ArticlePDF Available
Should we build more large dams? The actual costs of hydropower
megaproject development
$
Atif Ansar
a,b,
n
, Bent Flyvbjerg
b
, Alexander Budzier
b
, Daniel Lunn
c
a
Blavatnik School of Government, University of Oxford, Oxford OX1 4JJ, UK
b
Saïd Business School, University of Oxford, Oxford OX1 1HP, UK
c
Department of Statistics, University of Oxford, Oxford OX1 3GT, UK
HIGHLIGHTS
We investigate ex post outcomes of schedule and cost estimates of hydropower dams.
We use the outside viewbased on Kahneman and Tversky's research in psychology.
Estimates are systematically and severely biased below actual values.
Projects that take longer have greater cost overruns; bigger projects take longer.
Uplift required to de-bias systematic cost underestimation for large dams is þ99%.
article info
Article history:
Received 25 February 2013
Accepted 29 October 2013
Available online 10 March 2014
Keywords:
Large hydropower dams
Schedule and cost estimates
Costbenet forecasting
Reference class forecasting
Outside view
abstract
A brisk building boom of hydropower mega-dams is underway from China to Brazil. Whether benets of
new dams will outweigh costs remains unresolved despite contentious debates. We investigate this
question with the outside viewor reference class forecastingbased on literature on decision-making
under uncertainty in psychology. We nd overwhelming evidence that budgets are systematically biased
below actual costs of large hydropower damsexcluding ination, substantial debt servicing, environ-
mental, and social costs. Using the largest and most reliable reference data of its kind and multilevel
statistical techniques applied to large dams for the rst time, we were successful in tting parsimonious
models to predict cost and schedule overruns. The outside view suggests that in most countries large
hydropower dams will be too costly in absolute terms and take too long to build to deliver a positive risk-
adjusted return unless suitable risk management measures outlined in this paper can be affordably
provided. Policymakers, particularly in developing countries, are advised to prefer agile energy
alternatives that can be built over shorter time horizons to energy megaprojects.
&2014 Elsevier Ltd. All rights reserved.
1. Large hydropower dam controversy
The 21st Century faces signicant energy challenges on a global
scale. Population and economic growth underpin increasing demand
for energy from electricity to transport fuels. Social objectives of
poverty alleviation, adaptation and mitigation of climate change, and
energy security present policy makers and business leaders with
difcult decisions and critical trade-offs in implementing sound
energy policies. Demand for electricity is, for example, slated to
almost double between 2010 and 2035 requiring global electricity
capacity to increase from 5.2 terawatt (TW) to 9.3 TW over the same
period (IEA, 2011). Currently, the de facto strategic response to these
big energy challenges is big solutionssuch as large hydropower
dams. Are such big solutions in general and large hydropower dams
in particular the most effective strategy, on a risk-adjusted basis, to
resolve global energy challenges? Might more numerous small
interventions be more prudent from the perspective of risk manage-
ment and maximizing net present value even when they entail
somewhat higher per unit cost of production?
Proponents of large dams envisage multiple benets. A big step-
up in hydropower capacity along with a long and varied list of
corollary benets: reducing fossil fuel consumption, ood control,
irrigation, urban water supply, inland water transport, technological
progress, and job creation (Billington and Jackson, 2006;ICOLD,
2010). Inspired by the promise of prosperity, there is a robust
Contents lists available at ScienceDirect
journal homepage: www.elsevier.com/locate/enpol
Energy Policy
0301-4215/$ - see front matter &2014 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.enpol.2013.10.069
This is an open-access article distributed under the terms of the Creative
Commons Attribution-NonCommercial-No Derivative Works License, which per-
mits non-commercial use, distribution, and reproduction in any medium, provided
the original author and source are credited.
n
Corresponding author. Tel.: þ44 186 528 8795; fax: þ44 186 528 8805.
E-mail addresses: atif.ansar@bsg.ox.ac.uk,atifansar@gmail.com (A. Ansar).
Energy Policy 69 (2014) 4356
pipeline of new mega-dams being developed globally after a two-
decade lull. The Belo Monte dam in Brazil, the Diamer-Bhasha in
Pakistan, Jinsha river dams in China, Myitsone dam in Myanmar, or
the Gilgel Gibe III dam in Ethiopia, all in various stages of develop-
ment, are unprecedented in scale.
Large dams are, however, controversial because they exert
substantial nancial costs (World Bank, 1996;World Commission
on Dams, 2000). Beyond the nancial calculus, large dams have
profound environmental (McCully, 2001;Scudder, 2005;Stone,
2011), ecological (Nilsson et al., 2005;Ziv et al., 2012), and social
(Bakker, 1999;Duo and Pande, 2007;Richter et al., 2010;
Sovacool and Bulan, 2011) impacts. Stone (2011, p. 817) reports
in Science that the Three Gorges dam in China is an environ-
mental banethat will cost over USD 26.45 billion over the next 10
years in environmental mitigation efforts. Despite their outsized
nancial and environmental costs, the purported benets of large
hydropower dams prove uncertain. For example, the World
Commission of Dams (2000, p. 30) reported that for large hydro-
power dams average [hydropower] generation in the rst year of
commercial operation is 80% of the targeted value”—a trend of
which the recently completed Bakun hydroelectric project in
Borneo is an alarming example (Sovacool and Bulan, 2011).
Similarly, Duo and Pande (2007) nd adverse distributional
impacts of large irrigation dams in India. Winners downstream
come with losers upstream yielding a more modest, if any, net
economic benet.
The scale of contemporary large dams is so vast that even for a
large economy such as China's the negative economic ramications
could likely hinder the economic viability of the country as a whole
if the risks inherent to these projects are not well managed (Salazar,
2000). Similarly, Merrow et al. (1988, pp. 23) warn that such
enormous sums of money ride on the success of megaprojects [such
as large dams] that company balance sheets and even government
balance-of-payments accounts can be affected for years by the
outcomes. Such warnings are not idle alarmism. There is mounting
evidence in civil society, academic research, and institutional
accounts that large dams have strikingly poor performance records
in terms of economy, social and environmental impact, and public
support (McCully, 2001;Scudder, 2005;Singh, 2002;Sovacool and
Bulan, 2011; WCD, 2000). There are acrimonious, and as yet incon-
clusive, debates in scientic literature and civil society about whether
large dams are a boon or a curse. Should we build more large
hydropower dams? How condent can planners be that a large bet
on a large dam will pay-off handsomely?
We investigate these questions with the outside viewor
reference class forecastingbased on the literature on decision-
making under uncertainty that won Princeton psycholo-
gist Daniel Kahneman the Nobel Prize in economics in 2002
(Kahneman and Tversky, 1979a,1979b;Kahneman, 1994) extended
and applied by Bent Flyvbjerg and colleagues to infrastructure
projects (Flyvbjerg et al., 2003;Flyvbjerg, 2009). We present
statistical and comparative evidence from the largest reference
class to-date of actual costs of large hydropower dam projects
(hereafter large dams unless stated otherwise). We nd that even
before accounting for negative impacts on human society and
environment, the actual construction costs of large dams are too
high to yield a positive return. Large dams also take inordinately
long periods of time to build, making them ineffective in resolving
urgent energy crises. Our evidence pertains primarily to large dams
and the results cannot be applied either to smaller dams or other
large energy solutions such as nuclear power without rst building
a separate reference classfor other types of power generation
technologies. Our ndings, however, point towards the general-
izable policy proposition that policymakers should prefer energy
alternatives that require less upfront outlays and that can be built
very quickly.
There is no doubt that harnessing and managing the power of
water is critical for economies but large dams are not the way to
do so unless suitable risk management measures outlined in this
paper can be affordably provided. Building on literature in decision
making under uncertainty in management, psychology, and
planning research, this paper further provides public agencies
(e.g. national planning and nance ministries, power and water
authorities), private entrepreneurs, investors, and civil society a
framework to test the reliability of ex ante estimates for construc-
tion costs and schedules of power generation alternatives. An
impartial and rigorous application of the reference class forecast-
ing methods proposed here can improve the selection and
implementation of new investments.
2. Delusion and deception in large hydropower dam planning?
Our approach to address the debates about whether or not to
build dams is to incorporate an evidence-based perspective that
reects how decisions among alternative options are actually
made and on what basis. Theoretical and empirical literature on
decision-making under uncertainty proposes two explanations
psychological delusion and political deceptionthat suggest deci-
sion-makers' forecasts, and hence ex ante judgments, are often
adversely biased (Tversky and Kahneman, 1974;Kahneman and
Lovallo, 1993;Flyvbjerg, 2003;Lovallo and Kahneman, 2003;
Kahneman, 2011).
First, experts (e.g., statisticians, engineers, or economists) and
laypersons are systematically and predictably too optimistic about
the time, costs, and benets of a decision. This planning fallacy
(Kahneman and Tversky, 1979b;Buehler et al., 1994) stems from
actors taking an inside viewfocusing on the constituents of the
specic planned action rather than on the outcomes of similar
actions already completed (Kahneman and Lovallo, 1993). Thus, for
example, the estimated costs put forward by cities competing to
hold the Olympic Games have consistently been underestimated
yet every four years these errors are repeated. Biases, such as
overcondence or overreliance on heuristics (rules-of-thumb),
underpin these errors.
Second, optimistic judgments are often exacerbated by decep-
tion, i.e. strategic misrepresentation by project promoters (Wachs,
1989;Pickrell, 1992;Flyvbjerg et al., 2002,2005,2009). Recent
literature on infrastructure delivery nds strong evidence that
misplaced political incentives and agency problems lead to awed
decision-making (see Flyvbjerg et al., 2009). Flyvbjerg et al. (2009, p.
180 ) further discuss that delusion and deception are complemen-
tary rather than alternative explanations for why megaprojects
typically face adverse outcomes. It is, however, difcult to disen-
tangledelusion from deception in practice. Using quasi-
experimental evidence from China, Ansar et al. (2013) suggest that
while better incentive alignment can help to lower the frequency
and, to a lesser extent, the magnitude of biases, it does not entirely
cure biases.
Be it delusion or deception, is decision-making in large hydro-
power dams systematically biased by errors in cost, schedule, and
benet forecasts? What is the risk that costs might outweigh
benets for a proposed dam? While the future is unknowable,
uncertain outcomes of large investments can still be empirically
investigated using reference class forecasting(RCF) or the out-
side viewtechniques (Kahneman and Lovallo, 1993;Flyvbjerg,
2006,2008). To take an outside view on the outcome of an action
(or event) is to place it in the statistical distribution of the
outcomes of comparable, already-concluded, actions (or events).
The outside view has three advantages: First, it is evidence-based
and requires no restrictive assumptions. Second, it helps to test
and t models to explain why the outcomes of a reference class of
A. Ansar et al. / Energy Policy 69 (2014) 435644
past actions follow the observed distribution. Third, it allows to
predict the uncertain outcomes of a planned action by comparing
it with the distributional information of the relevant reference
class. The theoretical foundations of the outside view were rst
described by Kahneman and Tversky (1979b) and later by
Kahneman and Lovallo (1993) and Lovallo and Kahneman (2003)
as means to detect and cure biases in human judgment. The
methodology and data needed for employing the outside view, or
reference class forecasting, in practice were developed by
Flyvbjerg (2006,2008) in collaboration with the Danish consulting
rm COWI (see Flyvbjerg and COWI, 2004).
2.1. Three steps to the outside view
The outside view, applied to large dams for the rst time here,
involves three steps: (i) identify a reference class; (ii) establish an
empirical distribution for the selected reference class of the
parameter that is being forecasted; (iii) compare the specic case
with the reference class distribution. We take a further innovatory
step of tting multivariate multilevel models to the reference data
to predict future outcomes. Our technique is an important
improvement in the methodology of the outside view that can
be generalized and applied to other large-scale and long-term
decisions under uncertainty. With de-biased forecasts managers
can make empirically and statistically grounded, rather than
optimistic, judgments (Dawes et al., 1989;Buehler et al., 1994;
Gilovich et al., 2002).
The outside viewas implemented by Flyvbjerg (2006,2008)
is not without limitations (see Sovacool and Cooper, 2013 for a
discussion specically about energy megaprojects). For example,
RCF focuses on generic risk inherent in a reference class rather
than specic project-level risk. We rectify against this limitation
by tting regression models in addition to using traditional
RCF methods in the result section below. Sovacool and Cooper
(2013, p. 63) further suggest that RCF may not provide sufciently
accurate indication of the risks of rare megaprojects the likes of
which have never been built before. Such out of the sample
problems are well noted in probability theory. They do not,
however, deny the fundamental usefulness of RCF. If anything
our results err towards conservative estimates of actual cost
overruns and risks experienced by large dams.
2.2. Measures and data
Following literature on the planning fallacy (Sovacool and
Cooper, 2013), the parameters central to our investigation and
multilevel regression analysis is the inaccuracy between man-
agers' forecasts and actual outcomes related to construction
costs, or the cost overrun, and implementation schedule, or
schedule slippage. Following convention, cost overrun is the
actual outturn costs expressed as a ratio of estimated costs
1
;cost
overruns can also be thought as the underestimation of actual
costs (Bacon and Besant-Jones, 1998;Flyvbjerg et al., 2002).
Schedule slippage, called schedule overrun, is the ratio of the
actual project implementation duration to the estimated project
implementation. The start of the implementation period is taken
to be the date of project approval by the main nanciers and the
key decision makers, and the end is the date of full commercial
operation.
Inaccuracies between actual outcomes versus planned forecasts
are useful proxies for the underlying risk factors that led to the
inaccuracies. For example, cost overruns reduce the attractiveness
of an investment and if they become large the fundamental
economic viability becomes questionable. Bacon and Besant-
Jones (1998, p. 317) offer an astute summary:
The economic impact of a construction cost overrun is the
possible loss of the economic justication for the project. A cost
overrun can also be critical to policies for pricing electricity on
the basis of economic costs, because such overruns would lead
to underpricing. The nancial impact of a cost overrun is the
strain on the power utility and on national nancing capacity in
terms of foreign borrowings and domestic credit.
Similarly, schedule slippages delay much needed benets,
expose projects to risks such as an increase in nance charges,
or creeping ination, which may all require upward revision in
nominal electricity tariffs. Financial costs and implementation
schedules, because of their tangibility, are also good proxies for
non-pecuniary impacts such as those on the environment or on
the society. Projects with a poor cost and schedule performance
are also likely to have a poor environmental and social track
record. A greater magnitude of cost and schedule overruns is thus
a robust indicator of project failure (Flyvbjerg, 2003).
In taking the outside view on the cost and schedule under/
overruns, our rst step was to establish a valid and reliable
reference class of previously built hydropower dams as discussed
above. The suggested practice is that a reference class ought to be
broad and large enough to be statistically meaningful but narrow
enough to be comparable (Kahneman and Tversky, 1979b;
Kahneman and Lovallo, 1993;Flyvbjerg, 2006). International
standard denesdamswithawallheight415 m as large. The
total global population of large dams with a wall height 415 m
is 45,00 0. There are 30 0 dams in the world of monumental scale;
these major damsmeet one of three criteria on height
(4150 m), dam volume (415 mil lio n m
3
), or reservoir storage
(425 km
3
)(Nilsson et al., 2005).
From this population of large dams, our reference class drew a
representative sample of 245 large dams (including 26 major
dams) built between 1934 and 2007 on ve continents in 65
different countriesthe largest and most reliable data set of its
kind. The portfolio is worth USD 353 billion in 2010 prices. All
large dams for which valid and reliable cost and schedule data
could be found were included in the sample. Of the 245 large
dams, 186 were hydropower projects (including 25 major dams)
and the remaining 59 were irrigation, ood control, or water
supply dams. While we are primarily interested in the perfor-
mance of large dam projects with a hydropower component, we
also included non-hydropower dam projects in our reference class
to test whether project types signicantly differ in cost and
schedule overruns or not. Fig. 1 presents an overview of the
sample by regional location, wall height, project type, vintage,
and actual project cost.
The empirical strategy of this paper relied on documentary
evidence on estimated versus actual costs of dams. Primary docu-
ments were collected from ex ante planning and ex post evaluation
documents of the:
1. Asian Development Bank;
2. World Bank, also see World Bank (1996) and Bacon and Besant-
Jones (1998);
3. World Commission of Dams (WCD), also see WCD (2000)
2
;
4. U.S. Corps of Engineers;
5. Tennessee Valley Authority;
1
Cost overruns can also be expressed as the actual outturn costs minus
estimated costs in percent of estimated costs.
2
Note that the World Bank, Asian Development Bank, and the WCD typically
report cost data in nominal USD. We, however, converted these data, adapting
methods from World Bank (1996: 85), into constant local currencies.
A. Ansar et al. / Energy Policy 69 (2014) 4356 45
6. U.S. Bureau of Reclamation, also see Hufschmidt and Gerin
(1970)
,3
and Merewitz (1973) on the U.S. water-resource con-
struction agencies.
The procedures applied to the cost and schedule data here are
consistent with the gold standard applied in the eldmore
detailed methodological considerations can be found in Flyvbjerg
et al. (2002),Federal Transit Administration (2003),Pickrell
(1989,1992),World Bank (1996) and Bacon and Besant-Jones
(1998) with which our data are consistent. All costs are total
project costs comprising the following elements: right-of-way
acquisition and resettlement; design engineering and project
management services; construction of all civil works and facil-
ities; equipment purchases. Actual outturn costs are dened as
real, accounted construction costs determined at the time of
project completion. Estimated costs are dened as budgeted, or
forecasted, construction costs at the time of decision to build. The
year of the date of the decision to build a project is the base year
of prices in which all estimated and actual constant costs have
been expressed in real (i.e. with the effects of ination removed)
local currency terms of the country in which the project is
located. We exclude from our calculations debt payments, any
ex post environmental remedial works, and opportunity cost of
submerging land to form reservoirs. This makes comparison of
estimated and actual costs of a specic project a like-for-like
comparison.
2.3. Analyses
We investigated the magnitude and frequency of cost and
schedule forecast (in)accuracies with a combination of simple
Fig. 1. Sample distribution of 245 large dams (19342007), across ve continents, worth USD 353B (2010 prices).
3
Hufschmidt and Gerin (1970) report data on over 100 dams built in the
United States between 1933 and 1967. The salient results of the study were that in
nominal USD terms dams built by TVA suffered a 22% cost overrun; U.S. Corps of
Engineers overrun was 124% for projects built or building prior to 1951, and 36% for
projects completed between 1951 and 1964; while U.S. Bureau of Reclamation
overrun was 177 per cent for projects built or building prior to 1955 and 72 per cent
for all projects built or building in 1960 (Hufschmidt and Gerin, 1970: 277). Despite
its large sample, Hufschmidt and Gerin (1970) do not report data broken down
project-by-project. The validity and reliability of these data could not thus be
established and were consequently excluded.
A. Ansar et al. / Energy Policy 69 (2014) 435646
statistical (parametric and non-parametric) tests and by tting
more sophisticated multilevel regression models sometimes
termed Hierarchical Linear Models (HLM).
Multilevel or hierarchically structured data are the norm in the
social, medical, or biological sciences. Rasbash et al. (2009, p. 1)
explain: For example, school education provides a clear case of a
system in which individuals are subject to the inuences of
grouping. Pupils or students learn in classes; classes are taught
within schools; and schools may be administered within local
authorities or school boards. The units in such a system lie at four
different levels of a hierarchy. A typical multilevel model of this
system would assign pupils to level 1, classes to level 2, schools to
level 3 and authorities or boards to level 4. Units at one level are
recognized as being grouped, or nested, within units at the next
higher level. Such a hierarchy is often described in terms of
clusters of level 1 units within each level 2 unit, etc. and the term
clustered population is used.Important for a hierarchical linear
model is that the dependent variable is at the lowest level of the
nested structure. Multilevel models are necessary for research
designs where data for observations are organized at more than
one level (i.e., nested data) (Gelman and Hill, 2007). Failing to use
multilevel models in such instances would result in spurious
results (Rasbash et al., 2009).
With respect to our data on dams, projects are nested in the
countries of their domicile. Like test scores of pupils from the
same school tend exhibit within-school correlation, similarly
outcomes of dam projects may exhibit within-country correla-
tion that needs to be properly modeled using a multilevel
model. We took this into account by modeling country as a rst
level random effect in a mixed effects multilevel model. The
models were made parsimonious by using stepwise variable
selection.
3. Results and interpretation
Our second step was to establish an empirical distribution for
the cost forecast errors of large dams. We collected data on 36
possible explanatory variables, listed in Table 1, for the 245 large
dams in our reference class.
Table 1
Variables and characteristics used in multilevel regressions on construction cost overrun and schedule slippage.
Project-specic variables
Project features
Hydropower or non-hydropower large dam project (dummy variable)
New power station or station extension (dummy variable)
Size
Generator unit capacity (MW)
Total project generation capacity (MW)
Dam height for new hydropower station (meters)
Hydraulic head for new hydropower station (meters)
a
Reservoir area created by project (hectares)
a
Length of tunnels (kilometers)
a
Cost
Estimated project cost (constant local currency converted to 2010 USD MM)
Actual project cost (constant local currency converted to 2010 USD MM)
Cumulative ination contingency (percentage)
Time
Year of nal decision to build
Estimated implementation schedule (months)
Year of start of full commercial operation
Actual implementation schedule (months)
Procurement
Estimated project foreign exchange costs as a proportion of estimated total project costs (percentage)
Competitiveness of procurement process, international competitive bidding amount as a proportion of estimated total project costs (percentage)
*
Main contractor is from the host country (dummy variable)
Country variables
Country (second level to control for within country correlation)
Political regime of host country is a democracy (dummy variable)
GDP of host country (current USD)
Per capita income of host country in year of loan approval (constant USD)
Average actual cost growth rate in host country over the implementation periodthe GDP deator (percentage)
MUV Index of actual average cost growth rate for imported project components between year of loan approval and year of project completion
Long-term ination rate of the host country (percentage)
Actual average exchange rate depreciation or appreciation between year of formal-decision-to-build and year of full commercial operation (percentage)
South Asian projects (dummy variable)
North American projects (dummy variable)
a
Denotes variables with a large number of missing values not used for regression analysis.
Fig. 2. Density trace of actual/estimated cost (i.e. costs overruns) in constant local
currency terms with the median and mean (N¼245).
A. Ansar et al. / Energy Policy 69 (2014) 4356 47
3.1. Preliminary statistical analysis of cost performance
With respect to cost overruns, we make the following observa-
tions:
1. Three out of every four large dams suffered a cost overrun in
constant local currency terms.
2. Actual costs were on average 96% higher than estimated costs;
the median was 27% (IQR 86%). The evidence is overwhelming
that costs are systematically biased towards underestimation
(MannWhitneyWilcoxon U¼29,646, po0.01); the magni-
tude of cost underestimation (i.e. cost overrun) is larger than
the error of cost overestimation (po0.01). The skew is towards
adverse outcomes (i.e. going over budget).
3. Graphing the dams' cost overruns reveals a fat tail as shown in
Fig. 2; the actual costs more than double for 2 out of every 10
large dams and more than triple for 1 out of every 10 dams. The
fat tail suggests that planners have difculty in computing
probabilities of events that happen far into the future (Taleb,
[2007] 2010, p. 284).
4. Large dams built in every region of the world suffer systematic
cost overruns. The mean forecasting error is signicantly above
zero for every region. Fig. 3 shows the geographical spread and
cost overruns of large dams in our reference class. Large dams
built in North America (N¼40) have considerably lower cost
overrun (M¼11%) than large dams built elsewhere (M¼104%).
Although after controlling for other covariates such as project
scale in a multilevel model, reported below, the differences
among regions are not signicant. We noted, three out of four
dams in our reference class had a North American rm advising
on the engineering and economic forecasts. Consistent with
anchoring theories in psychology, we conjecture that an over-
reliance on the North American experience with large dams may
bias cost estimates downwards in rest of the world. Experts may
be anchoringtheir forecasts in familiar cases from North
America and applying insufcient adjustments(Flyvbjerg
et al., 2009;Tversky and Kahneman, 1974), for example to
adequately reect the risk of a local currency depreciation or
the quality of local project management teams. Instead of
optimistically hoping to replicate the North American cost
performance, policymakers elsewhere ought to consider the
global distributional information about costs of large dams.
5. The typical forecasted benet-to-cost ratio was 1.4. In other
words, planners expected the net present benets to exceed
the net present costs by about 40%. Nearly half the dams
suffered a cost overrun ratio of 1.4 or greater breaching this
threshold after which the asset can be considered strandedi.e.
its upfront sunk costs are unlikely to be recovered. This is
assuming, of course, that the benets did not also fall short of
targets, even though there is strong evidence that actual
benets of dams are also likely to fall short of targets (WCD,
2000;McCully, 2001;Scudder, 2005).
4
6. We tested whether forecasting errors differ by project type
(e.g., hydropower, irrigation, or multipurpose dam) or wall type
(earthll, rockll, concrete arch, etc.). Pairwise comparisons of
percentage mean cost overrun and standard deviations as well
as non-parametric MannWhitney tests for each of the para-
meters show no statistically signicant differences. We con-
clude that irrespective of project or wall type, the probability
distribution from our broader reference class of 245 dams
applies as in Fig. 2.
7. We analyzed whether cost estimates have become more
accurate over time. Statistical analysis suggests that irrespec-
tive of the year or decade in which a dam is built there are no
signicant differences in forecasting errors (F¼0.57, p¼0.78).
Similarly, there is no linear trend indicating improvement or
deterioration of forecasting errors (F¼0.54, p¼0.46) as also
suggested in Fig. 4. There is little learning from past mistakes.
By the same token, forecasts of costs of large dams today are
likely to be as wrong as they were between 1934 and 2007.
We also explored the absolute costs of large hydropower dams
(N¼186). A large hydropower dam on average costs 1800 million
in 2010 USD with an average installed capacity of 630 MW. One
MW installed capacity on average costs 2.8 million in 2010 USD.
Fig. 3. Location of large dams in the sample and cost overruns by geography.
4
A more comprehensive inquiry into planned versus actual benets of dams is
postponed until a future occasion but data available on 84 of the 186 large
hydroelectric dam projects thus far suggests that they suffer a mean benets
shortfall of 11%.
A. Ansar et al. / Energy Policy 69 (2014) 435648
A preliminary univariate analysis, which makes no attempts to
take into account any covariates, shows that increase in the scale
of a dam, e.g., measured as height of the dam wall, increases the
absolute investment required exponentially, e.g. a 100 m high dam
wall is four times more costly than a 50 m wall (R
2
¼0.27, F¼92.5,
po0.01). An even stronger relationship can be seen between
installed capacity MW and actual costs (R
2
¼0.70, F¼461.1,
po0.01).
Furthermore, the rate of cost overrun outliers increases with
increase in dam size either measured in installed hydropower
generation (r¼0.24, p¼0.01) or wall height (r¼0.13, p¼0.05).
Since there is a signicant correlation between dam height and
hydropower installed capacity (r¼0.47, po0.01), evidence sug-
gests that larger scale in general is prone to outlying cost overruns.
We further investigate the effects of scale on cost overruns by
tting multilevel models (Models 1 and 2) reported below.
3.2. Preliminary statistical analysis of schedule performance
Not only are large dams costly and prone to systematic and
severe budget overruns, they also take a long time to build. Large
dams on average take 8.6 years. With respect to schedule slippage,
we make the following observations:
8. Eight out of every 10 large dams suffered a schedule overrun.
9. Actual implementation schedule was on average 44%
(or 2.3 years) higher than the estimate with a median of 27%
(or 1.7 years) as shown in Fig. 5. Like cost overruns, the
evidence is overwhelming that implementation schedules
are systematically biased towards underestimation (Mann
WhitneyWilcoxon U¼29,161, po0.01); the magnitude of
schedule underestimation (i.e. schedule slippage) is larger
than the error of schedule overestimation (po0.01).
10. Graphing the dams' schedule overruns also reveals a fat tail as
shown in Fig. 5, albeit not as fat as the tail of cost overruns.
Costs are at a higher risk of spiraling out of control than
schedules.
11. There is less variation in schedule overruns across regions than
cost overruns. Large dams built everywhere take signicantly
longer than planners forecast. North America with a 27% mean
schedule overrun is the best performer. A non-parametric
comparison using a Wilcoxon test (p¼0.01) suggests that
projects in South Asia have signicantly greater schedule
overruns (M¼83%) than rest of the world taken as a whole
(M¼42%). We investigate this further with a multilevel model
below (Model 3).
12. There is no evidence for schedule estimates to have improved
over time.
We tested whether implementation schedules and project
scale are related. A preliminary univariate analysis, which makes
no attempts to take into account any covariates, shows that
increase in the scale of a dam, e.g., measured as estimated cost
of construction, increases the absolute actual implementation
schedule required exponentially (R
2
¼0.13, F¼36.4, po0.01). Large
scale is intimately linked with the long-term (see Model 2 below).
The actual implementation schedule, reported here, does not take
into the account lengthy lead times in preparing the projects.
Dams require extensive technical and economic feasibility analy-
sis, social and environmental impact studies, and political negotia-
tions. The actual implementation cycles are far longer than the
average of about 8.6 years, as shown in our data, that it takes to
build a dam. These lengthy implementation schedules suggest that
the benets of large dams (even assuming that large dam generate
benets as forecasted) do not come onlinequickly enough. The
temporal mismatch between when users need specic benets
and when these benets come online is not to be downplayed
(Ansar et al., 2012). Alternative investments that can bridge needs
quickly, without tremendous time lags, are preferable to invest-
ments with a long lead-time and hence duration risk (Luehrman,
1998;Copeland and Tufano, 2004).
3.3. Multilevel regression analysis of cost and schedule performance
Means, standard deviations, and correlations of the variables
used in the multilevel regressions are shown in Table 2.
We tted multilevel regression models with projects nested by
country as a second level to incorporate within-country correla-
tion. The models were tted using the lmeprocedure in the
nlmepackage in R software. This function ts a linear mixed-
effects model in the formulation described in Laird and Ware
(1982) but allowing for nested random effects. The within-group
errors are allowed to be correlated and/or have unequal variances.
We found it necessary to transform variables to remove excessive
skewness as noted in Table 2. Using stepwise variable selection,
we are not only able to t explanatory models for cost and
overruns and estimated duration but also practicably parsimo-
nious models for predicting them.
Table 3 summarizes the results from multilevel model examin-
ing predictors of cost overruns (Model 1). Model 1 identies
Fig. 4. Inaccuracy of cost estimates (local currencies, constant prices) for large
dams over time (N¼245), 19342007.
Fig. 5. Density trace of schedule slippage (N¼239) with the median and mean.
A. Ansar et al. / Energy Policy 69 (2014) 4356 49
the estimated implementation schedule and the long-term ina-
tion rate in the country in which the project is built as highly
signicant variables. An increase in estimated duration of one
year contributes to an increase in cost overrun of approx. 56
percentage points depending on the country whilst holding the
ination rate constant (see Fig. A1). Note that an R-squared
measure, which is customary to report for single-level regressions
as explained proportion of variance, cannot be applied to
multilevel models (Recchia, 2010).
5
The usual diagnostics, based
upon the model residuals, were satisfactory.
The rst nding in Model 1 is that the larger the estimated
implementation schedule the higher the cost overrun (p¼0.016),
with all other things being equal, is particularly noteworthy for
two reasons.
First, Model 1 suggests that planners' forecasting skills decay
the longer in the future they are asked to project the risks facing a
large dam. Material information about risks, for example, related
to geology, prices of imports, exchange rates, wages, interest rates,
sovereign debt, environment, only reveal in future shaping episode
to which decision-makers are blindex ante (Flyvbjerg and
Budzier, 2011). We discuss some qualitative case examples to
illustrate this statistical result and its broader implications in the
next section.
Second, preliminary analysis had suggested that estimated
implementation schedules depend on the scale of a planned
Table 2
Descriptive statistics and correlations (N¼245).
Variable Mean S.D. 1 2 3 4 5 6 7 8 9
1. Cost Overrun
a
2.0 3.6
2. Schedule slippage
a
1.5 0.7 0.17
nn
3. Estimated schedule (months)
b
73.1 33.8 0.16
n
0.23
nn
4. Actual schedule (months)
b
102.7 55.7 0.27
nn
0.43
nn
0.76
nn
5. Yeardecision to build 1971.1 13.2 0.02 0.05 0.21
nn
0.25
nn
6. Yearcompletion 1979.6 12.7 0.14
n
0.10 0.03 0.08 0.94
nn
7. Project type dummy 0.8 0.4 0.14
n
0.08 0.10 0.02 0.02 0.02
8. Democracy dummy 0.4 0.5 0.00 0.14
n
0.16
n
0.20
nn
0.45
nn
0.38
nn
0.00
9. Estimated cost (USD MM 2010 constant)
b
699.6 1215.5 0.03 0.09 0.4 8
nn
0.37
nn
0.02 0.13
n
0.37
nn
0.04
10. Actual cost (USD MM 2010 constant)
b
1462.2 4032.5 0.38
nn
0.02 0.50
nn
0.43
nn
0.02 0.17
nn
0.38
nn
0.03 0.93
nn
11. Height of dam wall (m)
c
77.3 51.6 0.10 0.10 0.26
nn
0.17
nn
0.10 0.16
n
0.34
nn
0.03 0.51
nn
12. Installed hydropower capacity (MW)
b
487.0 1255.3 0.16
n
0.19
nn
0.22
nn
0.08 0.13
n
0.16
n
0.69
nn
0.14
n
0.59
nn
13. Length of dam wall (m)
b
1364.1 2061.9 0.12 0.07 0.25
nn
0.30
nn
0.19
nn
0.08 0.07 0.08 0.37
nn
14. Tunnel length (m)
b
3500.0 7869.5 0.13 0.12 0.04 0.16 0.06 0.01 0.23 0.05 0.11
15. Manufactures unit value index CAGR
d
6.0 5.4 0.01 0.03 0.25
nn
0.18
nn
0.12 0.18
nn
0.08 0.08 0.13
16. GDP (nominal USD B)
b
1221.1 253.4 0.05 0.25
nn
0.36
nn
0.17
n
0.29
nn
0.37
nn
0.13 0.13 0.19
n
17. Per capita income (2000 constant USD)
b
4132.8 5198.6 0.23
nn
0.15
n
0.11 0.01 0.37
nn
0.40
nn
0.07 0.48
nn
0.07
18. Long-term ination (%)
b
17% 0.2 0.29
nn
0.04 0.09 0.11 0.22
nn
0.19
nn
0.24
nn
0.37
nn
0.13
n
19. Forex depreciation (%)
e
18% 70.3 0.30
nn
0.04 0.03 0.00 0.29
nn
0.29
nn
0.16
n
0.20
nn
0.21
nn
20. South Asia dummy 0.1 0.3 0.25
nn
0.18
nn
0.17
nn
0.26
nn
0.04 0.07 0.06 0.20
nn
0.11
21. North America dummy 0.2 0.4 0.28
nn
0.06 0.21
nn
0.13
n
0.57
nn
0.55
nn
0.09 0.52
nn
0.06
Variable 10 11 12 13 14 15 16 17 18 19 20
11. Height of dam wall (m) 0.51
nn
12. Installed hydropower capacity (MW) 0.60
nn
0.47
nn
13. Length of dam wall (m) 0.38
nn
0.03 0.13
14. Tunnel length (m) 0.01 0.05 0.22 0.18
15. Manufactures Unit Value Index CAGR 0.12 0.08 0.02 0.02 0.02
16. GDP (nominal USD) 0.19
n
0.10 0.09 0.04 0.29 0.31
nn
17. Per capita income (2000 constant USD) 0.14
n
0.08 0.11 0.02 0.09 0.01 0.29
nn
18. Long-term ination (%) 0.22
nn
0.06 0.33
nn
0.07 0.41
n
0.15
n
0.03 0.24
nn
19. Forex 0.29
nn
0.09 0.29
nn
0.02 0.37
n
0.16
n
0.00 0.26
nn
0.64
nn
20. South Asia dummy 0.19
nn
0.08 0.03 0.20
nn
NA 0.09 0.01 0.46
nn
0.10 0.11
21. North America dummy 0.03 0.10 0.16
n
0.19
nn
NA 0.18
nn
0.33
nn
0.60
nn
0.44
nn
0.31
nn
0.15
n
a
One over (1/x) transformed.
b
Log transformed.
c
Sq. rt. (x).
d
Cb rt. (x).
e
x
0.25
transformed to remove excess skewness for regression analysis and to calculate correlations.
nn
po0.01.
n
po0.05.
Table 3
Model 1Signicant variables for cost accuracy for large dam projects (constant
local currency).
Variable Regression
coefcient
Standard
error
t-Stat 2-Tailed
signicance
Intercept 1.402 0.185 7.560 0.000
Log estimated duration
(months)
0.100 0.041 2.424 0.016
Log of country's long-term
ination rate (%)
0.085 0.029 2.930 0.005
Note: Dependent variable is cost forecast accuracy, which is the estimated/actual
cost ratio (i.e. 1/xof the cost overrun to remove excessive skewness), based on 239
observations. Since the dependent variable in Model 1 is the inverse of the cost
overrun a negative sign on the coefcients of both signicant variables suggests
that an increase in the estimated duration or long-term ination rate increases the
cost overrun.
5
Recchia (2010, p. 2) explains further why a R-squared measure cannot be
used for a multilevel model. A single-level model includes an underlying
assumption of residuals that are independent and identically distributed. Such an
assumption could easily be inappropriate in the two[or multi]-level case since
there is likely to be dependence among the individuals that belong to a given
group. For instance, it would be difcult to imagine that the academic achieve-
ments of students in the same class were not somehow related to one another.
Also see Kreft and Leeuw (1998) and Goldstein (2010).
A. Ansar et al. / Energy Policy 69 (2014) 435650
investmenti.e. bigger projects take longer to build. Support of
this preliminary result was found by tting a multilevel model
(Model 2) that examines the predictors of estimated implementa-
tion schedule. Model 2 shows that height (p¼0.02), installed
capacity (MW) (p¼0.02), and length (p¼0.04) of the dam wall
are signicant variables associated with the estimated implemen-
tation schedule. The effect of these covariates can be seen from the
coefcients in Table 4: a greater height, installed capacity, or
length contribute to longer implementation schedules. We inter-
pret Model 2 as follows. Estimated implementation schedule acts
not only as a temporal variable but also as a surrogate for scalar
variables such as wall height (which is also highly correlated with
installed capacity). The larger the dam, the longer the estimated
implementation schedule, and the higher the cost overrun.
Taken together, the multilevel models for cost overruns and
estimated schedule suggest that longer time horizons and increas-
ing scale are underlying causes of risk in investments in large
hydropower dam projects.
The second nding in Model 1 is that the higher the long-term
ination rate of the host country the higher the cost overrun
suffered by a dam (p¼0.02). The long-term ination rate was
calculated by tting a linear model to the log of the time series of
the GDP deator index of each country. The slope of this tted line
can be interpreted as the annual average growth rate of the log
ination for each country. This slope is a different constant for
each country with some countries such as Brazil with a consider-
ably higher long-term ination rate, and hence greater propensity
to cost overruns, than China or the United States. Moreover, this
slope is stable in the short-run (it takes years of high or low
ination to change this slope) and hence our estimate can be
assumed to be a reliable predictor. Recall that the cost overrun is
being measured in constant terms (i.e. with the effects of ination
removed); yet Model 1 suggests that the ination trajectory of a
country, which we interpret as a surrogate of the overall macro-
economic management, is an important risk when making durable
investments. The multilevel model nally suggests that once
country specic factors have been taken into account the factor
that drives cost overrun is the planning horizon.
Finally, we t a multilevel model (Model 3) to examine predictors
of schedule overruns. Model 3 identies the following signicant
variables:whetherornotacountryisademocracy;thepercapita
income of the country in 2000 constant USD in the year of the
decision to build; the planned installed capacity (MW); and planned
length of the dam wall (meters). Avid dam building countries in
South Asia, at various stages of democratic maturity, have also one of
the poorest schedule performances in building dams. We controlled
for this fact by including a dummy variable for South Asia in the
model as a covariate with an interaction effect with the democracy
dummy. Democracy in South Asia is signicant in explaining
schedule overruns. The South Asia dummy, however, does not come
out to be signicant. The effect of these covariates and the interaction
effect can be seen in Table 5.
First, democracies' forecasts about implementation schedules of
large dams are systematically more optimistic than autocracies even
after controlling for systematically higher schedule overruns in India
andPakistan.Thesizeofthecoefcient is large suggesting that
political process has profound impact on the schedule slippage. We
tested whether democracies take longer than autocracies to build
large dams by tting a model to explain the actual implementation
schedule (Model 4). Model 4, summarized in Table 6,showsthat
effects of political regime on the actual schedule are not signicant.
In other words, while democracies do not take longer to build large
dams than autocracies in absolute terms, democracies appear to be
more optimistic. Given its vast scope, we defer a further investigation
of this important result to a future inquiry. We note, however, that
theories of delusion and deception in the planning of large infra-
structure projects (Flyvbjerg et al., 2009) would interpret this as
evidence of ex ante political intent among democratically elected
politicians to present a rosier picture about large dams than they
know the case to be.
Second, countries with a higher per capita income in constant
2000 USD in the year of decision to build tend to have lower
Table 4
Model 2Signicant variables for estimated construction schedule for large dam
projects (months).
Variable Regression
coefcient
Standard
error
t-Stat 2-Tailed
signicance
Intercept 3.444 0.197 17.464 0.000
Sq rt of dam wall
height (m)
0.029 0.012 2.414 0.017
Log of dam wall length (m) 0.058 0.027 2.153 0.033
Log of hydropower
installed capacity (MW)
0.016 0.007 2.141 0.034
Note: Dependent variable is log of the estimated construction schedule, based on
239 observations.
Table 5
Model 3Signicant variables for schedule slippage for large dam projects.
Variable Regression coefcient Standard error t-Stat 2-Tailed signicance
Intercept 0.405 0.163 2.483 0.014
Democracy dummy
a
0.134 0.055 2.439 0.016
Log of country's per capita income in year of decision to build (constant USD) 0.065 0.019 3.334 0.001
Log of dam wall length (m) 0.027 0.013 2.081 0.039
Log of hydropower installed capacity (MW) 0.018 0.006 3.207 0.002
South Asia dummy 0.211 0.113 1.874 0.066
Democracy in South Asia interaction effect 0.239 0.113 2.114 0.036
Note: Dependent variable is 1/xof the actual/estimated schedule ratio, based on 239 observations.
a
Dummy based on the Polity2 variably of Polity IV regime index. Score of þ10 to þ6¼democracy; score of þ5to10 ¼autocracy.
Table 6
Model 4Signicant variables for estimated construction schedule for large dam
projects (months).
Variable Regression
coefcient
Standard
error
t-Stat 2-Tailed
signicance
Intercept 17.712 6.401 2.767 0.007
Log of dam wall length
(m)
0.105 0.029 3.567 0.001
Year of actual project
completion
0.011 0.0 03 3.358 0.001
Note: Dependent variable is log of the actual construction schedule, based on 239
observations.
A. Ansar et al. / Energy Policy 69 (2014) 4356 51
schedule overruns than countries with lower per capita income.
We concur with the interpretation of Bacon and Besant-Jones
(1998, p. 325) that the best available proxy for most countries
is [the] country-per-capita income[for] the general level of
economic support that a country can provide for the construction
of complex facilities. This result suggests that developing coun-
tries in particular, despite seemingly the most in need of complex
facilities such as large dams, ought to stay away from bites bigger
than they can chew.
Third, the evidence appears to be contradictory with respect to
scale. While a greater dam wall length contributes to a higher
schedule overrun, a higher MW installed capacity has the opposite
effect. Model 3 in Table 5 shows that the size of coefcients for the
two signicant variables related to physical scalei.e. Log of dam
wall length (m) and Log of hydropower installed capacity (MW)
is approximately the same but with the opposite sign.
6
In attempting to interpret this result our conjecture is as
follows. Dam walls are bespoke constructions tied to the geolo-
gical and other site-specic characteristics. In contrast, installed
capacity is manufactured off-site in a modular fashion. For
example, the 690 MW installed capacity of the recently com-
pleted Kárahnjúkar project in Iceland was delivered with six
generating units of identical design (6 115 MW). We p r o p o s e
that project components that require onsite construction, e.g.
dam wall, are more prone to schedule errors than components
manufactured off-site, e.g. generation turbines. Project designs
that seek to reduce the bespoke and onsite components in favor
of greater modular and manufactured components may reduce
schedule uncertainty.
This conjecture is supported by Model 4 in Table 6 ,which
shows that the actual construction schedule, in absolute terms,
is signicantly increased with an increase in the length of dam
wall. In contrast, MW installed capacity does not have an effect
on the absolute actual construction schedule suggesting that
construction schedules are more sensitive to on-site construc-
tion than to components manufactured in factories. Note that
lower installed capacity does not necessarily equate with a
smaller dam. For example, it is not rare for a large multipurpose
dam to have a low MW installed capacity when, for instance, the
dam is primarily being used for irrigation or ood management
purposes.
4. Qualitative case examples and policy propositions
The statistical results reported in the preceding sections show
that cost and schedule estimates of large dams are severely and
systematically biased below their actual values. While it is beyond
the scope of this paper to discuss wider theoretical implications,
the evidence presented here is consistent with previous ndings
that point to twin problems that cause adverse outcomes in the
planning and construction of large and complex facilities such as
large hydropower dams: (1) biases inherent in human judgment
(delusion) and (2) misaligned principal-agent relationships or
political incentives (deception) that underlie systematic forecast-
ing errors. In the context of large dams, we argue that large scale
and longer planning time horizons exacerbate the impact of these
twin problems. We now present a few qualitative examples of
risks large dams typically face to illustrate the statistical results
reported above. We jointly draw on the statistical analyses and
qualitative analyses to distill propositions of immediate relevance
to policy.
Globally, experts' optimism about several risk factors con-
tribute to cost overruns in large dams. For example, the planning
documents for the Itumbiara hydroelectric project in Brazil
recognized that the site chosen for the project was geologically
unfavorable. The plan optimistically declared, the cost estimates
provide ample physical contingencies [20% of base cost] to
provide for the removal of larger amounts [of compressible,
weak, rock] if further investigations show the need(World
Bank, 1973). This weak geology ended up costing þ96% of the
base cost in real terms. Itumbiara's case is illustrative of a
broader problem. Even though geological risks are anticipatable
there is little planners can do to hedge against it. For example,
exhaustive geological investigation for a large dam can cost as
much as a third of the total cost (Hoek and Palmieri, 1998); at
which point still remains a considerable chance of encountering
unfavorable conditions that go undetected during the ex ante
tests (Goel et al., 2012).
Policy proposition 1. Energy alternatives that rely on fewer site-
specic characteristics such as unfavorable geology are preferable.
Similarly, in the Chivor hydroelectric project in Colombia, the
planning document was upbeat that there will be no changes in
the exchange rate between the Colombian Peso and the U.S.
dollar during the construction period (19701977) stating, No
allowance has been made for possible future uctuations of the
exchange rate. This approach is justied by recent experience in
Colombia where the Government has been pursuing the enligh-
tened policy of adjusting [policy] quickly to changing conditions
in the economy(World Bank, 1970). In fact, the Colombian
currency depreciated nearly 90% against the U.S. dollar as shown
in Fig. 6.
Since over half the project's costs covers imported inputs, this
currency depreciation caused a 32% cost overrun in real Colombian
Peso terms. Currency exposure arises when the inputs required to
build a project are denominated in one currency but the outputs in
another, or vice versa. The outputs of dams, such as electricity, are
denominated in the local currency. Similarly, any increases in tax
receipts a dam may enable for the host government also accrue
in local currency. A large portion of inputs to build a dam,
particularly in developing countries, however, constitute imports
paid for in USD. Since the USD liabilities also have to eventually be
Fig. 6. Depreciation of the Colombian Peso 19702010.
6
Note that the dependent variable in Model 3 is forcast accuracy, the inverse of
schedule overrun (i.e. 1/xof the schedule overrun or Estimated/Actual schedule).
Thus a negative sign on the Log of dam wall length (m) suggests that an increase in
wall length decreases the inverse of the schedule overrun. In other words, increase
in wall length increases schedule overrun.
A. Ansar et al. / Energy Policy 69 (2014) 435652
paid in local currency, currency exposure consistently proves to be
ascal hemorrhage for large projects.
Policy proposition 2. Energy alternatives that rely on fewer imports
or match the currency of liabilities with the currency of future
revenue are preferable.
Although, following convention, our cost analysis excludes
the effects of ination, planners ought not to ignore the risks
of unanticipated ination(Pickrell, 1992, p. 164). Episodes of
hyperination in Argentina, Brazil, Turkey, and Yugoslavia caused
staggering nominal cost overruns, e.g. 7-times initial budget for
Brazil's Estreito dam (19651974), or 110-times initial budget for
Yugoslavia's Visegrad dam (19851990), which had to be
nanced with additional debt. Effects of unanticipated ination
magnify the longer it takes to complete a project. For example,
during the planning phase of Pakistan's Tarbela dam, it was
assumed that ination would not have a signication impact on
the project's costs. The appraisal report wrote: A general con-
tingency of 7½% has been added in accordance with normal
practice for works of this size and duration(World Bank, 1968).
The project, launched in 1968, was meant to start full commercial
operation in 1976, but the opening was delayed until 1984. Actual
cumulative inationinPakistanduring19681984 was 380%; the
actual cost of the dam in nominal terms nearly four times the
initial budget. In the case of Tarbela, unanticipated ination was
a product of delays in a project's construction timetable and a
higher-than expected ination rate(Pickrell, 1992, p. 164). For
our reference class, 8 out of 10 large dams came in late with an
average delay of 2.3 years. Moreover, forecasters expected the
annual ination rate to be 2.5% but it turned out to be 18.9%
(averages for the entire sample). Large dams have a high
propensity to face unanticipated ination.
Policy proposition 3. The best insurance against creeping ina-
tion is to reduce the implementation schedule to as short a horizon
as possible. Energy alternatives that can be built sooner and with
lower risk of schedule overruns, e.g. through modular design, are
preferable.
Large dams are typically nanced from public borrowing.
While our calculations exclude debt-servicing, cost overruns
increase the stock of debt but also the recurring nancing costs
that can further escalate if interest rates go up. The optimistic risk
assessments of the costs of large dams are consistent with
explosive growth of Third World debt(Bulow and Rogoff,
199 0;Mold, 2012). For example, the actual cost of Tarbela dam,
the majority of which was borrowed from external sources,
amounted to 23% of the increase in Pakistan's external public
debt stock between 1968 and 1984; or 12% for Colombia's Chivor
dam (19701977) as shown in Table 7.
These case examples reinforce the essential message of our
statistical results: bigger projects entail uncontrollable risks,
which even when anticipatable cannot be adequately hedged.
We do not directly negate the presence of economies of scale or
learning curvesi.e. declining average cost per unit as output
increases. Instead our argument is that any economies of scale
embedded in large scale are being acquired for a disproportio-
nately increased exposure to risk that can cause nancial impair-
ment. Companies and countries with insufcient capacity to
absorb adverse outcomes of big bets gone awry often face
nancial ruin.
Policy proposition 4. Energy alternatives that do not constitute a
large proportion of the balance sheet of a country or a company are
preferable. Similarly, policymakers, particularly in countries at lower
levels of economic development, ought to avoid highly leveraged
investments denominated in a mix of currencies.
5. Forecasting the actual costs and schedules using reference
class forecasting (RCF)
As discussed in the methods section, the third step of the
outside viewor RCF techniques is to compare a specic
venture with the reference class distribution, in order to estab-
lish the most likely outcome for the specicventure.Thusif
systematic errors in the forecasts generated using the inside
viewof previous ventures are found, decision-makers should
apply an uplift or downlift to the inside viewforecast in order
to generated a de-biased outside viewforecast. For example,
empirical literature has established that rail projects suffer a
cost overrun of 45% on average (Flyvbjerg, 2008;alsosee
Table 8). The 50th percentile cost overrun for rail projects is
40% and the 80th percentile is 57%. Based on these ndings, RCF
techniques suggest that decision-makers ought to apply a 57%
uplift to the initial estimated budget in order to obtain 80%
certainty that the nal cost of the project would stay within
budget (Flyvbjerg, 2008,p.16).Ifdecision-makersweremore
risk tolerant then they could apply a 40% uplift to the initial
estimated budget but then there will remain a 50% chance that
the proposed project might exceed its budget.
In line with the RCF techniques, the third and nal step of our
investigation on dams was to derive a good predictor of cost and
schedule overruns for proposed large dams based on the distribu-
tional information of the reference class. This predictor serves to
correctthe systematically biased ex ante cost and schedule
estimates by adjusting them upwards by the average cost or
schedule overrun (see Kahneman and Tversky, 1979b;Flyvbjerg,
2006,2008).
First, using traditional RCF (Flyvbjerg, 2006,2008), we traced
the empirical distribution of cost and schedule overruns of large
dams. Second, we use multilevel Models 1 and 3, described
above, for predicting cost and schedule overruns. Models 1 and
3 prove to be practicably parsimonious models for two reasons:
First both models are tted with variables known ex ante.
Second, both models were successfully tted with only a few
signicant variables making it practicable to collect the data
needed to make a prediction. For example, Model 1 on cost
overruns has only two signicant variablesestimate schedule
and the long-term ination rate of the host country. Data on
both these variables is readily available for any proposed large
dam making it possible to predict the cost overrun before
construction begins. We illustrate the usefulness of our predic-
tive models with an example below.
With respect to cost overruns, using traditional RCF (Flyvbjerg,
2006,2008), we nd that if planners are willing to accept a 20%
risk of a cost overrun, the uplift required for large dams is þ99%
(i.e. double experts' estimates) as seen in Fig. 7; and þ176%
Table 7
Total stock of public net external debt (USD current, MM).
Year Colombia Pakistan
1968 3252.4
1970 1296.6
1977 2699.6
1984 9692.8
Debt increase over the implementation schedule 1403.0 6440.5
Cost of mega-dam over the relevant period
(USD current MM)
Chivor dam Tarbela dam
168.7 1497.90
Cost of dam as percentage of debt increase 12.0% 23.2%
A. Ansar et al. / Energy Policy 69 (2014) 4356 53
including unanticipated ination. If planners are willing to accept
a5050 chance of a cost overrun, the uplift required is 26%
(32% outside North America).
In terms of cost overruns, Fig. 7 also illustrates that large dams
are one of the riskiest asset classes for which valid and reliable
data are available. Compare, for example, Fig. 7 with reference
class forecasts previously conducted for rail, road, tunnel, or bridge
projects (Flyvbjerg, 2006,2008) also summarized in Table 8.
Second, using our multilevel Model 1 we were able to derive
predictions for cost overrun (in constant local currency) and
schedule overrun respectively.
Experts estimate, for instance, that Pakistan's Diamer-Bhasha
dam, whose construction began shortly after the 2010 oods, will
cost PKR 894 billion (USD12.7B in 2008 prices and exchange
rates and about 9% of Pakistan's 2008 GDP) (WAPDA, 2011). The
dam is forecasted to take 10 years from 2011 and become
operational in 2021. Using our rst approach, the reference class
forecast for cost overruns suggests that planners need to budget
PKR 1,788B (USD25.4B) in real terms to obtain 80% certainty of not
exceeding the revised budget. Including the effects of unantici-
pated ination the required budget is PKR 2,467B (USD35.0B) or
about 25% of Pakistan's 2008 GDP. A future sovereign default in
Pakistan owing to this one mega-dam is not a remote possibility.
Using our second approach, our multilevel Model 1 predicts
that given the 10 year estimated duration and a long-term
ination rate of about 8% the expected (average) cost overrun of
a large dam in Pakistan will be 44% (PKR 1,288B or USD 18.3B).
Combining the two methods, a conservative estimate for the cost
overrun on the Diamer-Bhasha dam is 44% at which point there
remains a 4 in 10 chance of the revised budget being exceeded.
Note, however, that if a dam of dimensions similar to Diamer-
Bhasha were being built in the US, Model 1 predicts that it would
only suffer a cost overrun of 16%, which the much larger US
economy could absorb without any lasting damage.
We applied a similar two-pronged forecast of schedule slip-
page. Using our rst approach, the reference class forecast for
schedule slippage suggests that planners for large dams around
the world need to allow for a 66% schedule overrun to achieve 80%
certainty that the project will be completed within the revised
implementation schedule. Since Diamer-Bhasha is expected to
take 10 years to build (20112021), planners need to adjust their
schedule estimate upwards to nearly 17 years (i.e. an actual
opening date of 2028). Using our second approach, our multilevel
Model 3 predicts that given that the dam's nal decision to build
was made in Pakistan by a democratically elected government,
when the per capita income was USD 497 in 2000 constant dollars,
a dam wall length of 998 m, and an installed capacity of 4500 MW,
the expected outcome is a 60% schedule overrun. Thus, using
either approach, Diamer-Bhasha can be expected to only open in
2027 when there remains a 20% risk of further delay. Pakistan is
facing an energy crisis today (Kessides, 2011). A dam that brings
electricity is 2027 will be a little late in coming.
Note, however, that if a dam of dimensions similar to Diamer-
Bhasha were being built in the US (with its high per capita income
of approximately USD 38,000), Model 3 predicts that it would face
a schedule slippage of a mere 0.05%. Recall that per capita income
Table 8
Comparing large dams with other infrastructure asset classes.
Category Types of projects Mean cost overrun Applicable capital expenditure optimism
bias uplifts (constant prices)
50th percentile 80th percentile
Roads Motorway, trunk roads, local road, bicycle facilities,
pedestrian facilities, park and ride, bus lane schemes,
guided buses
20% 15% 32%
Rail Metro, light rail, guided buses on tracks,
conventional rail,
high speed rail
45% 40% 57%
Fixed links Bridges, tunnels 34% 23% 55%
Building projects Stations, terminal buildings 451%
a
Standard civil engineering 344%
a
Non-standard civil engineering 666%
a
Mining projects 14%
b
Thermal power plants 6%
c
Large dam projects Large hydropower, large irrigation, ood control,
multipurpose dams
96% 26% 99%
Nuclear power plants 207%
d
109281%
d
a
Based on Mott MacDonald (2002).
b
Based on Bertisen and Davis (2008).
c
Based on Bacon and Besant-Jones (1998, p.321), included for an approximate comparison purposes only, reference class probability distribution not available.
d
Based on Schlissel and Biewald (2008, p.8) review of the U.S. Congressional Budget Ofce (CBO) data from Energy Information Administration, Technical Report DOE/
EIA-0485 (January 1, 1986).
2
4
6
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Acceptable chance of overruns
Required uplift
(as multiple of estimated cost)
Fig. 7. Required uplift for large dam projects as function of the maximum
acceptable level of risk for cost overrun, constant local currency terms (N¼245).
A. Ansar et al. / Energy Policy 69 (2014) 435654
is a useful proxy for the economic support that a country can
provide for the construction of complex facilities. This suggests
that rich and not developing countries best attempt very large
energy projects, such as large dams. Even so, richer countries
should also consider alternatives and should adopt the risk
management measures of the outside view illustrated here to
choose prudently among energy alternatives.
Using their insidecost estimates, the net present benets to
cost ratio of the dam according to experts is 1.43 (WAPDA, 2011).
Even assuming experts' calculations about potential benets are
accurate, although this is a doubtful assumption, the de-biased
cost forecasts require an uplift of 4499% in constant prices
suggest that the benets to cost ratio will be below one. The
Diamer-Bhasha dam is a non-starter in Pakistan. This is without
even discussing potential effects of ination and interest rates,
potential social and environmental costs, and opportunity cost
Pakistan could earn by committing such vast amount of capital to
more prudent investments.
Our reference class forecasting techniques suggests that other
proposed large dam projects such as Belo Monte, Myitsone, or the
Gilgel Gibe III among many others in early planning stages are
likely to face large cost and schedule overruns seriously under-
mining their economic viability. Large dams also exert an oppor-
tunity cost by consuming scarce resources that could be deployed
to better uses, sinking vast amounts of land that could have
yielded cash ows and jobs from agricultural, timber, or mineral
resources. Risks related to dam safety, environment, and society
further undermine viability of large dams. Decision-makers are
advised to carefully stress test their proposed projects using the
risk management techniques of the outside view proposed here
before committing resources to them.
The outside view techniques applied to large dams above have
broader application in energy policy by helping public agencies
(e.g. national planning and nance ministries, power and water
authorities), private entrepreneurs and investors a framework to
improve upfront selection among alternatives. The problems of
cost and schedule overrun are not unique to large hydropower
dams. Preliminary research suggests that other large-scale power
projects using nuclear, thermal, or wind production technologies
face similar issues. Our research of large hydropower projects
reveals that there is a serious dearth of valid and reliable data on
the risk proles of actually completed energy projects across the
board. Much of the data in existing literature are drawn from
surveys and interviews of dubious validity. At times, interest
groups, seeking to promote a particular kind of scale or technol-
ogy, also report distorted data. There is thus an urgent need to
empirically document, in a comprehensive global database, the
risk proles of energy infrastructure assets of large, medium, and
small scales across production technologies. For example, compar-
ing the likely actual cost, schedule, and production volumes of a
large hydropower dam project versus an on-site combined heat
and power generator.
We propose that prior to making any energy investment, policy
makers consult a valid and reliable outside viewor reference
class forecast(RCF) that can predict the outcome of a planned
investment of a particular scale or production technology based
on actual outcomes in a reference class of similar, previously
completed, cases. Rigorously applying reference class forecasting
to energy investments at various scales and production technol-
ogies will yield the following contributions:
Create transparency on risk proles of various energy alter-
natives, from not only the perspective of nancial cost and
benet but also environmental and social impacthard evi-
dence is a counter-point to experts' and promoters' oft-biased
inside view.
Improve resource allocation through outside-in view to estimate
costs, benets, time, and broader impacts such as greenhouse
gas emissions incurred in building a project and emission
created or averted once a project becomes operational.
A comprehensive global dataset that can create such transpar-
ency on risk proles of energy alternatives does not yet exist. We
have sought to bridge this precise gap by providing impartial
evidence on large hydropower dam projects. As a venue for further
research we hope valid and reliable data on the actual cost,
schedules, benets, and impacts of other production technologies
will become available to enable comparative analysis with novel
implications for theory and practice.
Acknowledgments
This work was supported by funding from Oxford University's
Centre of Major Programme Management at Saïd Business School.
The authors are grateful to the United Nations Environment
Programme (UNEP) and Messrs. Jeremy Bird and Bruce Aylward
who made access to data from the World Commission of Dams
possible. We also thank Mr. John Besant-Jones at the World Bank
for advice on data collection. Professor Lord Robert May of Oxford,
Professor Nassim Taleb, Professor Jim Hall, and two anonymous
referees provided helpful comments on earlier versions of the
paper.
Appendix A. Visual representation of Model 1 (reported in
Table 3)
See Fig. A1.
References
Ansar, A., Flyvbjerg, B., Budzier, A., 2013. Do Hard StatesMake Better Decisions?
Incentives versus Biases. Unpublished manuscript. Saïd Business School, Uni-
versity of Oxford, 42 pp.
Fig. A1. .
A. Ansar et al. / Energy Policy 69 (2014) 4356 55
Ansar, A., Flyvbjerg, B., Budzier, A., 2012. Big is Fragile: The Effects of Temporal
Uncertainty and Sunk Costs on the Outcomes of Big Ventures. Unpublished
manuscript. Saïd Business School, University of Oxford, 58 pp.
Bacon, R.W., Besant-Jones, J.E., 1998. Estimating construction costs and schedules:
experience with power generation projects in developing countries. Energy
Policy 26, 317333.
Bakker, K., 1999. The politics of hydropower: developing the Mekong. Political
Geogr. 18, 209232.
Bertisen, Jasper, Davis, Graham A., 2008. Bias and Error in Mine Project Capital Cost
Estimation. The Eng. Economist 53 (2), 118139 .
Billington, D.P., Jackson, D.C., 2006. Big Dams of the New Deal Era: A Conuence of
Engineering and Politics. University of Oklahoma Press.
Buehler, R., Grifn, D., Ross, M., 1994. Exploring theplanning fallacy: Why people
underestimate their task completion times. J. Pers. Soc. Psychol. 67, 366.
Bulow, J., Rogoff, K., 1990. Cleaning up third world debt without getting taken to the
cleaners. J. Econ. Perspect. 4, 3142.
Copeland, T., Tufano, P., 2004. A real-world way to manage real options. Harv. Bus.
Rev. 82, 9099.
Dawes, R.M., Faust, D., Meehl, P.E., 1989. Clinical versus actuarial judgment. Science
243, 16681674.
Duo, E., Pande, R., 2007. Dams. Q. J. Econ. 122, 601646.
Federal Transit Administration (FTA), 2003. Predicted and Actual Impacts of New
Starts Projects: Capital Cost, Operating Cost and Ridership Data. U.S. Depart-
ment of Transportation, Washington, DC.
Flyvbjerg, B., 2003. Delusions of success: comment on Dan Lovallo and Daniel
Kahneman. Harv. Bus. Rev. 81, 121122.
Flyvbjerg, B., 2006. From Nobel Prize to project management: getting risks right.
Project Manage. J. 37, 5.
Flyvbjerg, B., 2008. Curbing optimism bias and strategic misrepresentation in
planning: reference class forecasting in practice. Eur. Plann. Stud. 16, 321.
Flyvbjerg, B., 2009. Survival of the unttest: why the worst infrastructure gets built
and what we can do about it. Oxford Rev. Econ. Policy 25, 344.
Flyvbjerg, B., Bruzelius, N., Rothengatter, W., 2003. Megaprojects and Risk: An
Anatomy of Ambition. Cambridge Univ Press, Cambridge.
Flyvbjerg, B., Budzier, A., 2011. Why your IT project may be riskier than you think.
Harv. Bus. Rev. 89, 2325.
Flyvbjerg, B., COWI, 2004. Procedures for dealing with optimism bias in transport
planning: Guidance document 200609 [WWW Document]. URL goo.gl/lLiNj.
Flyvbjerg, B., Holm, M., Buhl, S., 2005. How (in) accurate are demand forecasts in
public works projects. J. Am. Plann. Assoc. 71, 131146.
Flyvbjerg, B., Holm, M.S., Buhl, S., 2002. Underestimating costs in public works
projects: error or lie? J. Am. Plann. Assoc. 68, 279295.
Flyvbjerg, B., Garbuio, M., Lovallo, D., 2009. Delusion and deception in large
infrastructure projects. Calif. Manage. Rev., 51.
Gelman, A., Hill, J., 2007. Data Analysis Using Regression and Multilevel/Hierarch-
ical Models. Cambridge University Press, Cambridge.
Gilovich, T., Grifn, D., Kahneman, D., 2002. Heuristics and Biases: The Psychology
of Intuitive Judgment. Cambridge University Press.
Goel, R.K., Singh, B., Zhao, J., 2012. Underground Infrastructures: Planning, Design,
and Construction. Butterworth-Heinemann, Waltham, MA.
Goldstein, H., 2010. Multilevel Statistical Models, 4th ed. Wiley.
Hoek, E., Palmieri, A., 1998. Geotechnical risks on large civil engineering projects.
In: Proceedings of the 8th Congres IAEG, pp. 7988.
Hufschmidt, M.M., Gerin, J., 1970. Systematic errors in cost estimates for public
investment projects. In: Margolis, J. (Ed.), The Analysis of Public Output.
Columbia University Press, New York.
ICOLD, 2010. International. Commission on Large Dams [WWW Document]. URL
http://www.icold-cigb.org/.
IEA, 2011. Policies and Scenarios [WWW Document]. URL http://www.iea.org/
publications/scenariosandprojections/.
Kahneman, D., 1994. New challenges to the rationality assumption. J. Inst. Theor.
Econ. 150, 1836.
Kahneman, D., 2011. Thinking, Fast and Slow, 1st ed. Farrar, Straus and Giroux, New
York.
Kahneman, D., Lovallo, D., 1993. Timid choices and bold forecasts: a cognitive
perspective on risk taking. Manage. Sci., 1731.
Kahneman, D., Tversky, A., 1979a. Prospect theory: an analysis of decision under
risk. Econometrica, 263291.
Kahneman, D., Tversky, A., 1979b. Intuitive prediction: biases and corrective
procedures. In: Makridakis, S., Wheelwright, S.C. (Eds.), Studies in the Manage-
ment Sciences: Forecasting. North Holland, Amsterdam.
Kessides, I.N., 2011. Chaos in Power: Pakistan's Electricity Crisis.
Kreft, I.G.G., Leeuw, J.d.e., 1998. Introducing Multilevel Modeling. Sage Publications
Ltd.
Laird, N.M., Ware, J.H., 1982. Random-effects models for longitudinal data. Bio-
metrics, 963974.
Lovallo, D., Kahneman, D., 2003. Delusions of success. Harv. Bus. Rev. 81, 5663.
Luehrman, T.A., 1998. Strategy as a portfolio of real options. Harv. Bus. Rev. 76,
89101.
McCully, P., 2001. Silenced Rivers: the Ecology and Politics of Large Dams. Zed
Books, London.
Merewitz, L., 1973. How do Urban Rapid Transit Projects Compare in Cost
Estimating Experience. Institute of Urban & Regional Development, University
of California.
Merrow, E.W., McDonnell, L., Argüden, R.Y., Corporation, R., 1988. Understanding
the Outcomes of Megaprojects: A Quantitative Analysis of Very Large Civilian
Projects. Rand Corp.
Mold, A., 2012. Will it all end in tears? Infrastructure spending and African
development in historical perspective. J. Int. Dev. 24, 237254.
Mott MacDonald, 2002. Review of Large Public Procurement in the UK, Study for
HM Treasury. HM Treasury, London. www.hm-treasury.gov.uk/d/7(3).pdf.
Nilsson, C., Reidy, C.A., Dynesius, M., Revenga, C., 2005. Fragmentation and ow
regulation of the world's large river systems. Science 308, 405408.
Pickrell, D.H., 1989. Urban Rail Transit Projects: Forecast Versus Actual Ridership
and Costs (No. PB90148693), Ofce of Grants Management, Urban Mass
Transportation Administration. Transportation System Center, U.S. Department
of Transportation, Cambridge, MA.
Pickrell, D.H., 1992. A desire named streetcar: fantasy and fact in rail transit
planning. American Planning Association. J. Am. Plann. Assoc. 58, 158176 .
Rasbash, J., Browne, W., Goldstein, H., Yang, M., Plewis, I., Healy, M., Woodhouse, G.,
Draper, D., Langford, I., Lewis, T., 2009. A User's Guide to MLwiN. University of
London, Institute of Education, Centre for Multilevel Modelling.
Recchia, A., 2010. R-squared measures for two-level hierarchical linear models
using SAS. J. Stat. Softw., 32.
Richter, B.D.a., Postel, S.b., Revenga, C.c., Scudder, T.d., Lehner, B.e., Churchill, A.f.,
Chow, M., 2010. Lost in development's shadow: the downstream human
consequences of dams. Water Alternatives 3, 1442.
Salazar, J.G., 2000. Damming the child of the ocean: the Three Gorges project.
J. Environ. Dev. 9, 160174.
Scudder, T., 2005. The Future of Large Dams: Dealing with Social, Environmental,
Institutional and Political Costs. Earthscan, London.
Schlissel, D., Biewald, B., 2008. Nuclear Power Plant Construction Costs. Synapse
Energy Economics, Inc.
Singh, S., 2002. Taming the Waters: The Political Economy of Large Dams in India.
Oxford University Press, New Delhi.
Sovacool, B.K., Bulan, L.C., 2011. Behind an ambitious megaproject in Asia: the
history and implications of the Bakun hydroelectric dam in Borneo. Energy
Policy 39, 48424859.
Sovacool, B.K., Cooper, C., 2013. The Governance of Energy Megaprojects: Politics,
Hubris and Energy Security. Edward Elgar, Cheltenham, UK.
Stone, R., 2011. Hydropower. The legacy of the Three Gorges Dam. Science 333, 817.
Taleb, N.N., 2010. The Black Swan: The Impact of the Highly Improbable with a new
section: On Robustness and Fragility, 2nd ed. Random House Trade
Paperbacks.
Tversky, A., Kahneman, D., 1974. Judgment under uncertainty: heuristics and biases.
Science 185, 11241131.
Wachs, M., 1989. When planners lie with numbers. J. Am. Planning Assoc. 55,
476479.
WAPDA, 2011. Diamer-Basha Dam Project Status as on 20 June 2011.
World Bank, 1968. Report and Recommendation of the President to the Executive
Directors on a Proposed Loan for the Tarbela Project and on Proposed Extension
of Closing Dates of Loan 266-PAK (Indus Basin Project), Pakistan (No. P-616).
The World Bank, Washington, DC.
World Bank, 1970. Staff Appraisal Report (pu-31a) of the Chivor Hydroelectric
Project, Interconexion Electrico S.A., Colombia (No. PU-31a). The World Bank,
Washington, DC.
World Bank, 1973. Staff Appraisal Report (150a-BR) of the Itumbiara Hydroelectric
Power Project, FURNAS, Brazil (No.150a-BR). The World Bank, Washington, DC.
World Bank, 1996. Estimating Construction Costs and Schedules: Experience with
Power Generation Projects in Developing Countries (No. WTP325). The World
Bank, Washington, DC.
World Commission on Dams (WCD), 2000. Cross-Check Survey: Final Report. Cape
Town, South Africa.
Ziv, G., Baran, E., Nam, S., Rodríguez-Iturbe, I., Levin, S.A., 2012. Trading-off sh
biodiversity, food security, and hydropower in the Mekong River Basin. Proc.
Natl Acad. Sci. 109, 5609.
A. Ansar et al. / Energy Policy 69 (2014) 435656
... Literature on energy megaprojects has been strongly influenced by Flyvbjerg's argument that megaprojects are consistently 'over budget, over time, over and over again', which he refers to as the 'Iron Law of Megaprojects'. This framework has been used to demonstrate cost overruns, complexity and delays in energy megaprojects in Canada, China and Brazil (Joseph et al. 2020, Ansar et al. 2014Kian et al. 2016). Other studies do not extensively engage with broader literature on megaprojects, but provide important insights about the challenges of largescale energy infrastructure (El Mekaoui et al. 2020). ...
... Studies on megaprojects perceive the lack of transparency as a critical governance challenge (Flyvbjerg 2014;. A number of scholars claim that the enormous amount of funds involved in the large-scale infrastructure and weak governance systems of developing nations lead to ample opportunities for unethical practices (Ansar et al. 2014;Gutierrez et al. (2019); Callegari et al., 2018). For example, Shell has been accused of being complicit in corrupt practices and ignoring human rights issues in relation to its energy projects in Nigeria (AI 2017). ...
... Within literature on megaprojects, a number of studies have critiqued the social costs of mega-plantations (Kenney-Lazar 2019), large-scale hydroelectric dams (Ansar et al. 2014;Callegari et al. 2018) and pipeline infrastructures (Ogwang and Vanclay 2021). Van de Graaf and Sovacool (2014) find that while megaprojects do little to improve energy access for vulnerable populations, they exact enormous costs in terms of restricting access to land and water resources and undermining livelihoods. ...
Article
Full-text available
For the first time since the end of British colonisation in 1947, the poorly integrated and politically volatile region of South Asia is about to embark on the development of several multilateral energy megaprojects. Historically, regional energy cooperation in South Asia has been undermined by a range of geopolitical issues. Political developments in the last decade have faciliated an unprecedented level of cooperation on energy, and currently several multilateral pipelines and hydroelectric projects are at advanced stages of implementation. Yet,, the progress in the physical and bureaucratic development of energy projects has not been complemented by a corresponding conceptualisation of the governance challenges of these initiatives. Within academic literature, there is a dearth of studies that analyses the critical aspect of collectively governing the proposed and under-construction energy megaprojects in South Asia. This gap is concerning as the complexity of energy megaprojects and weak governance systems in developing countries can lead to enormous economic, social and environmental externalities. This article aims to identify the governance challenges of South Asia’s energy megaprojects. To this end, the article draws on literature on global energy governance and energy megaprojects and primary data from interviews undertaken with policymakers. The article synthesises literature on energy cooperation in South Asia with existing knowledge on megaprojects and global energy governance. In doing so, the paper addresses one of the key gaps in literature, which is the lack of studies that appraises megaprojects in the context of interactions between domestic and international politics. Due to the growing importance of energy megaprojects to regional integration and energy transition, the findings of this article have relevance towards policies on global governance and sustainable development.
... The adaptation implementation could be impeded by conflicting interests among stakeholders. A particular example is the long-lasting debate on whether to construct reservoirs to mitigate water crisis (e.g., Ansar et al., 2014;Berga, 2016;Poff et al., 2015; 37 Chapter 2. Literature review Zarfl et al., 2015). In addition, how to motivate water managers to positively participate in adapting to global change is rather challenging. ...
Thesis
Understanding the vulnerability of water management under global change is the premise for designing adaptation actions. A comprehensive assessment of current water management vulnerability to future changes hinges on new tools that are able to represent human impact on water resources and innovative frameworks that are able to generate new insights to inform adaptation designing. Therefore, this dissertation sets out to (1) develop and improve models to represent water resources, water demand, and water management in an integrated hydrological modelling framework; (2) apply a "scenario-neutral" bottom-up framework and a "scenario-led" top-down framework to identify and investigate plausible vulnerability and impact under global change. These developments and applications are demonstrated by taking the Neste water system in French Pyrenees as a case study.
... A general simple calculation of the payback period shows a pay-off period of less than 5 years. However, Ansar, Flyvbjerg [29] showed that in most cases budgets for such large-scale projects do go above the original estimates. ...
... NASA is a public sector entity and thus subject to extensive scrutiny. To its credit, NASA has made data access to convenient and transparent, where common practice among sponsors of big projects is to hide or obfuscate data (Ansar, 2017, Ansar et al., 2016Flyvbjerg, 2014). Duraiswamy (2017, p 26, Table 3.2) collected NASA data from NASA budget disclosures, U.S. government audit reports, U.S. congressional hearings, logs of space launches, industry data, company disclosures, academic articles, news articles, and documentary programs (Congressional Budget Office, CBO, 2005;Emmons et al., 2007;Butts and Linton, 2009;Government Accountability Office, GAO, 2009, 2010, 2013, 2015, 2017; Office of the Inspector General, 2012). ...
Article
Full-text available
How should government and business leaders solve big problems? Ought policy responses to occur in bold leaps or multitudinous methodical moves? Here we show that one-off major projects, with a high level of bespoke content, are prone systematically to poorer outcomes than projects built with a repeatable platform strategy. Repeatable projects are cheaper, faster, and scale in volume and variety at much lower risk of failure. We arrive at these conclusions using comparative evidence—NASA vs SpaceX—on cost, speed-to-market and schedule, and scalability outcomes of their respective space missions. Our reference class dataset consists of 203 space missions spanning 1963–2021, of which 181 missions belong to NASA and 22 belong to SpaceX. We find that SpaceX’s platform strategy was 10X cheaper and 2X faster than NASA’s bespoke strategy. Moreover, SpaceX’s platform strategy was less risky, virtually eliminating cost overruns. We further show that achieving platform repeatability is a strategically diligent process involving experimental learning sequences. Sectors of the economy where governments find it difficult to control spending or timeframes or to get benefits quickly enough—e.g. health, education, climate, defence—are ripe for a platform rethink.
Article
Full-text available
There are few scientific studies evaluating the impact of the loss of wetlands on the banks of tropical rivers on assemblages of small non‐flying mammals. To understand the possible deleterious effects of hydroelectric construction in tropical forests on this group of mammals, we used data from 2 years of monitoring carried out during the period before the filling of a hydroelectric plant reservoir in the Brazilian Amazon, and related them to vegetation, soil and topography. We captured 659 individuals of 20 small‐mammal species. The species assemblage composition in the flooded areas was a subset of species that occurred in both floodable and non‐floodable areas, and only one species was captured exclusively in the flooded area. Species composition was influenced by the proportion of sand, by soil nutrient concentration and distance from water bodies. We conclude that there is no evidence that the flooding of low‐lying areas along the Madeira River would negatively affect the assemblage of non‐flying small mammals in the short term because the remaining areas have similar assemblages of small mammals as those destined for flooding. Whether the area lost will be important for population dynamics will depend on the conservation of the remaining areas. Existem poucos estudos científicos avaliando o impacto da perda de áreas úmidas nas margens de rios tropicais sobre as assembleias de pequenos mamíferos não voadores. Para entender os possíveis efeitos deletérios da construção de hidrelétricas em florestas tropicais sobre esse grupo de mamíferos, utilizamos dados de dois anos de monitoramento realizado durante o período anterior ao enchimento de um reservatório de hidrelétrica na Amazônia brasileira e os relacionamos com a vegetação, solo e topografia. Capturamos 659 indivíduos de 20 espécies de pequenos mamíferos. A composição da assembleia de espécies nas áreas alagadas foi um subconjunto das espécies presentes nas áreas não alagáveis, e apenas uma espécie foi capturada exclusivamente na área alagável. A composição de espécies foi influenciada pela proporção de areia, pela concentração de nutrientes no solo e pela distância dos corpos d'água. Concluímos que não há evidências de que a inundação de áreas baixas ao longo do rio Madeira afetaria negativamente a assembleia de pequenos mamíferos não‐voadores no curto prazo, pois as áreas remanescentes apresentam assembleias de pequenos mamíferos semelhantes às destinadas a inundação. Se a área perdida será importante para a dinâmica populacional dependerá da conservação das áreas não inundadas.
Chapter
Global warming is becoming a very challenging issue for the current world. With more than 800 million people all over the world, global warming seems to be chiefly anthropogenic. It is estimated that the global temperature of our planet will be enhanced by 1.5°C in the coming decades as the greenhouse gases continue to pollute our atmosphere. Gases like carbon dioxide, methane nitrous oxide, and sulfur dioxide are being continuously released into our atmosphere. Withered crops, starving humans and animals, heat waves, heavy rains, and shrinking glaciers are giving a chilling picture of climate change. Apart from this, urbanization and industrialization resulting in land degradation and soil erosion are further worsening the situation. As far as forecasting and management are concerned, scientists all over the world are trying to understand the phenomena and helping people by predicting, forecasting, and diminishing the disaster. They are focusing on risk and hazard assessment, research, and educational programs on climate variability, improvement in forecast skills, and teaching from disasters. Climate predictions aid in the prediction of big weather occurrences that affect many facets of human life. In this chapter, the authors will try to focus on forecasting and management of disasters happening due to extreme climate change all over the world.
Article
This analysis of the human rights to safe drinking water and sanitation (HRtWS) uncovers why some groups around the world are still excluded from these rights. Léo Heller, former United Nations Special Rapporteur on the human rights to water and sanitation, draws on his own research in nine countries and reviews the theoretical, legal, and political issues involved. The first part presents the origins of the HRtWS, their legal and normative meanings and the debates surrounding them. Part II discusses the drivers, mainly external to the water and sanitation sector, that shape public policies and explain why individuals and groups are included in or excluded from access to services. In Part III, public policies guided by the realization of HRtWS are addressed. Part IV highlights populations and spheres of living that have been particularly neglected in efforts to promote access to services.
Article
Does climate change adaptation require that investments are designed to be more robust? What about when climate change is more uncertain? What if the climate changes faster? This decision problem is difficult if the design of the investments is irreversible for their lifetime, for instance, in the construction industry. We study an irreversible design decision when the investment starts, combined with an irreversible option to abandon. The design determines the investment’s robustness to sustain detrimental conditions. We find that for short-lived investments, optimal robustness decreases if the climate changes faster, and increases if uncertainty is higher. For long-lived investments, these effects reverse. This has implications for decision makers who plan infrastructure adaptation, for instance, that adverse climate change does not require more robust investments under the identified circumstances.
Article
Construction projects usually cover long implementing periods, and exert profound influence on human society and development, therefore leading to increasing concerns on dynamic sustainability over their life cycle. Sustainability of construction projects can only be achieved when the sustainability in economic, social and environmental dimensions are adequately addressed. It is therefore important to develop an effective model for understanding dynamic sustainability of construction projects with coordination perspective incorporated. In this paper, the dynamic sustainability assessment model of construction projects is proposed by adopting system dynamic and coupling coordination degree methods collectively, and the weighting values between dimensions are assigned from coordination perspective. The sustainability of a construction project is measured not only by the total value of sustainability in all dimensions, but also the coordination of sustainability between various dimensions. Moreover, according to the assessment results, an adjustment procedure is designed for simulating the formulation process of improvement strategy. The effectiveness of the developed sustainability assessment model is demonstrated with its application in three scenarios based on a case study. The results of case demonstration show that the overall sustainability value of case project is 0.63, and this value can be increased to 0.69 when effective strategies are adopted. Sustainability of the case project in environmental and social dimensions at operation and demolition stages should be enhanced. In applying the assessment model, dynamic sustainability of construction projects can be properly realized by project managers, and therefore proposing effective strategies to improve their sustainability and raise long-term value.Graphical abstract
Preprint
Micro-hydrokinetic turbines are hydropower turbines that can utilize the current of the river to generate power without constructing a dam. Recent advances in micro-hydrokinetic turbine technology have made their use an attractive alternative to traditional hydropower systems that require the construction of a dam. In this paper, we introduce a novel and practical analysis of continental to global scale hydrology and hydraulics that we use to assess continental United States (CONUS) stream locations with optimum hydraulic characteristics for an example micro-hydrokinetic turbine (Volturnus). We then utilize the boundaries of CONUS military installations to determine which installations possess the greatest potential to harvest hydropower using the Volturnus. For the Volturnus, our results indicate that although less than 1% of CONUS stream locations possess the hydraulic characteristics necessary for the Volturnus to function optimally, this translates into roughly 231-549 millon kWh per week. Around 27-42 CONUS installations possess streams that are sometimes capable of producing hydropower with the Volturnus and 12-23 of those military installations possess optimum hydraulic conditions for the Volturnus year round. In our analysis, reducing either the minimum velocity or minimum depth threshold of the Volturnus nearly double the proportion of CONUS locations with optimum Volturnus hydraulic characteristics.
Article
Full-text available
This article presents results from the first statistically significant study of traffic forecasts in transportation infrastructure projects. The sample used is the largest of its kind, covering 210 projects in 14 nations worth US$59 billion. The study shows with very high statistical significance that forecasters generally do a poor job of estimating the demand for transportation infrastructure projects. The result is substantial downside financial and economic risks. Such risks are typically ignored or downplayed by planners and decision makers, to the detriment of social and economic welfare. For nine out of ten rail projects passenger forecasts are overestimated; average overestimation is 106 percent. This results in large benefit shortfalls for rail projects. For half of all road projects the difference between actual and forecasted traffic is more than plus/minus 20 percent. Forecasts have not become more accurate over the 30-year period studied. If techniques and skills for arriving at accurate demand forecasts have improved over time, as often claimed by forecasters, this does not show in the data. The causes of inaccuracy in forecasts are different for rail and road projects, with political causes playing a larger role for rail than for road. The cure is transparency, accountability, and new forecasting methods. The challenge is to change the governance structures for forecasting and project development. The article shows how planners may help achieve this.
Book
Based on extensive original research, this book explores the technical, social, political, and economic dimensions of four Asian energy megaprojects: a regional natural gas pipeline network in Southeast Asia, a series of hydroelectric dams on the island of Borneo, an oil pipeline linking Europe with the Caspian Sea, and a very large solar energy array in the Gobi desert. © Benjamin K. Sovacool and Christopher J. Cooper 2013. All rights reserved.
Code
R package for Data Analysis using multilevel/hierarchical model
Book
This book explains the history and politics of dam building worldwide. It describes the many technical, safety and economic problems that afflict the technology, and explores the role played by international banks and aid agencies in promoting it. The author also examines the rapid growth of the international anti-dam movements, and stresses how replacing large dams with less destructive alternatives will depend upon opening up the dam industry's practices to public scrutiny.