ECONOMIC ANALYSIS OF THE DEREGULATION
- THROUGH
OF COACH SERVICES IN SWEDEN M O D E L S I M U L A T I O N S O F H I S T O R I C & HYPOTHETICAL
COMPETITIVE S I T U A T I O N S
Kjell lansson
Swedish Institute for Transport and Communications Analysis
P.O. Box 3118, S-103 62 Stockholm, Sweden
Inge Vierth
Swedish Transport and Communications Research Board
P.O. Box 5706, S- 114 87 Stockholm, Sweden
John McDaniel
Swedish Institute for Transport and Communications Analysis
P.O. Box 3118, S-103 62 Stockholm, Sweden
IINTRODUCTION
During the last five years a number of new policy measures have influenced Swedish
passenger transport. In 1989 the Swedish national state-owned railway operator, SJ,
was separated into a national state-owned, welfare oriented, track authority
(Banverket, BV) and a state-owned, commercially oriented, operator SJ. The Swedish
air industry was fully deregulated on January 1st 1992. On 1st of January 1993 the
Swedish coach industry was partially deregulated.
These policy changes led to increased competition between long-distance public
transport modes. However, the Swedish Railways, SJ, still has a monopoly for longdistance rail passenger transport.
Over the last two years a Swedish Parliamentary Transport and Communications
Committee has been working on a new transport policy. On behalf of this committee
the Swedish Institute for Transport and Communications Analysis, S K A , has carried
out a number of studies. One of these studies concerned an assessment of the partial
deregulation of coach services 1st of January 1993. The report was presented in
March 1997. In May 1997 the Ministry of Transport asked SIKA to analyse the
possible effects of full deregulation, an ongoing work which will be reported in
September 1997.
Section 2 describes the general prerequisites and assumptions. Section 3 includes a
summary of the report presented to the Transport and Communications Committee
concerning partial deregulation. In section 4 we present the prerequisites for the
ongoing study on behalf of the Ministry of Transport.
181
2 P R E R E Q U I S I T E S AND ASSUMPTIONS
2.1 Basic Analytical Method
Basically we regard the passenger transports as a system, where passengers can
choose among lines and operators and where a single journey may involve several
lines and operators. The method used is simulation, employing a computerised routenetwork-analysis package, Vips. Some properties of this package are described in
section 2.4 below.
In this analysis we assume that:
• Operators compete independently, without regarding modes as being
complementary.
• Passengers regard the modes as alternatives, but also as being complementary.
• Passengers are not homogenous, but have different preferences and face different
prices.
For each policy change, historic or hypothetical, we calculate:
*
*
•
•
•
•
•
Demand for each public transport mode and service
Consumer surplus nationally
Consumer surplus by region (county or town level)
Revenues and costs for each operator
External effects and taxes
State finances and excess burden
Total welfare
Travel standard is expressed in terms of generalised cost per journey, that is price plus
weighted travel time converted into monetary units.
Whereas the modal split between pubhc transport modes is calculated within the
Vips-system, we use an elasticity model for the split between public transport and car.
The elasticity with respect to price is assumed to be -0.8, and the elasticity with
respect to generalised cost is proportional, that is -0.8 times generalised cost over
price.
The Swedish Transport and Communications Committee has proposed to increase the
price of petrol by SEK 0.1 (appr. 1 penny) per year during 23 years. It is assumed that
tectmological improvements and possible regulations will decrease the fuel
consumption per kilometre, in which case we assume that the price of car travelling
remains unchanged.
2.2 User Categories and Time Values
We take into account that various passenger categories have separate valuations of
price, travel time and comfort. These differences are reflected in separate values of
time per category. We use one set of time values for the overall national analysis and
a slightly different set for the analysis o f rail/coach competition between the two cities
of Karlstad and Gothenburg, where the travel distance is below average. In the rabies
below the weights for ride (in-vehicle) time, wait time, transfer time and w a l k time are
related to ride time for InterCity train, normalised to unity (one). The values o f time
and the weights w h i c h reflect the perceived convenience o f various categories for each
mode are derived from stated-preference studies carried out on b e h a l f o f SIKA.
Values and weights are summarised in the table below. Within parenthesis are given
the slightly modified values which are used in the ongoing study on b e h a l f o f the
Ministry. The reason for the modified values is that improved information on
passenger loads from the state railways, SJ, and the biggest coach operator Swebus,
w h i c h meant that the calibration showed that the values o f waiting tinge should be
slightly increased to m a k e model calculated loads more in line with the operators'
load figures.
Weight for
Ride time weight
Shar Time value
wait time transfer- ICX2000 coach air
e
ride
time
and
train
pax (train)
walk time
SEK/hour
Private journeys
Working,
high value
Working,
low value
Students
Pensioners
0.20
105
0.4 (0.7)
2.0
1.0
0.9
1.50
1.1
0.50
80
0.4 (0.7)
2.0
1.0
0.9
1.50
1.1
0.15
0.I5
30
30
0.2 (0.4)
0.2 (0.4)
2.0
2.0
1.0
1.0
0.9
0.9
1.50
1.50
1.1
1.1
<100 kmjourneys
(35)
1 . 0 0 140
(0.4)
(2.0)
(1.0)
(0.9)
(1.50)
1.0 (1.1)
2.0
1.0
0.9
1.50
Business journeys
1.1-2.5
Table 1: Time values and weights for the national network, in work for the Transport and
Communications Committee(and for the Ministry of Transport).
Share
pax
Weight for
Time value
wait time
ride
time
(tram)
transfer- and
walk time
Ride time weight
train
coach
SEK/hour
Private journeys
Working,
low value
Students
Pensioners
0.54
60
0.30
2.0
1.0
1.50
0.32
0.14
22
22
0.13
0.13
2.0
2.0
i.O
1.0
1.50
1.50
Business journeys
1.00
140
1.0
2.0
1.0
1.50
Table 2: Time values and weights for the Karlstad-Gothenburg study, in work for the
Transport and Communications Committee.
2.3 Responses and costs with respect to demand changes
Since various measures would cause increased or reduced demand for certain lines
and operators, the operators would have to respond in one way or the other.
Here we assume that train supply can always be modified by the introduction/removal
o f carriages per train. The reason is that Swedish trains seldom use the m a x i m u m
number of carriages permitted.
For air we assume that the response would be either change in aircraft size or change
of service frequency, where service frequency change is only applied when demand
changes by more than 10 per cent. Change in frequency is however used only for the
calculation of user benefits. With regard to costs, we assume that air and coach costs
change in proportion to 70% of the demand change.
2.4 Some Properties in VIPS
Importance of the use of time table
In Vips it is possible to use the basic assumption that travellers do not use a time table
when planning their trips, which is suitable for route network analysis o f frequent
urban services. However, VIPS is also capable of utilising the assumption that
travellers can use the time table. The impact of recognising this behaviour will result
in radically different results when compared to the assumption that a time table is not
used:
• Travellers are likely to choose between alternative places of departure (bus stops,
stations, airports) and mode specific services when using a time table, whereas
only one alternative is otherwise the case. Travellers then also consider travel time
and price of the journey in their choice o f mode and route.
Specific standard o f stop or service
Different points o f departures and services display varying comfort. Therefore an
option exists to adjust the weights for waiting times and riding times accordingly. For
example, a comfortable service can be given a lower than average travel time weight.
Fare analyses
Vips allows for analysis of service specific fares. Either the fare is a fimction of the
distance travelled on each route, or the fare is specified on a station-station basis. Both
fare structures affect the choice of route and the generalised cost. The result of a route
network analysis is total revenues, for the different modes and for the individual
services. Both the financial results of the operators and the social net benefit are
obtained.
Specified transfer times
Normally transfer times are calculated as half the headway. However, where intervals
between departures are long this may cause unrealistically long waiting time costs,
184
especially since waiting time typically has a high value of time related to in-vehicle
time. Vips allows specification of transfer times between all pairs of routes at each
stop. Each route pair m a y thus give rise to 8 different specified transfer times, since
the two directions of each route must be taken into account.
2.5 U n c e r t a i n t i e s
Note that all results presented are based on calculations which have some degree o f
uncertainty..The main cause of uncertainty is the travel matrix, which is a hybrid o f a
synthetic matrix and a matrix based on two national surveys. Another cause of
uncertainty is the insufficient information provided by the coach operators on actual
passenger loads, which are used for calibration of the model. Swebus and the state
railways, S J, have though provided good information.
3 PARTIAL COACH DEREGULATION
On the 1st of January 1993 the Swedish coach industry was partially deregulated. This
means that the "burden-of-proof' concerning "damage of the railway" has been
reversed. SJ must now in principle prove that a new coach line will seriously harm the
profits o f the railway operation. Due to the deregulation the supply in terms of coach
kms has increased by 39% between 1992 and 1996.
In section 3.1 we present our calculations o f the impact o f coach deregulation on
national level, between the years 1992 and 1996. We do this by comparing the
performance of the air, rail and coach system in 1996 with the air system in 1996, the
rail system in 1996 and the coach system in 1992. In 3.2 we present a case study for
the corridor Gothenburg-Karlstad. In 3.3 we analyse how intemalisation of external
effects m a y influence the welfare impacts of coach deregulation.
3.1 C o a c h d e r e g u l a t i o n 1992-1996 - n a t i o n a l level
M o d a l split
According to the model analysis, the choice of mode seems to have changed very
little, ceteris paribus, following the partial deregulation of the coach industry.
Coach 1992;
rail, air Share Air, rail, coach 1996
1996
Million. pass. Kin
X2000
577
Night Ixain 1105
IC-train
3669
SAS
1992
Other air
887
Coach
170
Total
8400
%
Million.pass. km
Share
%
6,9
13,2
43,7
23,7
10,6
2,0
100,0
573
1103
3605
1987
881
278
8429
6,8
13,1
42,8
23,6
10,5
3,3
100,0
Table 3: Calculated modal split (in passenger kms and shares), due to the development of
coach operation
Consumers
The dominating effect is that low-value of time (low-income) travellers have gained,
since the gain in price is larger than the loss in terms of travel time. Note that a
positive sign indicates improvement.
Business
140 SEK/h
Generalised cost/trip, SEK 0,5
Generalised cost/trip, %
0,0%
Consumer surplus,mkr
3,5
of which fare
3,7
of which time
-0,2
Private
Private
Private
Sum
105 SEK/h
1,8
0,1%
5,5
4,7
0,8
80 SEK/h
2,3
0,2%
17,8
18,9
-1,0
30 SEK/h
2,6
0,5%
12,2
22,1
-9,9
39,0
49,4
-10,4
Table 4: Calculated change in generalised cost and consumer surplus per category, due to
the development of coach operation.
Welfare
The net benefit, change in consumer and producer surplus, is close to zero, which
means that the dominating effect is redistributional.
3.2 C o a c h Deregulation Karlstad - G o t h e n b u r g
Background
From the 1950"s until September 1995, SJ and the SJ-owned coach operator Swebus
were, under the old regulations, the only operators between the cities o f Karlstad and
Gothenburg - a corridor 250 k m long.
Under the new regulations a new coach operator entered the market - Saffle Rese
Service (SRS). They charged a price that was about 40% below Swebus, which in turn
caused Swebus to respond by setting a price slightly below SRS's.
In April 1995 SJ responded by decreasing its fare by 40%.
The analysis comprises these situations as well as three hypothetical ones. The first
hypothetical case is that SJ is assumed to reduce the number of departures while the
coach operators increase their departures. The second situation is that SJ is assumed to
close down their operation completely. The third situation is that the coach operation
is assumed to cease entirely, while SJ is assumed to maintain their low fare in order to
keep out competition.
The table below summarises the cases, indicating number of departures and prices for
three categories o f private travellers. Note that the price of 1st class tickets are
assumed to be unchanged, which they have been throughout the "price war".
186
2nd class Fare Gothenburg-Karlstad,
SEK
round trips/ Working Students
Pensioners
weekday
Situations
Actual cases
6 trains, 3 coaches
(till 9/95)
SJ
Swebus
6
3
208
185
143
95
143
95
6 trains, 5 low-price coaches
(from 9/95)
SJ
Swebus
SRS
6
3
2
208
95
100
143
76
80
143
76
80
6 low-price trains, 5 low-price coaches
(as present, from 4/96)
SJ
Swebus
SRS
6
3
2
135
95
100
95
76
80
135
76
80
4 low-price trains, 8 low-price coaches
SJ
Swebus
SRS
4
5
3
135
95
100
95
76
80
135
76
80
only 20 low-price coaches
Swebus
SRS
10
10
95
100
76
80
76
80
only 6 low-price trains
SJ
6
135
95
135
Hypothetical cases
Table 5: Situations for Gothenburg- Karlstad
Welfare Results
The analysis shows that the worst situation would be that S$ decides to cease
operation, leaving all operations to the coach operators. The reason is that the users,
especially business travellers and h i g h value o f time private travellers, would lose the
option o f taking the train which is faster and has a lower value o f time than coach.
The best situation seems to be where SJ is the only operator. Note, however, that this
situation is hardly sustainable. I f SJ were left alone it would be unlikely that they
w o u l d keep the low price level, since the higher price level yields a higher profit.
...compared with the present Consumers
situation
surplus
Producers
surplus
External
effects
Taxes,
excess
burden
Sum
6 trains, 3 coaches
-8,2
6 trains, 5 low-price coaches
-5,3
4 low-price Wains, 8 low-price-9,9
coaches
only 20 low-price coaches
-20,4
only 6 low-price trains
-2,1
5,4
2,1
3,9
0,3
-0,9
1,3
1,1
1,4
-0,4
-1,4
-2,7
-5,1
2,5
5,2
-0,3
1,4
1,6
-0,5
-16,6
4,0
Table 6: Calculation of welfare, in million SEK per year
3.3 I n t e r n a l i s a t i o n o f E x t e r n a l E f f e c t s and Rail Subsidies
Since one of the proposals put forward b y the Parliamentary Transport and
Communications committee was intemalisation o f external effects, meaning price
increases for all modes, especially for car, coach and air, one important aspect is
whether such intemalisation may affect the welfare effects of coach deregulation.
Another issue is whether subsidisation of rail services may affect the impacts of coach
deregulation.
Introduction
SIKA analysed:
• Intemalisation of external effects through taxation on a vehicle kilometre basis.
• Reduced prices for 2nd class train journeys, via subsidies.
• Combination o f intemalisation and price decreases of 2nd class train joumeys.
With respect to intemalisation, it is assumed that tax revenues are used for reducing
other taxes, something which in turn reduces excess burden.
SIKA analysed intemalisation based on its recommendations on values of exhaust
gases and accidents, but also based on three levels of carbondioxide valuation, one of
which is presented here: CO 2 is valued at SEK0.83 /kg, which corresponds to the
Commissions aim for 2010, and which gives rise to the following tax increases.
InternaHsafion 0.83
External
cost
SEK/veh.km
Train
Air
Coach
Car
5,60
11,00
1,92
0,31
Tax
1996
SEK/veh.km
3,80
4,20
0,78
0,35
Tax
increase
SEK/veh.km
1,80
6,80
1,14
0,04
Table 7: Internalisation at the level SElf 0,83/kg C02
MaTg'mal costs for rail ope~af~on ate xela~,~e~ to~ 6ue to ~ g e economies of scale,
especially where it is possible to accommodate demand by more carriages per train.
The price of train tickets are today around 50% over the marginal cost. For air and
coach the (discount) prices seem to be fairly equivalent to the marginal cost per
passenger of a larger plane and a marginal coach departure respectively.
There is thus an argument for subsidising train tickets but not air and coach tickets.
Assuming that the government has a budget constraint, we analyse subsidies related
to price decreases only for the more price sensitive 2nd class tickets on InterCity
trains.
Results of internalisation: Modal split
The analysis indicates that the main effect on modal choice of internalisation would
be a demand shift from air to train. The reason why SAS would lose more than other
air operators is that SAS operate longer distances for which the tax and price increases
are higher.
Base 1996
Pass.kms
Result of internalisation 0.83
Pass.km
Share
Share
million
%
million
X2000
719
7,2
787
Night train
1135
11,3
1208
IC-train
4280
42,8
4288
SAS
1916
19,1
1597
Other air
1038
10,4
1010
Coach
391
3,9
363
Regional bus
529
5,3
504
TotaIpublic transport 10006
100,0
9757
Car
16860
62,8
17034
Table 8: Calculated modal shares, in million passenger kilometres
%
8,1
12,4
43,9
i6,4
10,3
3,7
5,2
i00,0
63,6
Cotlsulners
According to the table below the less in terms of consumer surplus, mainly in terms of
price, is quite substantial.
Business
Private
Private
Private
140 SEK/h 105 SEK/h 80 SEK/h 30 SEK/h
General. cost, SEK/trip -21,7
-13,7
-14,3
-12,4
General.. cost, %/trip
-1,5
-1,0
-I,3
-2,3
Cons. Surplus, mill. SEK -159
-43
-111
-57
of which fare
-178
-29
-69
-50
of which time
19
-14
-43
-8
Table 9: Calculated changes in generalised cost and consumer surplus
Trips<100 km
Sum
35 SEK/h
-2,2
-1,5
-74
-70
-4
-444
-395
-49
Welfare
The net economic gain is quite substantial. The main benefit stems from state fmances
and hence reducing the excess burden.
Consumer surplus
(of winch fare)
(of winch time)
Producer surplus
Net state finances
Excess burden
External effects
Sum
SEK, million per year
.444
-395
-49
68
762
229
43
656
Table 10: Calculated change in welfare
The results thus indicate that intemalisation of extemal effects is socially beneficial
without regarding that such intemalisation is assumed to be beneficial for its o w n
sake. The interpretation is that the current transport system is too large. Taxation
implies reduction of expensive air services, while at the same time it is cheap to
increase supply of rail services by adding the number of carriages. Another conclusion
is that rail becomes more competitive and less vulnerable by coach deregulation.
Results o f subsidies and price decreases on 2nd class InterCify trains, given
internalisation: Modal split
The combination internahsafion and rail subsidies indicate that the fast train X2000
would also gain some demand, while the loss for air is quite substantial
Subsidies and price decreases on 2nd
class IC-trains, given
internalisation 0.83
Pass.km
Share
Base 1996
Pass.kin
million
X2000
719
Night train
1135
IC-train
4280
SAS
1916
Other air
1038
Coach
391
Regional bus
529
Total public transport 10006
Car
16860
Table 11: Calculated modal shares,
Share
million
%
7,2
769
11,3
1138
42,8
4741
19,1
1557
10,4
985
3,9
311
5,3
508
100,0
10011
62,8
16857
in million passenger kilometres
%
7,7
11,4
47,4
15,6
9,8
3,15,1
100,0
62,7
Consumers: national level
The users (in aggregate terms) are more or less unaffected. The main effect is that
business travellers would lose (no price reductions assumed), while private travellers
would benefit.
Business
Private
Private
Private
140 SEK/h 105 SEK/h 80 SEK/h 30 SEK/h
General cost, SEK/trip -21,7
7,5
8,5
7,3
General. cost, %/trip
-1,5
0,6
0,8
1,4
Cons. surplus,mill. SEK -159
24
68
34
of which fare
-178
56
151
43
of whichtime
19
-32
-84
-8
Table 12: Calculated changes in generalised cost and consumer surplus
Trips<100 kin
Sum
35 SEK/h
0,1
0,1
4
30
-26
-29
102
-132
Consumers: Regional level
In figure 1 we also present the results on a regional level in terms of generahsed cost
(expressed in minutes) for private travellers with value of fime equal to SEK 80/hour.
The pattern is more or less the same for the other groups except business travellers
who will not benefit from the price decreases.
As a matter of course, an imphcit tax shift through intemahsation would disbenefit
travellers in all parts of Sweden and price decreases would benefit travellers in all
parts of Sweden. The combination would mainly disbenefit travellers to and from
northern Sweden and the very south of Sweden, while the majority of areas would
benefit, as can be seen in table 13 below.
Welfare
The main welfare effects now stem from the improved state finances mad hence
reducing the excess burden. The welfare improvement is calculated to be some SEK
190
80 million less if price reductions would only comprise the commercially non-viable
train lines.
Consumersurplus
(of which fare)
(of whichtime)
Producer surplus
Net state finances
Excessburden
External effects
Sum
Table 13: Calculated change in welfare
SEK, million per year
-29
102
- 132
59
529
159
85
802
3.4 Conclusions concerning the partial deregulation
• The development of coach service is neutral from an aggregate welfare point of
view, but low income (low value of time) private travellers have gained, while SJ
has lost a little revenue.
• Coach competition may in certain cases imply that SJ would cease to operate lines
which become non-viable from a profit point of view even if they are motivated
from a welfare point of view.
• An implicit tax shift through internalisation of external effects is socialty
beneficial. The reason is that the taxation reduces the total amount of resources to
the transport sector, whereby the resource saving outweighs the loss for the users.
• Intemalisation of external effects also mitigate the problems that coach
deregulation may cause the railway.
• Subsidies aimed at price decreases are welfare improving. The reason is that the
excess burden loss and the small extra cost of train carriages are smaller than the
calculated resource saving for air transport.
On the basis of the studies carried out by SIKA the parliamentary transport and
communications committee has proposed, for example:
To raise the price of petrol and diesel by SEK 0.1 per year during the next 23 years.
• To internalise the external effects for air- rail- and coach transport according to the
values of carbondioxide which stem from the taxation of petrol.
• To subsidise commercially non-viable train services in a competitive tendering
process, which also aims to reduce the prices on 2nd class InterCity trains.
•
4 ANALYSIS OF FULL DEREGULATION
4.1 Background
In May 1997 the Ministry of Transport engaged SIKA for a further analysis of
extended coach deregulation. The task in short is:
• To analyse the impacts of full deregulation.
• To compare the welfare impacts of full deregulation with those of a integrated, coordinated planned coach, rail and air network.
4.2 Assumptions
Intemalisation o f extemal effects at the level SEK 0.83/kg carbondioxide shall be
assumed.
The results shall be presented in terms of:
• Producer surplus for air, rail, coach, local and regional operators.
• Consumers surplus, firstly assuming separate values of time for each passenger
group, secondly assuming (for the welfare calculation, but not for the behavioural
calculations) one common value for all groups.
• External effects
• State finances and excess burden
• Net social benefit, according to the two sets of values of time.
Figure 1: Change in generalised cost on city level for private travellers with
VoT= SEK 80/hour. (I cm = 130 minutes)
Legend: Blue/dark = increased generalised cost Yellow/light = decreased generalised cost
193