Road investment benefits over and above transport cost savings and gains to generated traffic; Do they exist? : Some extensions of the perfect market economy model of VTI meddelande 616, 1990

VTInotat

Hummer: T 107 Datmm: 1991-04-16

Titel: ROAD INVESTMENT BENEFITS OVER AND ABOVE TRANSPORT COST

SAVINGS AND GAINS TO GENERATED TRAFFIC; DO THEY EXIST? - some extensions of the perfect market economy model

of VTI meddelande 616, 1990

Forfattare: Imdad Hussain

Avdelning: Trafikavdelningen

Projektnummer: 73012-7

Projektnamn: Indirekta effekter av véginvesteringar i

inriktningsplanering och objektanalys Uppdragsgivare: VTI

Distribution: fri/nyférvérv/begrénsad/

(Lb)

V5 ml) Tram

Statens vag- och traflkmstltut

Instltutgt Besok. O/aus Magnus vag 32 Lmkopmg

C O N T E N T S

INTRODUCTION

EXTENSIONS OF THE PERFECT MARKET ECONOMY MODEL -the freight transport model

The case of commodity tax

The case of wage tax on employers The case of tax on intermediate goods The case of increasing returns to scale THE MODEL WITH AN ALLOWANCE FOR PASSENGER TRAFFIC - The case of commuting

Model with 25 per cent income tax Model with 50 per cent income tax SUMMARY AND CONCLUSIONS

Appendices 11 13 15 18 20 21 24

INTRODUCTION

Practical road appraisal manuals mostly use computer based

cost-benefit models (see, for example, OBJEKTANALYS, and COBA, the

CBA models used by The National Road Administration boards in

Sweden and Great Britain respectively). "Benefits to existing

traffic" completely dominate the benefit side of such manuals.

According to these models, the traffic benefits of a road

improvement are usually measured in terms of:

° savings in accident costs

° savings intravel time

° savings in vehicle operating costs ° increased travel comfort

Historically, a close relationship between land use activities

and transportation system improvements has been perceived. The

intuitive ideas about the significance of transport

infra-structure for the development of national economies existed

already during the days of Adam Smith (1776). His famous statement: "the division of labour must always be limited by the

extent of market ...", seems to reflect his belief that

improvements in transport facitilies, by opening a more

extensive market for every sort of industry, would affect

industrial growth and specialization. The mechanisms determining

the rules of specialization and the distribution of income are

implicit in the classical theory of firm and household location

as well as in the comparative cost theory of international trade (Straszheim, 1972).

In the post world war II literature, the role of infrastructural

road investments in the growth of industries andtheir impact

upon the development of depressed regions has been a topic of continuous interest.

During the last three decades, the bulk of literature in this

area of research has grown tremendously and engineers,

economists, planners, geographers and operations researchers

have all contributed to the work. Due to the cross-disciplinary

nature of the subject, each discipline has considered the

complexities involved in the matter from slightly different

perspectives. Examples of work in this area are Tinbergen

(1957), B08 and Koyck (1961), Friedlaender (1965), Brown and

Harral (1965), Wilson et a1 (1966), Mohring and Williamson

(1969), Cleary and Thomas (1973), Dickey and Sharpe (1974), Dodgson (1974), Peaker (1976), Thornton (1978), Sharp (1980),

Friedlaender and Spady (1981), Liew and Liew (1984), Dodgson

(1984), Mackie and Simon (1986), Quarmby (1989), Aschaur (1989), Anderstig and Mattson (1989), Peterson (1989), Binder (1989), Forkenbrock and Foster (1990) and Hulthén (1991). On the basis of conclusions drawn from the previous research in this area, it

may be learnt that the thesis that infrastructural road

investments might play a major role in the economic development

of regions with sufficient unexploited resources in the form of

land, labour and capital, is agreed upon in general. Whether the

same can be true in the case of mature and industrially

developed economies with full employment of resources, is,

however, a question which still lacks a comprehensive and clear

answer .

The practical way of measuring road investment benefits in terms of transport cost savings to existing traffic is regarded as a

reasonable attempt to approximate the partial equilibrium

benefit measure which in fact should also include gains to newly generated traffic. In order to explicate the question of whether road investments really give rise to substantial benefits beyond

the transport cost savings and gains to generated traffic, a

general equilibrium approach may be used. In this type of

analysis, benefits may be measured by studying the variations in

the structure of production and consumption in an economy. Such

variations are brought about by changes in relative prices of

the goods with a direct impact on consumer welfare.

In an earlier VTI report, (see Hussain and Jansson, 1990), it

has been argued with the aid of a general equilibrium model,

that in an economy with perfect markets, a partial equilibrium analysis of the benefits will almost exactly be equivalent of the general equilibrium.measure of the benefits.

The purpose of this paper is to examine whether a departure from

the perfect market model assumption, by letting certain market

imperfections into the model, will considerably change the

results and conclusions of the earlier analysis. The paper

consists of two parts. The first part of the paper treats the

freight transport model with no passenger traffic. In the second

part the model with interregional commuting possibility is

introduced. Finally, the results of the models are discussed and summarized.

EXTENSIONS OF THE PERFECT.MmRKET ECONOMI.MODEL _ the freight transport model

The partial equilibrium benefit measure discussed in the perfect

market model is represented by a change in the area to the left

of the transportation demand curve, wheras the general

equilibrium measure is computed in terms of the equivalent

variation net of the value of leisure lost due to increased

production.

The already established equivalence between the two benefit

measures is due to the absence of any price distorting mechanism in the model economy. The market prices, in this situation, also

reflect the social value of production factors used and

commodities produced, thus guaranteeing the equality of the

private and social evaluation of the economys output. However,

the real world situation does not comply to these conditions and

the economies are subject to several types of price distorting

market imperfections which affect private production and

consumption decisions. Therefore, the equivalence between the

private and the social evaluation may break down. There may be many factors contributing to this, but the presence of commodity

and other taxes and non-convex production technologies are

common examples (see Lundgren, 1985). In such cases, prices do

not function as effective signals to buyers and sellers and,

therefore, the resource allocation is disturbed.

The purpose of this chapter is to test the validity of the

earlier conclusions drawn about the nature andsize of transport

investment benefits under different conditions characterizing a

market failure. The analysis is carried out, firstly by

incorporating different types of taxes in the basic model.

Secondly, increasing returns to scale technology is introduced

in the model. In an attempt to extend the model further, the

possibility of commuting between region 1 and region 3 is

allowed for.

The case of commodity tax

As a departure from the perfect market economy model, a 25 per

cent flat rate commodity tax on consumers in all regions is

introduced. An important assumption is that unit transport

costs for the transportation of a consumer good between two

regions is paid in terms of a certain amount of the same good.

Thus, it is not only the goods actually consumed which are

taxed, but also the goods which are bought to pay for the

transports, as the buyer always stands for the transport bill.

The entire amount of tax revenues collected in a region is

transferred back as a subsidy to the consumers of the same

region. Moreover, the commodity tax also affects the labour

supply since this is assumed to be a function of the real wage

which is computed as a ratio of money wages to general consumer price levels in each region.

The impact on resource allocation in the economy, of a 25 per

cent commodity tax, ceteris paribus, may be reflected by

com-paring ordinal utility indices of the various regions, Computed

in the initial equilibrium situations for the perfect market

economy as well as the commodity tax models. Such a comparison

Table 1: Changes in the ordinal utility indices

The Perfect The Tax

Region Market Model Model AU

1 63.92 57.51 -6.41

2 56.25 51.30 -4.95

3 57.16 52.15 -5.01

AU = Change in utility indices due to a 25 percent commodity tax on consumers

The figures given in the table above clearly indicate an adverse

change in the levels of consumption for all regions despite the

fact that the entire amount of tax revenues was transferred back

to the consumers. Nevertheless, this is an expected result and

there is nothing sensational in that.

With this in hand, it is high time to make an inquiry into

whether such disturbances in the resource allocation due to the

commodity tax also has implications for the relationship between

partial and general equilibrium.measures of benefits due to

transport improvements. In this case, a benefit measurement

analysis, similar to that conducted in the perfect market case,

is carried out for different levels of reduction in unit

transport costs.

A summary of the results from the benefit measurement

calcula-tions for the perfect market model as well as for the commodity

tax model, is presented in tables 2 and 3, respectively, for

comparison. The numerical computation lists of the model are

provided in appendix one.

An explanation to the variable notation in the table columns follows below:

At = percentage change in unit transportation cost

TS = transport cost savings for existing traffic

PE = partial equilibrium benefit measure

EV = Equivalent variation (general equilibrium benefit

measure)

NEV = EV-DUL: total net welfare gains

DUL = monetary value of leisure forgone

ANI = change in national income at constant prices

% = An indication of the percentage ratio of the

variable to ANI

The results shown in table 2 have been published earlier (see

Hussain and Jansson, 1990), with the minor exception for columns 8 to 11. These columns show the relative size of the different

benefit measures in relation to the changes in national income

measured at preinvestment prices. For example, the figures given in column 10 indicate that greater reductions in transport costs

lead to a slight decline in the relative size of NEV. This may

be a reflection of the assumption that the derivative of the

marginal utility of money is negative. Similarly, a slight

decline in the relative sizes of DUL, TS and PE is noted. The

main conclusion of the perfect market model, as previously

mentioned, is that transport cost savings to existing traffic

and gains to generated traffic almost exactly represent the

general equilibrium benefit measure, NEV. Moreover, the

con-clusion holds well in case of small as well as large reductions in unit transport costs.

A comparison between columns three and six shows that the

partial benefit measure of transport cost savings for existing

traffic and gains to newly generated traffic, now regularly

underestimates the net equivalent variation. The degree of

cost reductions. There is a clear indication that the pure

traffic generation is rather negligible. Transport cost savings

underestimate the net welfare gains by 2.6 per cent for a 15 per

cent reduction in unit transport costs. This gap becomes larger

for greater reductions in transport costs and is about 4 per

cent for a 50 per cent transport cost reduction.

The relative size of net equivalent variation with respect to

national income change is greater compared to the perfect market model case. This may be explained by a smaller relative share of

DUL in EV, in the case of the commodity tax model. The relative

sizes of TS and PE, in this case, are lower compared to that of

NEV and decline for larger transport cost reductions. However

NEV and ANI seem to have a stable relationship to each other.

Ta bl e 2: Re sul ts of th e pe rf ec t ma rk et mo de l 10 11 TS PE EV DU L N E V TS % PE % NE V% DU L% 15 % 44 .5 44 .8 56 11 45 55 81 81 .5 81 .8 20 25 % 74 74.8 93 18 75 93 79 .6 80 .4 80 .7 19 .4 35 % 10 3. 5 105 13 0 25 10 5 13 0. 4 79 .4 80 .5 80 .5 19 .2 50 % 14 7. 5 151. 8 18 7 36 15 1 189 78 80 .3 79.9 19 60 %

VTI Notat T 107 Ta bl e 3: Re sul ts of th e' co mm od it y ta x mo de l 10 11 -A t TS PE EV DU L N E V TS% PE % NE V% DU L% 15 % 50 .1 50 .4 60 8. 6 51 .4 60 83 .5 84 85 .7 14.3 25 % 83 .4 84 .3 10 0 14 .5 85 .5 10 0 83 .4 84 .3 85 .5 14 .5 35 % 11 6. 5 11 8. 2 14 0 20 .1 11 9. 9 14 1 82 .6 83 .8 85 14 .2 50 % 16 6. 1 16 9. 6 20 1 28 .4 172. 6 20 3 81 .8 83 .6 85 14 60 %

11

The case of wage tax on employers

A similar comparison is again conducted under the condition of a wage tax on employers. The employers are now supposed to pay 25 per cent of their total wage bill in tax to a hypothetical tax authority. The tax revenues, so collected, are returned to the

consumers as transfer payments according to the assumption of

the model. In this case too, however, the revenues raised in a

region are returned to the consumers of the same region. The tax leads to similar final allocational effects as in the commodity

tax case, however, it causes the marginal cost curves of the

producers to shift upwards. The welfare gains arising from

different levels of transport cost reductions are calculated

according to various benefit measures and are given in table 4. Columns 8 to 11 show thepercentage ratio of TS, PE, NEV and DUL

to ANI, respectively. These ratios are similar to the

corre-sponding ratios in the commodity tax model case.

The figures in the table also show that the discrepancy between

the partial and the general equilibrium.masures of the benefits is slightly greater as compared to the model with commodity tax. This is particularly true in cases of larger reductions in the

unit transport costs. For example, for a 35 per cent reduction

in transport costs, the transport cost savings related to exis-ting traffic and the gains to generated traffic underestimate the net equivalent variation by 1.4 per cent in the comodity tax

model, whereas the corresponding figure for the wage tax model

is 3.4 per cent. Nonetheless, the relative size of the NEV with

respect to ANI remains almost the same according to the two

models. The numerical computations for various levels of

transport cost reductions in this model are given in appendix

VTI Notat T 107 Ta bl e 4: Th e ca se of wa ge ta x on pr od uc er s 10 11 TS PE EV DU L N E V TS % PE% NE V% DU L% 15 % 49 .9 50 .2 60 8. 7 51 .3 60 83 .2 83 .7 85 .5 14 .5 25 % 83 83 .9 10 0 14 .4 85 .6 10 0 83 83 .9 85 .6 14 .4 35 % 11 6. 2 11 7. 9 14 2 20 .1 12 1. 9 14 2 81 .8 83 85 .8 14 .2 50 % 16 5. 6 16 9 20 5 28 .5 176. 5 20 6 80 .4 82 85 .7 13 .7 60 %

13

The case of tax on intermediate goods

In this model, a 25 per cent tax on intermediate goods is

levied. The producers of the consumer goods, thus pay in tax, an

amount equal to 25 per cent of the market value of the raw

materials used in production. As in earlier cases, the tax

revenues are returned to the consumers as transfers. An obvious

impact of this type of tax is to reduce the absolute level of

consumption and production in the economy and its regions. Also,

the level of the benefits arising from a certain reduction in

the unit transport costs is considerably reduced as a result of

the decline in production. However, the comparison between the

partial and the general equilibrium measures of benefits shows a

discrepancy between the two measures which is about the same

size as in the case of the wage tax model. The ratios of TS, PE

and NEV to ANI are about the same size as in the previous tax

model. However, theratio of DUL to ANI is greater in this model

for all levels of reductions in the unit transport costs. The

fact that the leisure becomes more valuable may be due to a

relatively greater increase in the consumer price level compared

to the real wage. Even in this case the relative size of NEV

with respect to ANI is greater than that in the perfect market

model. The numerical computation lists for the model are

VTI Notat T 107 Ta bl e 5: Th e ca se of a ta x on in te rm ed ia te go od s 10 11 TS PE EV DU L N E V TS % PE% NE V% DU L% 15 % 40 .4 40 .6 50 8. 8 41 .2 48 84 .2 84 .6 85 .8 18 .3 25 % 67 .3 68 .0 84 14 .7 69 .3 83 81 81 .9 83 .5 17 .7 35 % 94 .0 95 .4 11 9 20 .7 98 .3 11 8 79 .7 80 .8 83 .3 17 .5 50 % 13 4. 1 13 6. 9 17 2 29 .3 14 2. 7 17 1 78 .4 80 83 .5 17 .1 60 %

14

15

The case of increasing returns to scale

In this case, the falling average total cost for producers, in

relation to the demand faced by them, indicates increasing

returns to scale in production. This is a situation in which

even a competitive market system does not guarantee an efficient level of production. This phenomenon is introduced by

incorpora-ting a total fixed cost factor in the production of consumer

goods.

The existence of the phenomenon may prevent the market from

generating an optimal level of output in two ways. Firstly, the

incentive to exploit economies of scale provides room for

building monopoly. The price, in such a situation, will then be set at a higher level than the marginal cost, leading to a lower level of production. Secondly, even if the price is set equal to

the marginal costs, the falling average total costs imply a

certain operating loss for the firm.

In the particular model used in this study, the marginal costs

for the finm are assumed to be constant. The first case is

characterized by the zero profit condition where the price is

equal to the average total costs. I have chosen to call this the "quasi-monopoly case". In the second case, the price is equal to

the marginal cost which implies a net operating loss to the

firm, which is equal to the total fixed cost. According to the

model this loss is financed by a consumer tax on the good. The

results of the benefit analysis for both cases are given in

tables 6 and 7 below.

This analysis provides several interesting results. By studying

the results of the quasi-monopoly model, it can be learnt that

the partial equilibrium measure of transport improvement

benefits now overestimates the general equilibrium benefit

measure for different levels of transport cost reductions up to

50 per cent. For a 60 per cent reduction in trnasport cost

how-ever, this relation is reversed as NEV exceeds PE by merely 0.4

16

between the partial and the general equilibrium.measures in this case. A 35 per cent reduction in unit transport costs appears to cause the greatest discrepancy (3.4 per.cent). The relative size of NEV with respect to ANI is 82.2 per cent for a transport cost

reduction of 15 per cent. It decreases to 80 per cent for a 25

per cent transport cost reduction and then increases slightly,

to reach 82.2 per cent for a 60 per cent transport cost

reduction.

The case of increasing returns to scale where the price is equal

to the marginal costs leads to various conclusions. In this

case, the PE measure underestimates the general equilibrium

measure NEV for all levels of transport cost reductions. For a

50 per cent transport cost reduction, the degree of

under-estimation amounts to approximately 4 per cent. Transport cost

savings at this level underestimate the NEV by almost 6 per cent

which also indicates certain gains to newly generated traffic.

The percentage ratios of TS, PE, NEV and DUL to ANI are about

the same sizes as in the case of perfect market economy model.

The numerical computation lists for the model are given in appendix four.

Ta bl e 6: Th e ca se of in cr ea si ng re tur ns to sc al e. (P = AT C) . 10 11 -A t TS PE EV DU L N E V TS % PE % NE V% DU L% 15 % 44 .8 45 .1 55 10 .6 44 .4 54 83 83 .5 82 .2 19.6 25 % 74.6 75.3 90 17 .2 72 .8 91 82 82 .7 80 18 .9 35 % 10 4. 3 10 5. 8 12 7 23 .9 10 3. 1 128 81 .5 82 .7 80 .6 18 .7 50 % 14 6. 6 151. 6 18 4 34 150 18 5 79 .2 81 .9 81 18 .4 60 % 17 8 18 2.5 22 4 40 .8 18 3. 2 223 79 .8 81 .8 82 .2 18.3 17

VTI Notat T 107 Ta bl e 7: In cr ea si ng re tur ns to sc al e (P = MC < AT C) 10 11 TS PE EV DU L N E V TS % PE % NE V% DU L% 15 % 44 .8 45 .1 57 11 .5 45 .5 56 80 80 .5 81 .3 20.5 25 % 74 .4 75 .2 95 19 76 94 79.1 80 81 20 .2 35 % 10 4 10 5. 6 13 5 26 .7 10 8. 3 13 5 77 78 .2 80 .2 19.8 50 % 14 8. 4 15 1. 7 19 5 37 .5 15 7. 5 19 5 76 .1 77 .8 80 .8 8 19 .2 60 %

19.

THE MODEL WITH AN ALLOWANCE FOR PASSENGER TRAFFIC -The case of commuting

The perfect market model and its various extensions treated in

the analysis so far, merely considered the transportation of

goods between regions. Those cases implicitly assumed the exis-tence of an interregionally immobile working population in each

region. Therefore, there was no scope for any kind of

inter-regional passenger transportation.

This model attempts to analyse the benefit measurement problems

of road investment appraisal as adressed earlier, by allowing

for the possibility of interregional commuting. This is acheived

by incorporating a very simple mechanism for the transportation

of labour from one region to the other, in the model.

In particular, it is assumed that one of the regions is endowed

with two categories of labour force. The labour of the first

category specializes in the production of local goods and the

second is suitable for the production of goods elsewhere, thus potentially expanding the labour market there. The supply of the

latter is assumed to be a function of the price level in the

home region, the wage prevailing in the region of work and the

transport costs. It is further assumed that commuters payan ad

valorem income tax to the region where they work. The entire

ambunt of tax revenues is transferred to the consumers in the

region of work.

The analysis is carried out for different levels of transport

cost reductions and includes the transportation of goods as well

as peOple. Two cases are considered with regard to income tax

rates of 25 per cent and 50 per cent and a summary of the

20

Mbdel with 25 per cent income tax

The figures shown in table 10 clearly indicate that, in this

case, transport cost savings for existing traffic underestimate

the increases in national income substantially. A 25 per cent

reduction in unit transport costs leads to the transport cost

savings which are approximately 66.5 per cent of the national

income changes. The corresponding figure for a 50 per cent

reduction in the transport costs is 61.5 per cent. Thus, the larger the transportation cost reductions, the larger the degree of underestimation. The general equilibrium.measure of benefits, NEV, stands for the net welfare gains in the economy as a whole, acccording to the definitional framework established earlier. A comparison between the partial measure (PE), of the benefits and

the (NEV) shows considerable discrepancy between the two

measures due to small as well as large transport cost

reductions. This discrepancy may be represented by a true

multiplier defined as the ratio of the NEV to PE, for different

levels of transport cost reductions. The true multiplier is then compared with Tinbergens multiplier defined as ANI/TS.

Table 8

-At True multiplier Tinbergen multiplier

15% 1.07 1.50

25% 1.08 1.53

35% 1.10 1.57

50% 1.13 1.63

The comparisons above suggest a significantly lower value of the

true multiplier than that of the Tinbergen multiplier for the

various transport improvements. However, a larger improvement in

transport facilities induces a slightly larger true multiplier

which is also tha case for the Tinbergen multiplier. 1¥T1_Netat T 107~

21

Mbdel with 50 per cent income tax

An obvious consequence of the higher rate of income tax is that

the magnitude of road improvement benefits is reduced in

absolute terms for all levels of transport cost reductions. The partial measure underestimates the general equilibrium measure of the benefits and this discrepancy is larger than the earlier case of a 25 per cent income tax.

The true multipliers and the Tinbergen multipliers are computed

to illustrate the degree of deviation between the partial

measure and the general equilibrium measure of the benefits.

Table 9

-At True multiplier Tinbergen multiplier

15% 1.17 1.58

25% 1.20 1.61

35% 1.25 1.66

50% 1.27 1.71

The values of the multipliers show that, for a given reduction

in transport costs, both the true multiplier and the Tinbergen

multiplier will be greater the higher the rate of income tax.

Moreover, the true multiplier gets greater for greater

reductions in unit transport costs which is also the case

regarding the Tinbergen multiplier. A summary of the results

from the two cases is provided in tables 10 and 11. The results also indicate that the ratios of TS, PE and NEV to ANI in the commuting model are much smaller compared to such ratios in the perfect market model, whereas the percentage ratio of DUL to ANI

is much greater in the case of the commuting model. The

numerical computation lists are provided in appendix five.

Ta bl e 10 : Th e ca se of in te rr eg io na l co mm ut in g fo r wo rk ( 25 % in co me ta x) . -A t TS PE EV DU L N E V AN I TS % PE % NE V% DU L% VTI Notat T 107 15 % 53 .2 54 .4 80 22 58 80 66 .5 68 72 .5 27 .5 25 % 88 .5 92 13 6 37 99 13 5 65 .6 68 .2 73 .3 27 .4 35 % 12 3. 8 13 0. 4 19 5 53 14 3 19 4 63 .8 67 .2 73 .7 27 .3 50 % 17 6. 4 19 0 28 7 73 21 4 28 7 61 .5 66 .2 74 .6 25 .4 60 %

Ta bl e 11 : Th e ca se of in te rr eg io na l co mm ut in gfo r wo rk (5 0 % in co me ta x) . 10 11 TS PE EV DU L NE V TS % PE % N E V % DU L% 15 % 44 .3 45 .4 70 17 53 70 63 .3 64 .9 75 .7 24 .3 25 % 73 .8 76 .8 12 1 29 92 11 9 62 64 .5 77 .3 24 .4 35 % 10 3. 2 10 9. 1 17 6 40 13 6 17 1 60 .4 63 .8 79 .5 23 .4 50 % 14 7.3 15 9.6 25 6 54 20 2 25 2 58 .5 63 .3 80 .2 21 .4 60 %

23

24

SUMMARY AND CONCLUSIONS

In an earlier VTI report, (Hussain and Jansson, 1990), it has been learnt that transport cost savings to existing traffic and

gains to purely generated traffic are almost equal to the true

net welfare gains computed by a general equilibrium analysis of

road investment benefits, in an economy with perfect markets. In this paper the relationship between the partial and the general

equilibrium benefit measures is examined under various

conditions of market failure. One important conclusion is that

different types of taxes lead to situations where the partial

measure underestimates the general equilibrium measure.

Nonetheless, the discrepancy is not very large and the partial

measure therefore appears to provide a good approximation of the general equilibrium.measure of the benefits in these cases. The case with decreasing average total production costs leads to a slightly larger gap between the two measures of road invest-ment benefits compared to the tax models. Nevertheless, the introduction of the possibility of job commuting between various

regions considerably alters these conclusions. If road

investments lead to a situation where potentially idle factors can be brought to use, as is assumed in the commuting model, the partial measure of the benefits considerably underestimates the

general equilibrium measure. However, inthis study the true

multiplier used to measure the degree of discrepancy between the

two benefit measures is much smaller compared to the

corresponding Tinbergen multiplier which is defined by the ratio

of national income increases to transport cost savings for

existing traffic, due to transport improvements. Moreover, it is important to note that the results of the commuting model depend on the assumption that the passenger traffic is comprised only of commuting in the model. The inclusion of passenger traffic in the model, in a broader sense, would lead to a relatively lower value of the true multiplier.