy can be an
opportunity, not only a constraint
Use politics to put subsidies on the table and in
perspective:


e.g. Andhra Pradesh: introduced metering so that
subsidy was visible, transparent
e.g. Punjab: free power, but part of a fixed agricultural
budget
What about the constraints (2)

Fiscal

Know your budget, know your incidence
(don’t forget the taxes)


CEM, PER
For equity: think about real tradeoffs,
balance inclusion/exclusion
What can make subsidies go
wrong (or right)?
We need to avoid bad design

Badly designed – Information was
inadequate/wrong

Within the relevant market:


Lack of attention to other markets


e.g. Cote d’Ivoire water
e.g electricity/irrigation in India
Question of provider

Keep in mind subsidies are about gov’t finance,
not provision
And the environment matters
too

Constraints change (good design
initially)



Budget
Political economy
e.g. Sri Lanka food

Budget constraints tighten


Possible opportunity to narrow targeting
Or political economy binds and benefits erode
In the end good design can
have a powerful impact

The Bangladesh secondary school stipend
 Tuition and stipend subsidy for girls for
educational costs, supplies, uniforms and
transport.
 Good design: priced well (~1/2 of costs), tied to
attendance. The subsidy was put very close to
the good it was aimed at.
 Covered 1.9 million girls in 2004, for around $25
million
 Resulted in enrollment increases of 8-12
percentage points.
To take stock:


Not all subsidies are bad… Bad subsidies are
bad.
To separate the good and the bad, think
systematically




Understand why we have this subsidy
Who is the subsidy for, who is getting it?
How is this subsidy improving access,
affordability, reducing distortions, or changing
behavior. Is this the best feasible way?
Benchmark on the counterfactual of no
intervention  make an informed decision
Now how do we decide what
to do?
Why?
Equity
Efficiency
Equity
Will a lump sum transfer work?
yes
No subsidy
no
Subsidy for whom?
Access
Affordability
 Identify targeting mechanism
 Identify political constraints
and opportunities
 Identify fiscal constraints &
tradeoffs
Identify
target
poor
Efficiency
Behavior
• know
elasticities
• identify
closest and
least
distortionary
policy handle
Reduce market distortions
Externalities &
other MF
Public
Goods
• know markets
• gov’t funds
these
• identify
closest and
least
distortionary
policy handle
• assess
optimal level
of provision
• identify funding mechanism, fiscal constraints
and tradeoffs
This is a PREM task


Subsidies are everywhere
PREM economists have a central role in
thinking about them systematically




Understanding within sector
Adding up across sectors and taxes
Helping identify tradeoffs for the gov’t
We have an existing set of tools that allow
us to tackle these issues