Technology and incentive XI Riunione regulation in the I talian - - PowerPoint PPT Presentation

Technology and incentive regulation in the I talian motorways industry

Luigi Benfratello Alberto Iozzi Paola Valbonesi

Università di Torino Università di Roma Università di Padova and Ceris-CNR ‘Tor Vergata’

XI Riunione Scientifica SI ET Napoli, 3-5 ottobre 2007

XI Riunione Scientifica SIET Università degli Studi di Napoli "Federico II“ Napoli, 3-5 ottobre 2007 Benfratello – Iozzi – Valbonesi Technology and incentive regulation in the Italian motorways industry

2

I ntroduction

Rather limited analysis of the technology adopted by motorways

concessionaires

However, motorways concessions are very common in the EU Several differences between concession agreements across and

also within EU countries

Size Ownership Financing Tariff regulation Institutions …

Italian motorways industry is a very good laboratory to study

some of these issues

Is there any “best” solution ??

XI Riunione Scientifica SIET Università degli Studi di Napoli "Federico II“ Napoli, 3-5 ottobre 2007 Benfratello – Iozzi – Valbonesi Technology and incentive regulation in the Italian motorways industry

3

The I talian motorways industry

More than 8.000 km of motorways (around 20% of the

network in EU15), most of it built before 1975.

24 concessionaires, managing (and sometimes

building) different portions of the motorways network

Concessions awarded at different stages (mostly

during the ’60s) and expiring between 2012 and 2050

Huge differences between concessionaires: e.g.

ASPI:

2.854,6 km (3.402,3 km with partecipates)

VE-PD: 41,8 km

In the late 90s, motorways in Italy experienced a

change in

regulatory regime ( from ROR to PC)

• wnership ( privatisation)

XI Riunione Scientifica SIET Università degli Studi di Napoli "Federico II“ Napoli, 3-5 ottobre 2007 Benfratello – Iozzi – Valbonesi Technology and incentive regulation in the Italian motorways industry

4

Our dataset

We build a unique dataset with information over the

1992-2004 period for 20 Italian motorways concessionaires.

The database contains

financial and traffic indicators (costs, revenues, inputs,

kms travelled, etc),

characteristics of the motorways (length, no. of

stonework, % of network with 3 lanes, etc)

concessionaires’ institutional characteristics (ownership,

type of regulation, etc). Currently working on updating the dataset to more

recent years…

XI Riunione Scientifica SIET Università degli Studi di Napoli "Federico II“ Napoli, 3-5 ottobre 2007 Benfratello – Iozzi – Valbonesi Technology and incentive regulation in the Italian motorways industry

5

Our analysis

Our main aim is to explore the technology adopted by

motorways concessionaires, especially with respect to

Economies of scale Economies of density Technical progress

Useful policy indications, not only for Italy:

Optimal size of a concession Toll levels and toll update mechanism (e.g. X factor) Need for further network expansion

We also control for possible effects on the productivity of

the concessionaires of the regulatory reform of the 90s.

Effects of the changes in ownership Effects of the changes in regulatory regime

XI Riunione Scientifica SIET Università degli Studi di Napoli "Federico II“ Napoli, 3-5 ottobre 2007 Benfratello – Iozzi – Valbonesi Technology and incentive regulation in the Italian motorways industry

6

Regulatory reform (1)

Regulatory reform during the ’90s, main provision is

Delibera Cipe of December 1996.

All concession agreements renewed after the definition

• f this new regulatory framework

Regulatory framework resulting from these provisions

rather incomplete and somehow contradictory

More recently, further provisions attemp to modify,

complete and clarify the regulatory framework

However, the analysis of these more recent provisions

is beyond the scope of this paper, since their effects is not captured by our data.

XI Riunione Scientifica SIET Università degli Studi di Napoli "Federico II“ Napoli, 3-5 ottobre 2007 Benfratello – Iozzi – Valbonesi Technology and incentive regulation in the Italian motorways industry

7

Regulatory reform (2)

Before ~1996:

Most concessionaries publicly owned Tariffs set at the start of the concession period to ensure

cover of all costs borne over the concession period

Tariffs increased uniformly (across concessionaires and types

• f vehicles) by law each year

Building costs covered partly with government funds and

partly with highway revenues

Also,

Preferences for publicly owned firms in the award of the

concessions

Loans taken up (for building purposes) by publicly owned

concessionaires covered by government backup

Excess profits to be transferred to the government

XI Riunione Scientifica SIET Università degli Studi di Napoli "Federico II“ Napoli, 3-5 ottobre 2007 Benfratello – Iozzi – Valbonesi Technology and incentive regulation in the Italian motorways industry

8

Regulatory reform (3)

After ~1996:

Massive privatisation programme New regulatory framework:

No discrimination according to ownership Prices set at the start of the concession period to ensure cover

• f all costs borne over the concession period

Prices updated each year according to a price cap formula,

with parameters (possibly) specific to each concessionaire

Price cap formula parameters reviewed every 5 years Pricing flexibility

Also,

No back-up for loans taken up for building purposes Share of revenues transferred to the government

XI Riunione Scientifica SIET Università degli Studi di Napoli "Federico II“ Napoli, 3-5 ottobre 2007 Benfratello – Iozzi – Valbonesi Technology and incentive regulation in the Italian motorways industry

9

Regulatory reform (4)

Regulatory regime switch Ownership change

20 18 17 16 1

Concessionaires under new regulatory regime

2003 2002 2001 2000 1998 2 1992 4 1996 6 1999 13 2003

Private concessionaires

XI Riunione Scientifica SIET Università degli Studi di Napoli "Federico II“ Napoli, 3-5 ottobre 2007 Benfratello – Iozzi – Valbonesi Technology and incentive regulation in the Italian motorways industry

10

Price cap mechanism

Average toll level set at the beginning of the concession to

cover cost over the concession period

Concessionaires free to vary tolls each year, provided they

satisfy the Laspeyre-type price cap formula

ΔRPI

change in the Retail Price Index

• X
• ffset productivity factor
• ΔQ

change in a composite quality index

• β

scaling positive factor

X normally redetermined every 5 years to reflect changes in

the economic environment (traffic, productivity, costs, etc)

Q X RPI q p q p

i t i t i i t i t i

Δ + − Δ ≤ ⎥ ⎥ ⎦ ⎤ ⎢ ⎢ ⎣ ⎡ −

∑ ∑

− − −

β 100 x 1

1 1 1

XI Riunione Scientifica SIET Università degli Studi di Napoli "Federico II“ Napoli, 3-5 ottobre 2007 Benfratello – Iozzi – Valbonesi Technology and incentive regulation in the Italian motorways industry

11

Quality regulation (1)

Quality correction term allows the

concessionaires to “sell” higher quality for higher prices

Quality index Q based on:

accidents roughness of road surface

Choice lead initially by data availability (but

not amended later on)

As for β …

XI Riunione Scientifica SIET Università degli Studi di Napoli "Federico II“ Napoli, 3-5 ottobre 2007 Benfratello – Iozzi – Valbonesi Technology and incentive regulation in the Italian motorways industry

12

Quality regulation (2)

ASPI Other concessionaires if ΔQ > 0 Other concessionaires if ΔQ < 0

XI Riunione Scientifica SIET Università degli Studi di Napoli "Federico II“ Napoli, 3-5 ottobre 2007 Benfratello – Iozzi – Valbonesi Technology and incentive regulation in the Italian motorways industry

13

Quality regulation (3)

Main features

Higher quality “buys” higher prices (and viceversa) For a given increase in the quality index, the effect on prices is

larger the higher is the starting level of quality

For a given reduction of the quality index, the effect on prices

is larger the smaller is the starting level of quality

Effects are magnified when the starting level of quality if above

an “acceptable” level (Q=60)

Rationale

For quality increases, the quality correction term is cost- rather

than welfare-related

It ensures the firm recovers the cost of quality-related investments,

very much in the spirit of cost-of-service regulation

XI Riunione Scientifica SIET Università degli Studi di Napoli "Federico II“ Napoli, 3-5 ottobre 2007 Benfratello – Iozzi – Valbonesi Technology and incentive regulation in the Italian motorways industry

14

Data (1)

Balance sheet data of 20 concessionaires over

the 1992-2004 period (costs, revenues, employees)

Other variables recovered from Aiscat

publications (kms travelled, network characteristics)

Ownership and regulation information from

concessionaires’ official reports

Unbalanced panel, 253 observations

XI Riunione Scientifica SIET Università degli Studi di Napoli "Federico II“ Napoli, 3-5 ottobre 2007 Benfratello – Iozzi – Valbonesi Technology and incentive regulation in the Italian motorways industry

15

Data (2)

XI Riunione Scientifica SIET Università degli Studi di Napoli "Federico II“ Napoli, 3-5 ottobre 2007 Benfratello – Iozzi – Valbonesi Technology and incentive regulation in the Italian motorways industry

16

Data (3)

XI Riunione Scientifica SIET Università degli Studi di Napoli "Federico II“ Napoli, 3-5 ottobre 2007 Benfratello – Iozzi – Valbonesi Technology and incentive regulation in the Italian motorways industry

17

Data (4)

XI Riunione Scientifica SIET Università degli Studi di Napoli "Federico II“ Napoli, 3-5 ottobre 2007 Benfratello – Iozzi – Valbonesi Technology and incentive regulation in the Italian motorways industry

18

Data (5)

XI Riunione Scientifica SIET Università degli Studi di Napoli "Federico II“ Napoli, 3-5 ottobre 2007 Benfratello – Iozzi – Valbonesi Technology and incentive regulation in the Italian motorways industry

19

Model and econometric results (1)

Total cost function (translog, three inputs)

XI Riunione Scientifica SIET Università degli Studi di Napoli "Federico II“ Napoli, 3-5 ottobre 2007 Benfratello – Iozzi – Valbonesi Technology and incentive regulation in the Italian motorways industry

20

Model and econometric results

Estimation method: Seemingly Unrelated

Regressions (cost function and two cost shares)

Some specifications include concessionaires’

effects in the cost function to capture individual heterogeneity

Cobb-Douglas models as a robustness check

XI Riunione Scientifica SIET Università degli Studi di Napoli "Federico II“ Napoli, 3-5 ottobre 2007 Benfratello – Iozzi – Valbonesi Technology and incentive regulation in the Italian motorways industry

21

Technological measures (1)

y d

ε ε 1 =

Three crucial measures:

• 1. Density economies

Measures the inverse of the % increase in total cost due to a % increase in

• utput, holding the network length

fixed. A value above 1 shows increasing returns to density, so that an increase in

• utput induces a less-than-

proportional increase in total costs.

XI Riunione Scientifica SIET Università degli Studi di Napoli "Federico II“ Napoli, 3-5 ottobre 2007 Benfratello – Iozzi – Valbonesi Technology and incentive regulation in the Italian motorways industry

22

Technological measures (2)

n y s

ε ε ε + = 1

Three crucial measures:

• 2. Scale economies

Inverse of the % increase in total cost due to a % increase in output and in the network length. A value above 1 indicates increasing returns to scale, so that an increase in output and network length induces a less-than-proportional increase in total costs.

XI Riunione Scientifica SIET Università degli Studi di Napoli "Federico II“ Napoli, 3-5 ottobre 2007 Benfratello – Iozzi – Valbonesi Technology and incentive regulation in the Italian motorways industry

23

Technological measures (3)

t TC

t

∂ ∂ = ln ε

Three crucial measures:.

• 3. Yearly rate of technical

progress

The % increase (or decrease) in total cost due to the passing

• f one more year

XI Riunione Scientifica SIET Università degli Studi di Napoli "Federico II“ Napoli, 3-5 ottobre 2007 Benfratello – Iozzi – Valbonesi Technology and incentive regulation in the Italian motorways industry

24

Econometric results (1)

FE FE FE

XI Riunione Scientifica SIET Università degli Studi di Napoli "Federico II“ Napoli, 3-5 ottobre 2007 Benfratello – Iozzi – Valbonesi Technology and incentive regulation in the Italian motorways industry

25

Econometric results (1)

FE

Coefficients positive, almost identical across the two models and highly significant Technological progress around 0.3-0.5% per year

The coefficients for network length and kms travelled have the expected sign and are highly significant.

Coefficients for kms travelled quite similar across models

Coefficients for network differ quite sensibly: high variability of the estimates due the slowly time variant network variable low confidence in results without FE

individual dummies are jointly significant, highlighting concessionaires' heterogeneity

XI Riunione Scientifica SIET Università degli Studi di Napoli "Federico II“ Napoli, 3-5 ottobre 2007 Benfratello – Iozzi – Valbonesi Technology and incentive regulation in the Italian motorways industry

26

Econometric results (1)

FE

Coefficients almost unchanged, only slightly lower technical progress Regulation dummy not significant Ownership dummy different in sign and magnitude. Without FE,

• mitted variable

bias due to unobserved heterogeneity correlated with regressors, e.g. positive correlation between characteristics of the network and

• wnership

dummies Add institutional variables

XI Riunione Scientifica SIET Università degli Studi di Napoli "Federico II“ Napoli, 3-5 ottobre 2007 Benfratello – Iozzi – Valbonesi Technology and incentive regulation in the Italian motorways industry

27

Econometric results (1)

Add network characteristics. Since almost time invariant, cannot estimate model with FE Number of stonework positevely affect costs network composition does not seem to affect costs

coefficient of the

• wnership dummy

not significant and less pronounced, showing that network characteristics partially explain the apparent cost disadvantage of private firms found in model (4).

XI Riunione Scientifica SIET Università degli Studi di Napoli "Federico II“ Napoli, 3-5 ottobre 2007 Benfratello – Iozzi – Valbonesi Technology and incentive regulation in the Italian motorways industry

28

Econometric results (2)

XI Riunione Scientifica SIET Università degli Studi di Napoli "Federico II“ Napoli, 3-5 ottobre 2007 Benfratello – Iozzi – Valbonesi Technology and incentive regulation in the Italian motorways industry

29

Scale and density elasticities

XI Riunione Scientifica SIET Università degli Studi di Napoli "Federico II“ Napoli, 3-5 ottobre 2007 Benfratello – Iozzi – Valbonesi Technology and incentive regulation in the Italian motorways industry

30

High density economies Significant technical progress Scale economies never exhausted in the sample Our controls

Positive impact of privatisation on productivity No impact of regime change on productivity

Summary of the econometric results

XI Riunione Scientifica SIET Università degli Studi di Napoli "Federico II“ Napoli, 3-5 ottobre 2007 Benfratello – Iozzi – Valbonesi Technology and incentive regulation in the Italian motorways industry

31

• In 2002, on the occasion of the price cap review for ASPI, harsh debate
• n how should bear traffic risk
• Different positions resulting in different ways of setting the X factor
• Other things being equal (and when the firms is already efficient),

if the concessionaire bears the traffic risk

X = technical progress

if the concessionaire is insured against traffic risk

X = technical progress

+ changes in cost due to changes in traffic

• Using our results

technical progress:

0.5% (lowest estimate)

change in costs due to density economies with the observed demand

increase (≈ 3%): 3% - [exp(βy*ln1.03) - 1]*100 = 3% - 1.5% = 1.5%

Some additional issues: X factor and traffic risk

XI Riunione Scientifica SIET Università degli Studi di Napoli "Federico II“ Napoli, 3-5 ottobre 2007 Benfratello – Iozzi – Valbonesi Technology and incentive regulation in the Italian motorways industry

32

Use of “internal” prices for our estimates, with known drawbacks:

Not “real” prices Capture of variability of other regressors

However, external prices are not a good alternative for this

sample

No cross-sectional variability !!

But, most importantly

As recognised in the literature for other regulated firms

(see Atkinson & Halvorsen, IER, 1984), concessionaires may base their decision on unobservable shadow prices reflecting the effect of regulation

Indeed, the quality adjustment in the price cap formula reduces

the actual cost of quality-related investment (e.g. resurfacing)

Some additional issues: prices

XI Riunione Scientifica SIET Università degli Studi di Napoli "Federico II“ Napoli, 3-5 ottobre 2007 Benfratello – Iozzi – Valbonesi Technology and incentive regulation in the Italian motorways industry

33

Final comments

• Study of technological aspects of the motorways industry
• Technical progress
• Existence of economies of scale and density
• Optimal network extension
• etc.
• Possible important effect on “better” regulatory policies
• Setting of the X factor
• Effects on revenues of chenges in traffic levels
• Reform of quality regulation
• etc.
• On more general issues: study of the effect of regime and
• wnership changes

XI Riunione Scientifica SIET Università degli Studi di Napoli "Federico II“ Napoli, 3-5 ottobre 2007 Benfratello – Iozzi – Valbonesi Technology and incentive regulation in the Italian motorways industry

34

Price cap mechanism (2)

In principle, the “correct” setting of X requires (at least):

actual cost determination demand forecast assessment of industry-wide (expected)

technological progress

assessment of firm-specific (expected) efficiency

gains, possibly through benchmarking

XI Riunione Scientifica SIET Università degli Studi di Napoli "Federico II“ Napoli, 3-5 ottobre 2007 Benfratello – Iozzi – Valbonesi Technology and incentive regulation in the Italian motorways industry

35

Price cap mechanism (3)

In practice,

X factor set to the same

value for a large set of firms

For other firms, different

values chosen according to specific circumstances:

new investments systematic previous

unbalance between cost and revenues

1,30

• 9,82

0,77 (7) 15 2004 1,40

• 9,82

0,83 (8) 17 2003

• 10,0

0,90 (8) 18 2002

• 10,0

0,98 (6) 16 2001 1,10

• 15,86

1,09 (5) 15 2000 Max Min Mode #. obs. X Year

XI Riunione Scientifica SIET Università degli Studi di Napoli "Federico II“ Napoli, 3-5 ottobre 2007 Benfratello – Iozzi – Valbonesi Technology and incentive regulation in the Italian motorways industry

36

Quality regulation (1)

Theory suggest that when there is a quality

correction term in the PC:

Composite quality index should reasonably

reflect service quality (as perceived by consumers)

Price changes should be based on impact

• f quality changes on consumers’ welfare

XI Riunione Scientifica SIET Università degli Studi di Napoli "Federico II“ Napoli, 3-5 ottobre 2007 Benfratello – Iozzi – Valbonesi Technology and incentive regulation in the Italian motorways industry

37

Quality regulation (2)

Practical effects Quality index Q based

• n:

accidents roughness of road

surface

Choice lead by data

availability

As for β …

1,61

• 0,97

0,45 16 2004 1,15

• 0,79

0,44 16 2003 1,19

• 0,93

0,31 18 2002 0,90

• 0.98

0,01 15 2001 1,39

• 0,22

0,38 13 2000

Max Min Mean # obs.

β Δq Anno