Public-Private Partnerships and Prices: Evidence from Water Distribution in France

1
Public-Private Partnerships and Prices Evidence
from Water Distribution in France

• Stéphane Saussier
• ADIS GRJM University of Paris XI Sceaux
• ATOM University of Paris I Sorbonne

2
Introduction

• A precise question Should the State (or local
  public authorities) contract with a private
  operator in order to provide a public service?
• A large range of transactions can possibly be
  organized through PPP.
• Possible advantages
• 1/lower capital constraints
• 2/incentive-based contracts
• 3/encouragement of innovation
• 4/lower production costs (scale and scope
  economies).

3
Introduction

• The efficiency of PPP remains an unresolved
  question
• At the theoretical level
• TCT Williamson 1976, Williamson 1999
• ICT Hart-Shleifer-Vishny 1997 Hart 2003
  Bennett Iossa 2002.
• AT Bentz Grout 2002.
• At the empirical level
• Guash al 2002, Estache 2003, Littlechild 2002,
  Zupan 1989,

4
Introduction

• The case of water supply in France
• Interest
• Local public authorities - plenty of data
• One institutional framework
• Empirical studies not conclusive DGS 1995,
  DGCCRF 1997, IFEN 2000, Cour des comptes 1997,
  2004.
• An empirical paper
• A transaction cost economics approach

5
I. A TCE approach

• A TCE approach would stress a simple idea  The
  common error to be avoided is to pronounce that
  governance structures are efficient or
  inefficient without reference to the transaction
  there are transactions for which the firm is
  superior to the market, its bureaucratic
  disabilities notwithstanding  O.E. Williamson
  2002, page 20.
• Possible drawbacks of PPP contractual costs
• Transaction costs may, in certain situations,
  overcome advantages in terms of production costs

6

I. A TCE APPROACH
What are the potential transaction costs?
Choice of the Private Operator
Contract renewal
t
Service specification
Contract execution
Competition for the field
Competition for the field
7
Franchising problems to overcome ex ante
I. A TCE APPROACH

• Specification of the franchised service and
  effective competitive bidding
• Imprecise specification
• adverse selection problem (the winners curse)
• Small number of bidders
• Price and quality do matter ? Artificial or
  obscure award criterion.

8
Franchising problems to overcome ex post
I. A TCE APPROACH

• Risks of opportunism
• Delays in construction and provision
• Efforts to renegotiate the contract
• Lower level of quality than promised
• Absence of responsiveness to consumers needs
• Price increase
• Risks of maladaptation
• Price-cost divergences
• Indeterminacies
• Those problems are particularly important when
  the transaction is uncertain and needs high level
  of asset specificity

9
Potential transaction costs solutions exist!
I. A TCE APPROACH

• Contractual solutions
• Choice of the operator
• Menu auctions
• Pre-qualification criteria
• Bids formulated in terms of a constant revenue
  stream (E-F-G 02)
• Allocation of risk
• Minimum level of quality required
• Execution stage
• Dissuasive but realistic penalty clauses
• Enforceable sanctions Monitoring and accounting
  procedures Capacity of expertise
• Open Book Policy Yardstick competition
  Information transparency

10
Potential transaction costs solutions exist!
I. A TCE APPROACH

• Institutional solutions
• Credibility of the contracting parties
• Reputation
• Institutional dotations (corruption, property
  rights )
• Laws that constrain contractual choices and
  potential opportunistic behavior

11
General conclusions from a TCE perspective
I. A TCE APPROACH

• No institutional choice is better than another a
  priori.
• The choice between PPP and direct public
  management depends on the characteristics of the
  transactions (asset specificity / uncertainty /
  complexity)
• Recent developments show that solutions may exist
  that qualify PPP in many situations (Littlechild
  2002 Engel, Fisher, Galetovic 02, )

12
II. Putting TCE to workData

• Data sets coming from the French Environment
  Institute and the French Ministry of Agriculture
• A representative sample of all the French local
  authorities
• 5,000 local authorities in 1998
• Restricted to 4,443 observations in our
  statistics
• Restricted to 3,613 observations in our
  econometric tests
• Account for more than 50 of the consumers
• The question Price, cost and organizational
  choices

13
The case of water supply in France
TCE TO WORK

• French local authorities decide / not the State
• A great variety of organizational choices

14
Putting TCE to workThe case of water supply in
France
TCE TO WORK

• Characteristics of transactions and potential
  transaction costs
• Asset specificity
• Physical investments are specific (site
  specificity)
• Complexity
• Mainly due to
• ? the difficulty of the treatments for producing
  drinkable water (e.g. the quality of the water
  before treatment)
• ? The difficulty of transporting water to the
  consumer (the density of population)

15
Institutional environment and potential
transaction costs
TCE TO WORK

• Drawbacks of PPP resulting from contract
  incompleteness can be reduced by the
  institutional framework
• The intuitu personae principle
• May avoid problems linked to the choice of the
  operator (Bajari-McMillan-Tadelis 2003)
• Risk of corruption/collusion
• PPP are  administrative  contracts
• An asymmetric position between contracting
  parties
• Ex post renegotiations take place within a rigid
  framework

16
In what situations could drawbacks overcome
advantages of a PPP?
TCE TO WORK
17
The case of water supply in France
TCE TO WORK

• The use of PPP

18
The case of water supply in France
TCE TO WORK

• Prices (for one cubic meter in francs) and
  geographical areas

19
Prices and organizational choices a first look
TCE TO WORK
Prices for local authorities that do not contract
out the service are lower The price difference
between lease and direct public management is 32
In euros for 120 m3, without tax

• Two possible interpretations
• Collusion prices - PPP are inefficient because of
  ex ante transaction costs
• PPP are specialized on complex transactions - PPP
  are efficient

20
Prices and organizational choices an econometric
test
TCE TO WORK
Other control variables used in the econometric
study the area (France is divided in 100
departments) the technology used in measuring
and controlling links on the network the
organizational choice made concerning the
cleaning of used water

21
Prices and organizational choices an econometric
test
TCE TO WORK

• Naïve OLS equation, controlling for other
  parameters than organizational choices that may
  have an impact on prices

?R2 0,50 ? Lease contracts account for 27 in
the price (average bill of 140 ) ?All types of
PPP lead to an increase of price paid by the
consumer
22
Prices and organizational choices an
econometrical test
TCE TO WORK

• Cost structure might be different from one
  organizational choice to another
• We estimate 2 equations
• Possible selection bias
• Organizational choices are not randomly chosen ?
  TCE propositions ? Heckman Method

23
In what situations could drawbacks overcome
advantages of a PPP?
TCE TO WORK

• Robust probit estimate of lease vs. direct public
  management
• Complex transactions (complex treatments, low
  density of population) are more probably
  organized through lease contracts
• Non-linear effect of the population size
• More than 75 of observed choices accurately
  predicted

24
Prices and organizational choices an
econometrical test
TCE TO WORK
Correcting for selection bias, direct public
management is still the most efficient
organizational choice
? However the price difference between lease and
direct public management decreases again (About
10 )
25
The impact of mis-specified modes of organization
TCE TO WORK

• Do transaction cost propositions affect
  performances?
• Do organizational choices matter?
• Theoretical Misfit variable defined as equal to
  1 as soon as local authorities do not respect TCE
  propositions (focus on complex transactions).
• 13 of our observations do not fit TCE
  propositions

26
The impact of mis-specified modes of organization
TCE TO WORK

• Local authorities that did not choose modes of
  organization that correspond to a TCE
  prescription have higher prices.
• This result is true whatever the  direction  of
  the error.
• This result confirms previous recent works that
  attempted to connect organizational choices and
  performances through the lens of TCE in Strategic
  Management fields (Masten et al. 91, Leiblein et
  al. 02, Mayer et Silverman 02, Sampson 04,
  Nickerson et al. 97, Mayer 00, Yvrande 04,
  Bigelow 03, Nickerson et Silverman 04)

27
Conclusions and perspectives

• Organizational choices matter!
• Transaction costs matter!
• More general conclusions concerning the relative
  efficiency of public-private partnerships
• Limits concerning our results to be resolved
• What about contractual choices?
• What about price and efficiency?
• Price is questionable
• Other measures of efficiency (quality,)