Transcript fixed effects

Regulatory Governance and Electricity
Investment in Developing Countries
Jon Stern
London Business School and City
University, London
March 2006
Regulatory Governance and Infrastructure
Industry Performance
• By 2000, regulatory agencies independent of
Ministries had been established
 for over 100 countries in telecoms
 for over 50 countries in electricity/energy
• Since 2000, many more have been established – and
many more Ministry regulators operate under powers
and duties established in a regulatory law
• Has this made a difference to industry performance?
Regulatory Governance and Infrastructure
Industry Performance: Telecoms
• A number of panel data studies for telecoms have shown that
better quality regulatory governance:
– Significantly increases investment in fixed lines (Gutierrez 2003 –
by about 20-30% in long run)
– Significantly increases efficiency as measured by labour
productivity (Gutierrez 2003 )
– If in place pre-privatisation, an independent regulator significantly
increases privatisation revenues (Wallsten 2003)
• And now, we have the results from similar research for the
impact of regulatory governance on developing country
generation capacity per head
(See Cubbin & Stern WB Policy Research Working Paper No 3535 and
WBER forthcoming)
Regulatory Governance and its Expected
Impact on Investment
For commercialised electricity companies (ie with
private investment or finance of investment on
commercial terms), we would expect:
1)
2)
3)
Establishment of explicit regulatory framework to result in
higher investment and capital stock;
The better the regulatory governance framework, the
larger the expected increase in investment and capital
stock; and
Larger impacts on investment and capital stock over time
as the regulator gains experience and reputation.
Are these confirmed when tested?
Data Used to Test Impact of Regulatory
Governance I
•
We have data on generation capacity (MW) for 28 developing countries
for a 21 year period 1980-2001
– 15 Latin American countries, 6 in Caribbean, 4 in Asia and 5 in
Africa.
•
Of the 28 countries, 26 had passed an electricity regulatory law by
2001 – but only 5 before 1995
•
By 2001, 17 countries had an autonomous regulator (all with a
regulatory law) and 12 countries had a Ministry regulator (all but 2 with
a regulatory law)
– Almost half the autonomous regulators were under 3 years old in
2001 (including all the African ones) while around 30% were over
10 years old
– The median age (50% point) of the autonomous regulators was just
under 5 years in 2001
Data Used to Test Impact of Regulatory
Governance II
• We also had regulatory data on
– Whether the regulator (Ministry or autonomous) was funded by
licence fees or from central government funds
– Whether the regulator was obliged to pay staff on civil service pay
scales
– The year in which the regulatory law was enacted
(All the regulatory data came from Preetum Domah 2001 survey)
• Other data used:
– GDP, population, debt levels, industry value added
– Country governance variables
– Henisz et al dated information on privatisation (minority,
majority and full) and “competition”
Measures of Regulatory Governance
For our 4 governance indicators with known starting dates: [Law
(Yes/No), Autonomous Regulator (Yes/No), Licence fee funding
(Yes/No), non-Civil Service Pay Scales (Yes/No)], we can:
a)
Test effect of each individually – OK as staring point but
leads to over-estimate of effect of each
b)
Combine in a regulatory index – for regulatory index, each
country is scores either Zero or 1,2, 3 or 4 in each year
A country switching from standard, centrally funded Ministry
regulator to autonomous, licence fee funded regulator in 1995 is
scored zero on the Index from 1980 to 1995 and either 3 or 4
from 1996 through 2001
Both measures provide useful information
Estimation Method
Since we have data for 28 countries over 20 years we can use
panel data methods to estimate a fixed effects model.
What does this mean?
It means that for each country, we can estimate a country
specific fixed effect which captures all the particular features of
that country relevant to energy use (eg climate, fuel use, etc),
institutional quality (courts, corruption, etc)
Resulting estimates show impact of regulatory governance on
generation capacity & investment controlling for observable
and unobservable country specific determinants of
generation investment
Main Electricity Regulatory Governance
Results
• Regulatory governance does matter
In long-run (ie after about 10 years or more), best quality
regulatory governance associated with about 15-20% higher
generation capacity per head
– Each 1 point increase in index implies 4-5% increase in generation
capacity per head in long run
• But, effects take time to build up
– Very little effect for 1st 3 years, under half final effect after 5 years
• Biggest single impact from having a regulatory law in place,
followed by licence fee funding
– Autonomy of regulator less powerful – may be consequence of high
proportion of only recently established autonomous regulators
Results on GDP Level, Privatisation,
Competition and Country Governance
• A 10% increase in real GDP per head (exchange rate basis)
associated with a 7-8% higher generation capacity per head in
long run
• Majority or full privatisation associated with 12-20% higher
generation capacity per head in long run
– ‘Competition’: legal right for IPPs to generate for resale associated
with 14% higher generation capacity per head in long run
– Both – particularly ‘competition’ variable - may primarily reflect
country’s commitment to electricity reform
• Country governance important but not hugely
– CHECKS political risk index significant but much smaller effect than
electricity regulatory variables
– Weak evidence of some effect of Kaufmann Rule of Law index
Implications of Results
How should we interpret these results?
•
The results mean that an average developing country with an average
fixed effects score could expect enough extra investment to give 1520% higher generation capacity per head after about 10 years
•
This would be with an average quality law, a regulator with an average
staffing levels, average country governance, etc – and an average
(unobservable) fixed effects score
•
Countries with above average quality laws, staffing levels, country
governance, etc could expect larger impacts than 15-20%
•
Countries with below average quality laws, staffing levels, country
governance, etc should expect smaller impacts than 15-20%
Final comments
• The results provide strong empirical support for arguments that
good electricity governance genuinely matters in practice for
electricity generation investment levels
– Results stood up with extensive testing of variants and of statistical
assumptions
– Confirms results from case studies
• Other results suggest that better regulatory governance
improves generation availability but only by a small amount
– Efficiency effects also likely to be present but we couldn’t test
• Institutions matter in practice – particularly their design and
governance quality
– This is true not just for electricity. This has been shown to be true
for economic growth, for telecoms and for other industries with
high capital requirements and sunk costs.
Links to Supporting Papers
Main Paper discussed:
Cubbin, J.S and Stern J., (2004), “Regulatory Effectiveness: The
Impact of Good Regulatory Governance On Electricity Industry
Capacity And Efficiency In Developing Countries”
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=695385
Final version World Bank Economic Review forthcoming, with (paid-for)
online availability
See also
Stern, J. and Cubbin J.S. (2003), “Regulatory Effectiveness: The
Impact of Regulation and Regulatory Governance Arrangements on
Electricity Outcomes – A Review Paper”
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=695386