You are on page 1of 21

Ann Reg Sci (1991) 25:19-39

--The

Annals o f - -

Regi0nalScience
Springer-Verlag1991

Transportation and telecommunications costs


S o m e implications of geographical scale Ilan Salomon 1 and Joseph Schofer 2
I Department of Geography, The Hebrew University, Mount Scopus, Jerusalem, 91905 Israel 2 The Transportation Center and Department of Civil Engineering, Northwestern University, Evanston, IL 60208, USA Received September 1989 / Accepted in revised form June 1990

Abstract. The decreasing costs of telecommunications and the often increasing costs of transportation have given rise to claims that information-intensive activities are becoming footloose. One of the assumptions underlying these claims is that the cost of distance in telecommunications is negligible or very low. This paper examines the relationship between distance and interaction (telecommunications and transportation) costs and rates, with particular emphasis on the effects of geographical scale. Focusing on data from Israel, it demonstrates that the costs of distance are persistent even in telecommunications systems; that for short distances or small regions, transportation costs are not necessarily higher than telecommunications costs; and that pricing of telecommunications services by governments (or PTT's) often does not reflect the costs of providing the services. This creates a cost distribution which differs from Euclidean geographical distances: discontinuities in the rate structure of telecommunications distort the distance-cost schedules and, by creating barriers, may affect location decisions. The paper also demonstrates that actual interaction costs are context specific and therefore no general model has been formulated. Instead, an accounting procedure, which can be used by decision-makers considering (re)location in specific contexts, is suggested.

1. Introduction

Economic activities depend on interactions among agents for their normal operation. The costs of interactions have been identified as important factors in classical location theory, and the transport costs of materials, final products and labor underlie many location models. In recent years, with the growing importance of information resources as major inputs and outputs of economic activities, attention is shifting to the special case of "transporting" information (e.g., Pye 1977; Goddard 1973; Kutay 1986; Johansen 1987; Dieperink and Nijkamp 1987).

20

I. Salomonand J. Schofer

Means for transporting information have been significantly enhanced in recent years with the advent of New Information Technologies (NIT). These include a variety of microelectronic-based technologies and applications which combine telecommunications and computers to accommodate information processing, transmission and storage across distant locations. Spatial interactions are maintained through a variety of communication modes, some of which are NIT-based while others are travel-based and require the co-location of the participants. In this paper we deal with the distinction between these two modes. The shift toward an information-intensive economy, coupled with the very rapid development of NIT, has given rise to a number of hypotheses about the spatial impact of these changes. One popular claim is that telecommunications will relax many of the current locational constraints, making activities more footloose, and thereby bringing about a decentralization of activities, an exodus from dense urban centers to the periphery. Alternatively, it has been suggested that NIT will bring about a restructuring of the spatial pattern of activities, not necessarily in the outward direction. With the introduction of another dimension of flexibility, firms can take advantage of telecommunications to entertain other location economies. Firms may be able, for example, to disperse some of their activities to peripheral locations, while reconcentrating others to headquarters in central cities where agglomeration economies still outweigh the economies offered in remote locations, without losing control over the remote activities. Each of these conflicting views is based on numerous assumptions which have not received much attention to date compared with the attention paid to policy preferences for achieving certain impacts. For example, the decentralization hypothesis is quite attractive to policy makers interested in promoting development in hinterlands to replicate the economic success experienced in other areas (see, for example, Saxenian 1985). Other policy makers, concerned about avoiding congestion in infrastructure-poor areas, may focus on opportunities for agglomeration. At present, empirical evidence of these spatial impacts is hard to find. This may be attributed to the fact that many of these technological developments are very recent, many confounding factors are acting simultaneously, and methodologicalty, it is difficult to identify the causal relationship between investments in NIT and development impacts (Wellenius 1984; Hudson 1984). Research efforts have thus focused on theoretical arguments and limited empirical studies which attempt to identify specific relationships. It is almost uniformly assumed that the costs (to the user) of overcoming distance in telecommunications are lower than the costs of travel-based communications. Thus, for information-intensive activities, telecommunications could relax distance-related constraints. This paper focuses on this particular assumption that is embedded in the hypotheses about the spatial impacts of NIT. Recently, there has been growing interest in the role of barriers to communications and transportation, as it has been realized that distance alone, albeit a barrier, cannot account for many directional variations in interaction densities. Of course, the ability to construct or remove such barriers has come into the foreground in the context of "One Europe". The findings of the present effort are

Transportation and telecommunicationcosts

21

of relevance to the study of barriers, as attention is paid to the effect of discontinuities in rate structure as well as in other spatial cost factors. The common assumptions used to support the use of telecommunications as a spatial policy tool are briefly described in Sect. 2. Section 3 presents the hypotheses of this study, followed in Sect. 4 by a discussion on the cost schedules o f the two modes of interaction considered here: telecommunications and transportation. Section 5 presents some empirical findings from Israel, followed by Sect. 6 which discusses the implications of these findings. Section 7 presents some conclusions.

2. NIT and development


The hypothesis that NIT can foster economic development is based on two major arguments. First, it is assumed that the availability of information will increase the productivity of activities located in peripheral regions. For example, farmers in underdeveloped rural areas will be able to obtain information on improved production technologies, such as the use of pesticides or fertilizers and hence will increase crop production. Similarly, access to timely market information will enable better scheduling of shipments to the market, thus increasing crop revenues. It is not obvious, however, that such information is best channelled through NIT (Darrow and Saxenian 1985). It may, for example, be more efficient to disseminate it through mass media (Massam 1989). The second argument is that N I T will encourage information-intensive activities to locate in peripheral regions, and through their multiplier effects replicate the economic boom experienced elsewhere (e.g., in Silicon Valley, Route 128 in Massachusetts, etc.). This will happen, it is presumed, as managers make long term choices about locating and relocating, based in part on their expectations about patterns and costs of communications. To the extent that NIT advances attract new firms, these will attract skilled and educated population groups, who may generate greater demand for services and amenities, thus producing the multiplier effects. Of-course, many other factors operate to determine the ability and willingness of labor to migrate to remote areas. The locational patterns and economic impacts o f high-tech industries have, in recent years, received the attention of both researchers (Malecki 1985; Dieperink and Nijkamp 1987; Felsenstein 1985; Scott and Stroper 1987) and policy makers. Policy makers sometimes overlook the assumptions embodied in the analyses and fail to identify the causal relationship which evolved into success in some places and failure in others. Three assumptions common in these arguments are briefly presented here: /) First, that for information-intensive activities and high-tech industries, information is the only important - or the principal - input and output. Hence, access to quality telecommunications can substitute for access to transportation facilities. This assumption treats all high-technology and information-intensive activities as having similar needs, which is a common mistake in the literature and even more so in the policy process. Information-intensive activities can be classified into three groups which exhibit different degrees o f dependence of location on interaction technologies.

22

I. Salomonand J. Schofer

High-tech manufacturing industries have a significant freight movement requirement (Mahmassani and Toft 1985; Button 1988); since it is the product, rather than the process, that is knowledge intensive, such firms may be quite dependent on good transportation connections. The locational constraints of activities which are "pure" information processors depend on the nature of the information they demand. Thus, a second type of information-intensive activity includes "manual" information processing units, such as bank and credit-card clearinghouses. These often locate where telecommunications access is good, and inexpensive, and where some other advantage is available, e.g., a large supply of reliable, low-cost labor. Of course financial clearinghouse activities still'ultimately depend on paper transactions. Thus, while funds may be transferred electronically, the physical movement of checks and bills underlies the entire process. This movement is typically accomplished by mail, truck, and (in the United States) by air cargo. The third type of information-intensive activities entails primarily less-structured interchange, and includes universities, research and development organizations, and some office activities (see Pye 1977). In these, there is a greater dependency on the ability to meet professionals from within and outside the organization, thus their location is less influenced by the introduction of NIT, and more so by the needs and desires of their critical resource, skilled people. Note that each of these types is very different in the quality of the labor force it employs and the transport network services required; thus, interaction costs, infrastructure requirements and the multiplier effects may be quite distinct. ii) Second, it is often implicitly assumed that telecommunications infrastructure and services are universally available. While this is technologically possible, economic and political considerations, including the low demand (and revenues) in peripheral regions, have resulted in spatial inequalities in the distribution of telecommunications in both developed and less developed countries (Henckel 1988; Clapp and Richardson 1984; Salomon 1988). The availability of NIT in remote areas is, of course, a necessary condition for any policy aimed at its utilization for development. For interaction to take place, there are minimum requirements as to the availability and quality of each candidate mode. iii) The third assumption is that telecommunications-based interactions can substitute for travel-based (face-to-face) interactions. This is based on the belief that reductions in telecommunications costs are sufficient conditions for the substitution (see for example, Webber 1984). It ignores the fact that there are basic differences in the nature of the information that can be transmitted by each of these modes. Substitutability depends on contextual factors such as purpose, familiarity between parties and a host of other factors (Salomon 1986; Salomon and Schofer 1988; Moore and Jovanis 1988). Only when the informational benefits expected from an exchange through a particular medium will not render one mode to be dominant will costs become a major factor in the choice between telecommunications and travel. In other words, when the two modes supply a communications quality which is comparable, then costs become a major consideration. This theory will be elaborated below. The following discussion assumes that cost considerations are an important decision factor in the choice between telecommunications and travel. This means

Transportation and telecommunicationcosts

23

that the interacting parties do not gain any marginal informational benefits from the fact that they can interact face-to-face. It is important to recognize that this situation may n o t be typical; that is, one of the competing modes may have significant non-cost advantages. Often, that mode will be transportation. Arguments in the literature do not address the possibility that the spatial impacts o f NIT differ according to geographic scale. Yet, the consequences o f relaxation of distance-related constraints are likely to differ between regions the size of Australia and those the size of Israel or the Netherlands. This is an issue which warrants attention from geographers and regional scientists; it is the focus of our main hypothesis.

3. Hypotheses
The major hypothesis suggested in this paper is that the impacts of the changing telecommunications environment are different in small geographical units than in larger ones. Because telecommunications and transportation cost schedules exhibit different sensitivities to distance and are characterized by discontinuities, the implications of the cost ratios between these modes in small geographical units may be different from those prevailing in larger ones. The term "small" refers to geographical units which are up to a few hundred kilometers in radius. They may be countries, or regions within a country. In such areas, a round trip from the periphery to the center (or vice versa), allowing sufficient time to conduct business, can be accomplished within a single day. We avoid a strict definition of size as it may vary due to particular circumstances of the shape of the region or the transportation infrastructure and services available. It is generally accepted that the costs of interaction via telecommunications are lower than the costs of travel-based, face-to-face interactions, largely because telecommunications costs are less sensitive to distance, and no time is lost for travel. The hypothesis presented here is that this advantage of telecommunication is of lesser importance in small units than it is in large ones. Consequently, telecommunications costs may have a smaller effect on the spatial distribution of economic activities in small countries. Several factors which may act in opposite directions in small areas need to be considered. First, the risk of making the wrong location decision, in the face of small cost differentials is lower in small regions than in large ones, where the cost differentials may be wide. This leads to the hypothesis that in smaller countries or regions, lower risk expectations may lead decision-makers to attempt to capitalize on options opened by NIT, thus favoring relocation. But, this will occur only if the periphery offers some advantages which are not available at the central areas, and which exceed the transaction costs. Another factor supporting dispersion in small countries is the fact that outward commuting, from the big city to peripheral locations, may be a viable option (Felsenstein 1985). This allows firms to settle their information-intensive activities in locations which are accessible to high-skilled labor even if that is not local. Taking advantage of other locational benefits, such as government incentives, low communications costs, and the availability o f reverse-commuting labor force, can thus offer the "optimal" location.

24

I. Salomon and J. Schofer

4. The cost structure of transportation and telecommunications

The cost of transporting materials, goods or labor has been identified in classical location theory as a dominant factor in determining location. The more recent identification of agglomeration economies also supports the notion that transport costs are dominant factors, the only difference being that the "goods" considered are less tangible: proximity to services, information sources and transactional activities. The decision-maker considering a (re)location is thus facing a "cost map" which (at least conceptually) identifies the total costs of conducting business from any point in a geographical space The cost at a given location is the sum of two components: the site-specific local costs (e.g., land, labor, services and the negative costs of government incentives) and the "costs of distance", the costs to gain access to non-local resources (e.g., information, raw materials, intermediate goods, specialized labor, and clients). It is useful to explore the implications for the distance cost map of the rapid evolution of interaction technologies, both physical transport and telecommunications. To that end, it is first necessary to clarify some of the differences between transport and telecommunication costs. From the user's perspective, the major differences are: i) the relationship between costs and rates, ii) the structure of actual costs, iii) the marginal costs of additional activities, and iv) the temporal availability of the two modes.

4.1 Costs vs. rates

To the extent that prices paid by users can be systematically and purposefully controlled, differences between those prices and actual costs can be used to influence market response. The availability of such pricing control is related to the degree of centralization of system management and billing. For example, in most transportation systems, users pay directly (and often immediately) for certain costs (e.g., fuel, fares). They may pay only indirectly for infrastructure, perhaps through separate taxes or user fees. Billing in telecommunications tends to be centralized, allowing the supplier greater flexibility in exercising pricing policies which are based less on actual costs of providing the service than is the case for many transportation systems. This facilitates the use of pricing to achieve certain policy objectives almost independent of true costs. The differential treatment of capital and operating costs offers an example. Because these costs are incurred, paid, and often financed in totally different time frames, wide disparities between true and billed costs can be introduced into the market. While ultimately both of these costs must be paid by someone, institutional structures can permit cross-subsidies and discounts to achieve certain policy objetives. While from the narrow standpoint of economic efficiency, marginal cost pricing offers the best pricing scheme for use of either telecommunications or transportation services, alternative economic and political development objectives often prevail, leading to "distorted" price structures. Such distortions seem more common in communications systems than in transportation.

Transportation and telecommunicationcosts 4.2 Cost structures

25

The operating costs o f both transportation and telecommunications are characterized by some sensitivity to distance. In the case of transportation, fuel, labor and maintenance costs are proportional to the amount of usage, hence are related to distance. In the case of telecommunications, the costs o f distance are not so obvious, and in fact the popular notion is that there is no cost to distance at all. However, long-distance telecommunications requires more equipment (transmission channels, repeaters, switches and direct labor for their operation and maintenance) and therefore there is a justification to charge for distance to recover the capital invested in the long-distance link and its operation. The sensitivity to distance in operating costs alone (without capital recovery) o f telecommunications is, of course, lower than in transportation. However, because there is a positive distance cost for transportation, teIecommunications providers (where competition is limited) may be inclined to impose significant distance charges because the market expects and will accept them. Transportation costs are clearly sensitive to distance, as well as to time-of-day, particularly in congested areas. Telecommunications rates, on the other hand, are more sensitive to the duration of the interaction but also to time of day (congestion) 1. This is done to ensure efficient use of the infrastructure. The differential sensitivity to distance and time between transport and telecommunication costs creates situations where the cost advantage of telecommunications may be diminished, as will be demonstrated below. Another difference arising from the basic rate structure of the two modes is the sensitivity to the number of participants in an interaction. In telecommunications, the marginal costs of more participants at a node is very low: a speakerphone, multiple handsets or a studio for teleconferencing allow more individuals to take part in an interaction. Connecting additional individuals who are located elsewhere usually involves a surcharge beyond the charge for the additional call. This surcharge is in most cases small. In the case of transport, however, the marginal costs of additional participants may be very high if individual value of time, travel fares and accommodations have to be paid, and lower if participants share costs of travelling in a single vehicle. 4.3 Marginal cost of additional activities The third difference mentioned above is in the costs o f marginal activities. Telecommunications have a fixed cost per contact made. Contacting a number of individuals at a given distance would multiply the costs. However, under the travel alternative, the marginal costs of contacting an additional party at the same location is smaller than if the contact is made from the original site. Consequently, in terms of the cost map, one must examine communications alternatives in terms o f locations, timing and marginal costs.

1 In many cases local call rates are not time sensitive, although there is a general trend toward imposing local measured service (Fuhr 1986; Levin and Cale Case t988).

26

I. Salomon and J, Schofer

4.4 Temporal availability


The fourth important difference is the temporal availability or immediacy of each mode. The availability of some transportation options is not continuous: rail and air services are offered only on a given schedule, and the inability to use them at any desired time implies additional costs to travel-based interactions that do not affect NIT-based interactions. It is, however, difficult to quantify this cost.

4.5 Prototypical cost schedules


Prototypical transportation and telecommunications cost schedules by distance are schematically shown in Fig. 1. Transport-based interactions costs generally increase monotonically with distance, up to a point where a discontinuity results from costs associated with subsistence and overnight accommodation. Discontinuities may also exist if the mode of transport considered has a rate structure typical of public carriers. For example, bus, train and airline fares are commonly structured with discontinuities with respect to distance, whereas the costs of operating an automobile is more typically that shown in Fig. 1. By contrast, telecommunications-based interaction costs are characterized by stepwise increases (Weinhaus and Oettinger 1988). The delineation of the discontinuities is determined by the supplier and/or the regulator. Note that the schedules shown in Fig. 1 assume the terminal (non-distance or duration-based) costs associated with the interaction to be identical. It is reasonable to assume that most trips are preceded by some coordination via telecommunications. That cost should be added to the costs of the trip. We ignore this and assume the equal terminal costs for simplicity, as will be discussed below. The discontinuity on the transport schedule (point A) reflects the distance beyond which it is not expected that a business person will make a round trip in one work day. This, of course, depends on the transport technology used and on norms prevailing in the particular country or institution. For example, where domestic airlines provide frequent service, the distance can be as long as 1000 km, while in countries where business travel relies on motor vehicles, we would expect that distance to be in the magnitude of 200 km (one way). Beyond this distance (or comparable travel time), the interaction costs must include the accommodation expenses, which widen the relative cost advantage of telecommunications.

I--

~ m m

o
I

.]TELEeOMMUNICATION
f

,"

.' I I

J
DISTANCE

Fig. 1. Schematic relationship of the costs of distance in telecommunications and transportation

Transportation and telecommunicationcosts

27

The discontinuities on the telecommunications-based schedule are determined by the supplier, primarily on revenue considerations, although technical and policy factors may also play a role. In most cases, these discontinuities follow the boundaries of "area codes", perhaps defined from a marketing perspective, but multiple rates also appear within an area code. While the rates usually reflect distance, a logical distance-driven relationship is not necessary. For example, one U.S. telecommunications supplier divides the continent to two regions along the Mississippi river. Communicating within a region for a distance of 2000 km will thus cost less than a 300 km interaction across the Mississippi. In 1984 another supplier offered rates totally insensitive to distance, while more recently there has been a trend to relate rates to distance. Deregulation of the U. S. telecommunications industry has facilitated such market-oriented pricing schemes. The spatial distribution of the costs shown in Fig. 1 is obtained by projecting the schedules to a geographical space. This produces concentric regions around a center in which the ratio between the costs of the two interaction modes exhibit discontinuities. For example, a location inside such a ring provides a greater advantage for telecommunications, whereas a location just outside of the ring reduces that advantage. The above discussion points to the fact that general relationships between costs of distance for telecommunications and transportation cannot be studied analytically because they vary with contexts and circumstances. Therefore, we have employed a case study approach for the purpose of location costs analyses. We argue that the decision-maker considering (re)location choice should calculate the total interaction costs involved in conducting business. It is not suggested that certain locations are preferred for all information-intensive activities. We can suggest that a firm with a particular spatial interaction pattern should consider locating in a region which accommodates its idiosyncratic interaction needs at the lowest cost.

5. Interaction costs: some empirical findings To demonstrate the distribution of interaction costs, some costs maps were developed for interactions in Israel. In developing this case study and its variants, many simplifying assumptions were made. However, varying most assumptions does not alter the major findings. A firm, it is assumed, seeks a location which minimizes its total (telecommunications and travel-based) interaction costs. Shefer (1986) reported that Israeli high-tech firms spend only about 2070 o f their total operating costs on telecommunications. (New firms, characterized by a greater share o f research and development activities, had a mean expense of 3.170.) Although this figure is small, it is not insignificant as a factor in location decisions. The actual expenses were not found to be related to location, although perceived costs seemed to vary among different locations. The case-study is based on data for the telephone system in Israel. The telephone, although not exactly a New Information Technology, is still the backbone for most telecommunications-based interactions in Israel and in many other

28

I. Salomon and J. Schofer

countries. Note that some NIT's, such as facsimile, use the telephone network. Also, the economics o f NIT seem to follow those o f the telephone. The telephone rate structure in Israel is based on two dimensions: time-of-day and distance. A "message unit" constitutes the basic charging unit. Its current price is 0.1242 NIS (New Israeli Shekels). Four levels of distance are defined by the duration o f one message unit. Local calls in most cases are not time dependent 2. Rate A applies for short distances beyond the local calling area, and is 72 sec. Rate B is 48 sec and applies to intermediate distances, and rate C is 16 sec long and applies to the longest distances (the above figures apply at the peak period, in which most businesses operate). In addition there is a time-of-day pricing scheme, where peak time is between 08:30 and 13:00 and is charged at the full rates. Between 13:00 and 21:00, an intermediate rate applies, at 507o of the peak rate, and 21:00 to 08:30, the discount rate applies, at one sixth o f the peak rate. In addition, subscribers are charged a flat monthly fee of 10 NIS, and residents of remote areas are charged only a fraction of that, as a compensation for their relatively higher usage charges, which may have nothing to do with the actual cost o f providing service. In the analysis below, we explore telephone costs in Israel under varying assumptions. The geographical units used are the 104 towns and villages with population of 5000 inhabitants or more (1983), which represent 87% of the Israeli population.
5.1 The distance costs

The telephone rate structure in Israel is based on the division of area codes, with an implicit assumption about distance pricing. Intra-area non-local calls are charged by rate A, which also applies to some inter-area calls. The rational is that neighboring area codes will be charged by either rate A or B and areas that do not have a c o m m o n border will be charged at rate C. We can find very wide variations in the distance for which each rate applies. For example, rate B is, in some locations, applied for a distance o f / 6 km and in others for a distance of 250 km. Such distortions seem to be a result of historical developments, i.e., the slicing off of new area codes from previous larger zones, without adherence to the criteria o f either distance or common border. In Fig. 2, the rates applied to each o f the 104 towns by the distance from the three metropolitan areas are shown (rates are symmetrical in terms o f directionality). One would expect a positive monotonic relationship between telephone rates and distance if the generally accepted theory holds. This figure shows that all of the rates rise with distance. The rates from Tel-Aviv (2a) show a true positive monotonic pattern. The rates from Jerusalem (2b) and Haifa (2c) also exhibit distance dependency, but both show some anomalies, where places further from the metropolitan area have lower telephone rates than some nearer places. These rate distributions show that there are locations which enjoy telephone rates which are lower than would have been implied by their distance. For interaction-intensive activities, such locations offer communications cost advantages at this time. 2 Localmeasured service is being introduced in areas served by digital switches.

Transportation and telecommunication costs

29

,3- A

I
2-

I ~ / ........ ~ " O ....... O 20 ~

TEL AVIV . . . .

40

60

80

t O0

120

140

! 60

-B

DISTANCE (kin)

i-

'ii'ii~ !ii ........................................

.i

f'
O
0 20 40 SO 80

JERUSALEM .......

f O0

120

,
! 40

160

DISTANCE (km)

V
HAIFA - -

o. o zo 4o 60 Bo ~oo ~zo
~4o

~6o

DISTANCE (kin)

F i g . 2. T h e t e l e p h o n e r a t e s t o t o w n s b y d i s t a n c e f r o m Tel-Aviv, J e r u s a l e m a n d Haifa

30

I. Salomon and J. Schofer

Another lesson from these distributions concerns the equity o f the rates. It is a common practice in regulated telephone systems to apply rates which provide some cross-subsidization within the system 3. This is evident, for example in the fact that the high-traffic link between Jerusalem and Tel-Aviv, only 60 km apart, is charged by the highest rate (C), which generates revenues which can be used to offset costs of more remote low-traffic links. If the true cost of service is strongly distance-based, cost-based rates should follow a straight schedule of the type illustrated by the straight line in Fig. 2. Locations under a line o f this form are subsidized by the rates applied to locations above that line. The actual shape of the official distance-price schedule will reflect political decisions about cross-subsidies, as well as true costs of service. In a governmentowned, or heavily regulated, telecommunications system, there need be little or no relationship between rates and distances. On the other hand, it is apparent that at least s o m e of the costs o f telecommunications services, for both capital and operations, are related to distance. This provides a managerial incentive to include some distance component in rates, an incentive which is supported by market expectations for this kind of rational basis for pricing.
5.2 M i n i m u m t e l e c o m m u n i c a t i o n s costs z o n e s

To identify a minimum costs location, an accounting procedure is suggested, whereby the sums of the products of telephone traffic volumes by rates to the various destinations are compared for different locations. That, o f course, does not ensure the identification of the real minima, unless all possible locations are tested. However, as a heuristic, especially in small systems, it is likely to provide a reasonable solution. For example, given the geographical structure of the Israeli economy, it is reasonable to assume that for most firms, the information sources/destinations are within the three large metropolitan areas. The costs of calling the three metropolitan areas is the summation of: C where: a
a ( w ~ t / r l + w ~ t / r 2 + w ~ t/r3)

= the costs of one message unit; w i = the weight of destination i (% of calls); the set w . . . w represents the i communications mix for a particular firm; t = the duration o f interaction (seconds); a firm-specific cost index for a particular location might be simplified by setting t = 1; r i ~-- the rate, expressed in duration of message unit; i = destination.

Users have different geographical patterns of communications. The relative share of volume to be exchanged with each destination is an attribute of the particular business type, and location. For example, let us first assume that all con3 Suchcross-subsidizationhas recentlybeen disallowed under the deregulation of telecommunications in the United States and the United Kingdom.

Transportation and telecommunication costs

31

HAIFA

06

TEL

O.MAUAKHI.~'X~JERU$ALEM -Oz

0 5 I*

"BE'ER SHEVA

057

!o.j
0 I0 20 : 30
r

km

Fig. 3. The division of Israel into area codes, 1988

32

I. Salomon and J. Schofer

tacts are in Tel-Aviv, the largest metropolitan area. The obvious minimum cost location would be within the boundary of the local telephone exchanges, where local calls are not time measured, but are charged at the rate of a single message unit. On the other hand, if we assume that the interactions of a firm are equally distributed between the three metropolitan areas, and we consider the telephone rates alone, then the minimum cost location would be at area code 085 which includes the towns of Ashdod and Kiryat Malachi (see Fig. 3). This area is outside of the triangle formed by the three metropolitan areas. This seemingly paradoxical situation is a result of the deformation of the distance-cost relationship in telecommunications rates. By contrast, in the case of transportation, the minimum cost location is at one of the three nodes, as the costs are directly related to distance, and in addition, there are usually terminal costs which can be reduced or eliminated if one locates at a terminal. In the case of telephone traffic there are some similar "terminal" costs, referred in the next section. Varying the assumption about the spatial distribution of interactions result in different preferred locations. When the weight of communications to Tel-Aviv increases to 507o or more, and the remainder is equally divided between Jerusalem and Haifa, then the minimum cost location is in Tel-Aviv. As Tel-Aviv is the center for many economic activities, the distribution of telephone rates provides yet another incentive for agglomeration. In summary, the analysis of the telephone rates shows that a firm considering (re)location might benefit from an examination of its interaction patterns to identify the location which provides its particular minimum telecommunication costs location. This is a location which is in part determined by the (current) structure o f telecommunications rates; anachronistic rates, which are not unusual, imply that the minimum communications cost location may not be intuitively obvious.

5.3 Consideration of transportation costs


Transportation costs exhibit greater sensitivity to distance than telecommunications costs. Given that in most cases the expenses for transport are borne directly by the user, there are usually fewer anomalies in the cost to distance ratio. When fares are charged by a supplier of transportation services, they may distort the actual costs, due to cross-subsidization, marketing or regulation considerations. In the current analysis, we assume that the decision-maker facing the choice among alternative interaction modes considers the costs of travelling by car 4. The car may be a company car or a private automobile, where a fixed rate per kilometer is applied for reimbursement. For July, 1988, the rate in Israel was 0.3685 NIS s per km.

4 Travelby transit is a less relevant option. In Israel, and elsewhere, transit is heavily subsidized, and its cost structure is irregular. More importantly, transit would not be the intercity mode of choice for the business traveler. In larger countries, the logical option might be air or rail travel. s The rate calculated by Cheshev (1988), of 0.67 NIS for a 1300cc vehicle includes income tax and social security. To exclude these, we subtracted an assumed rate of 40% and 5%, respectively.

Transportation and telecommunicationcosts


40
TRANSPORTATION TELEPHONE (60 MIn.) ......... .,,.- a" ............. ,." s,,,s~SwdraS*'a ,,$/

33

30

TELEPHONE (25 MIn.)

,,s Sos

~20
ae a

o-"

o
as" as~

,,

.....

ooo, .......

,o ........

.o ......

O,

. , . Jl, - . ,
0 lb

............................... . ...................
2'0 3b 4'o s'o 6'o
70

DISTANCE (km.)

Fig. 4. Transportation and telephone rates by distance from Tel-Aviv

To the cost of travel it is necessary to add the cost of time spent travelling. As a working assumption we used the value of time of 10 NIS per hour, and a constant speed figure, so that the time loss is only a function o f distance. This biases the costs in favor of close locations, as travel time within metropolitan areas are proportionately longer, due to congestion. Moreover, we ignored the "terminal costs", namely the time costs for vehicle access and parking which consumes greater shares in short, urban trips. The cost schedules for transportation and telephones by distance from TelAviv are shown in Fig. 4. The telephone costs are for 25-min and 60-min call durations. The figure shows only the range of up to 70 km, because the schedules diverge for greater distances as suggested in Fig. 1. The schedules intersect in the range of about 10 km and again at about 48 km. This means that for these particular distances, under the assumptions incorporated in this numerical example, it is more economical to travel for a one hour meeting than to telecommunicate. Here we mean economical in the sense of financial costs, since we have assumed above that the same message is communicated, and with an equal interaction duration. Such intersections have been found in similar analyses we conducted for various telecommunications modes in the United States for relatively short distances, o f up to 300 km or so (Salomon et al. 1989). While the distance at which they occur varies according to the assumptions, it is clearly a phenomenon associated with short distances. Sensitivity analyses showed that these results, the occurrence of travel costs lower than telecommunications costs for certain distances, were robust under the following conditions:

34

I. Salomon and J. Schofer Increasing the value of time up to 20%. Testing three different speed levels, to reflect distance from the congested metropolitan area. Combining increases in value of time and speed of travel.

None of these changes modified the basic results. Beyond the 12 NIS per hour value of time, the travel option does become consistently more costly than the telephone cost. This expected result suggests that, strictly from a cost perspective, travel for higher paid personnel is more expensive than communications. It is important to recognize, of course, that we are assuming that the productivity of their actions is the same under both modes.

5.4 Other costs

In addition to the transportation and telecommunication costs, there are numerous other location-specific costs which affect the cost distribution map. Some of these are indirectly associated with telecommunications. For example, Shefer (1986) points out that although the expense for telecommunications in the high-technology firms he surveyed was only two percent of the operating costs, managers at the firms believed telecommunications costs were important decision inputs. He suggests that the direct costs are only a fraction of the costs perceived by the users. Indirect costs include the costs of poor communications, such as problems or miscommunications in placing orders, sales etc. and the costs of labor or automatic dialling equipment employed to establish communications in a congested area. We hypothesize that in Israel, a major perceived cost component is associated with network congestion. Analogous to the terminal costs in transportation, the costs of "getting through" at the peak period may be high, although it may not be experienced as a direct financial expense as in transportation. As telephone congestion levels are likely to be geographically differentiated, the congestion effect may play a role in the cost map considered by location decision-makers. At least at an anecdotal level, some telephone links in Israel seem to be more congested than others, and hence we would expect that high-volume firms will avoid locating at their ends. On the other hand, many of the entities with which they communicate may be in high density locations, so avoiding congestion costs may not be a simple matter of selecting the right location. Spatial costs differentials are partially off-set by spatial incentives provided by governments. Such policies are in effect in some European countries, including the Netherlands, the Federal Republic of Germany and the United Kingdom. In Israel, government spatial policies provide capital cost assistance to encourage firms to locate in preferred areas. The magnitude of these incentives can dwarf the telecommunications costs factor for those firms which have decided to locate in developing areas (Shefer 1986). The costs of conducting business should thus include not only the transportation and telecommunication costs but also the indirect costs of interaction over space and other costs that reflect level-of service and spatial incentives provided by governmental intervention.

Transportation and telecommunication costs

35

O Fig. 5. Telephone rates from Jerusalem

20

30kin

6. The implications
We have demonstrated above that telecommunications and transportation costs schedules exhibit different sensitivities to distance and to the duration of a contact. The premise of this paper is that decision-makers involved in location choice face a "cost map" which reflects information on the spatial distribution of the costs of conducting business. Careful analyses of the relative costs of transportation and telecommunications would reveal the existence, at least in the short-run, of preferred locations which provide local cost minima. Similar situations where found in the Netherlands (Dieperink and Nijkamp 1987) and in the United States (Salomon et al. 1989). Given the existence of locations which, due to discontinuities, offer lower costs than would be expected on the basis of a distance function, firms ought to evaluate their interaction costs in detail. In some locations firms may enjoy low telephone rates combined with some other spatial incentives. The geographical division into area codes does not overlap with the division to "preferred development areas" designated by the government, nor with the distance from Tel-Aviv. Therefore, it is possible for a firm to seek a location which provides the best combination of government incentives, telecommunications and transportation costs. This is clearly the advantage that the town of Yavne (see Fig. 5) offered in the past (until it was removed from the preferred locations list of the Ministry of Commerce and Industry in 1982). Today, locations in the West Bank such as Ariel combine the best combination in terms of the various location factors considered

36

I. Salomon and J. Schofer

here due to a series o f political decisions. Figure 5 demonstrates the distortions of telephone rates relative to distance from Jerusalem in the central part of Israel. As can be seen, the telephone rates from Jerusalem to the 085 and the 051 areas are lower than those to the closer 08 and 03 areas. New settlements in relatively sparsely populated regions present a special case in terms o f their telephone rates. Their connection to the telephone network can take different forms. If a cable network exits in the region, they are physically connected to it. However, a common short-term solution where cable is unavailable is to connect them via microwave radio link to a remote exchange. In that case they become part of an area code which is not geographically contiguous with them, as can be seen for the Ariel area in Fig. 5. Subscribers in Ariel enjoy the A rate to Tel-Aviv, some 30 km away, while subscribers in Bet-Shemesh, at a similar distance are charged the high C rate to Tel-Aviv. When combined with the incentives given to new settlements in the West Bank, this provides a significant advantage for locating in Ariel, at least for the present. However, as in the case o f Yavne, such advantages, being an outcome o f current government policies, can easily be eliminated in the future. The costs map is probably a fuzzy mental map, at best, for decision-makers, since we assume that most do not invest the resources required to generate an actual map, given the small Share o f telecommunications costs out o f the total operating expenses and the importance of other factors in the location decision. It is probably based on a combination of anecdotal information, some factual information on telephone rates and actual telephone bills, and the general spatial distribution o f the firm's interactions. As fuzzy as it may be, this concept still provides the decision-maker with some relevant background information for choice. The decision-maker may face two basic choice situations. In the short-run, he or she may change the interaction pattern of the firm so as to reduce the costs. This may imply changes in the modes used, the timing and the contact pattern. In the long-run, in addition to the above potential savings, he or she may decide to relocate. Even in the case that a map is produced, there remains substantial room for uncertainty. External sources of uncertainty include the possibility that in the dynamic technological environment, cost reductions will be realized, (as is so often suggested in this context) or that the regulatory environment will change, thus affecting the economics of interaction (the case o f Yavne, described above). As the decision-maker in a firm has no control of these, he or she may decide to ignore them, although the subjective expectations may affect the choice outcome. Of greater importance, perhaps, are the internal uncertainties about future patterns of a firm's operations, particularly after a major locational change 6. The actual pattern of communications in a new location may be difficult to anticipate and could be substantially different from the expectations developed prior to the move. New clientele and suppliers, previously unknown, may prove to change the geographic patterns of interactions. The actual patterns of interaction will evolve following the implementation o f a location change. Before that, the decision-

For a formal discussion on this aspect of uncertainty see Papageorgio and Pines (1988).

Transportation and telecommunicationcosts

37

maker has only a set of expectations with regard to the spatial, temporal and content dimensions of both intra- and inter-organizational communications. The nature and degree of uncertainty associated with future needs and costs probably vary with the type of location decision being made. Three broad classes o f location choices, in order o f increasing uncertainty about future communications patterns, are: 1) those made by established firms which are considering relocation; 2) those made by firms considering branching of some of their activities to additional locations; and 3) those made by new firms looking for their initial bases of activities. Those firms facing decisions involving less uncertainty about their future communications patterns may be better able to benefit from careful evaluation of communications costs maps. Firms having more uncertainty in this dimension may find this exercise useful only if the communications costs map varies substantially over the relevant space, and where the variations are likely to remain stable for some time to come. Why is uncertainty relevant to the current context? In making a locational decision based on the expected costs, the decision-maker is likely to consider the risk, or the costs o f making the wrong decision. The findings presented above imply that in small countries or regions, the risk level is substantially lower than in larger ones, due to the relative costs of travel vs. telecommunicating. Communicating with remote locations either face-to-face or by telecommunications does not involve significant cost differentials in small regions. While the cost differentials between the two modes of interaction in small regions are relatively small, and hence we do not expect major locational impacts of NIT, the risk involved in taking advantage of NIT is also small. Thus, decisionmakers in small countries or regions might be more willing to make a location decision based on the availability of NIT. This would depend on the relative magnitude of the risk compared to the gains obtained in relocation. For the individual decision-maker, the above findings imply that a heuristic procedure for identifying minimum cost locations should include the following steps: 1) First, identify current patterns of communications of the organization, both transportation and telecommunications. In concept, at least, these patterns are in the form of origin-destination-duration tables. 2) Extrapolate these patterns to the future, based on forecasts or scenarios of the ways in which the organization will change in the future, particularly when it (re)locates. It is entirely conceivable that different future patterns will results from different locationat alternatives. 3) Examine the communications costs, rates schedules and other spatial incentives, identify barriers, such as discontinuities in rates or in government supplied subsidies. Along these discontinuities are likely to be the minimum cost locations which can then be examined in detail. 4) Estamine communications costs likely to be experienced at locations at or near the discontinuities identified in step 3. Look for minimum or low cost locations which are otherwise feasible. 5) Reconsider the communication pattern (origin-destination tables) with regard to the candidate locations, to determine whether the pattern might be substantial different at those places. Recycle to step 4 as necessary.

38

I. Salomon and J. Schofer

6) Assess the robustness of candidate locations in the face of uncertainties in communication patterns; technology evolution and costs trends; and government policy and regulatory prospects.
Decision-makers, of course, consider a host of other factors beyond interaction costs. They most likely do not see the two communications modes as fully substitutable, as we have assumed above. They may assign greater weight to the transportation costs, because they are explicit, paid directly as a function of travel purchased, and likely to be increasing or stable, at least in relative terms, while telecommunications generally have opposite characteristics. In a large country these considerations might weaken the attraction of remote locations.
7. Conclusions

There are several practical implications of the findings presented above. First, it is clear that there is no uniform transportation-telecommunications cost plane in the real world. This implies that there is room for a case by case analysis of the actual cost map. Second, the current price schedules create an irregular map, which means that it may be possible to find places which offer communication cost advantages, at least for the present. Third, interaction costs are only a small share of total operating costs and there are many non-cost related factors which affect location. Hence, the exact nature of communications costs may be of only limited importance. Thus, in the near term, NIT should not be expected to produce major alterations in the spatial patterns of the economy. The policy implications of the findings presented above are of two types. First, from a government or supplier's perspective, it is clear that rates can be constructed so as to create spatial differences to suit policy objectives. This may include investments in infrastructure to offer advanced services to less favored areas and relax congestion in dense traffic areas. Decisions about such investments should be sensitive to the relative (un)importance of communications costs in location decisions. Policy makers should also recognize that private decision makers will be concerned about uncertainty, and particularly about the possibility that locational incentives might later be easily removed, thus negating any future benefits. On the other hand, from the individual firm's perspective, there is room for close examination of the cost map so as to capitalize on the prevailing rate structure and available technologies, among other locational factors. However, evidence presented here suggest that the impacts of telecommunications rates (relative to transport costs) will probably be quite small, at least for small regions. Policies aimed at fostering regional development can utilize telecommunications technologies as a tool for that end. However, the impacts of NIT on development are still not properly understood. Naive forecasts about the impacts of telecommunications rely on assumptions which, when examined carefully, seem to contradict the conventional wisdom on the decentralizing effect of these technologies. This paper has shown that at least one common assumption, that telecommunications costs are lower than transport costs, does not necessarily act in the direction of relaxing location constraints.

Transportation and telecommunication costs

39

Acknowledgement. The authors wish to thank Uri Ben-Zion, Daniel Shefer and Frank Boekema for
comments to an earlier version, to two anonymous reviewers for very stimulating and constructive comments, and to Michal Kidron for producing the maps.

References
Button K (1988) High-technology companies and their transport needs. Prog Planning 29 (2):81 - 146 Cheshev (1988) Monthly report on prices, indices and economic indicators. Tel-Aviv (Hebrew) Clapp J, Richardson H (1984) Technological change in information processing industries and regional income differentials in developing countries. Int Reg Sci Rev 9 (3):24i-256 Darrow K, Saxenian M (1985) Worshipping a false god? Dev Seeds Change 1:67-70 Dieperink H, Nijkamp P (1987) A multiple criteria location model for innovative firms in a communication network. Econ Geogr 63 (I):66-73 Felsenstein D (1985) The spatial organization of high technology industries in Israel. M.A. Dissertation, Department of Geography, The Hebrew University of Jerusalem Fuhr J (1986) The rationale for local measured service. Telecommun Policy 10:193- 194 Goddard JB (1973) Office linkages and location. In: Diamond DR, McLaughlin J (eds) Progress in planning, vol 1, No 2. Pergamon Press, Oxford Henckel D (1988) New technologies and regional disparities in the Federal Republic of Germany. Presented at the International Conference on Information Technologies and the New Meaning of Space, Sponsored by the International Sociological Association, RC24, Frankfurt Hudson H (1984) When telephones reach the village. Ablex, Norwood, NJ Johansen B (1987) Information technology and the viability of spatial networks. Papers Reg Sci Assoc 61:51-64 Kutay A (1986) Effects of telecommunications technology on office location. Urban Geogr 7 (3):243-257 Levin S, Cale Case J (1988) Local measured service: advantages and implementation. Telecommun Policy 2:27-35 Mahmassani H, Toft G (1985) Transportation requirements for high technology industrial development. J Transport Eng 111 (5):473-484 Malecki EJ (1985) Industrial location and corporate organization in high technology industries. Econ Geogr 61:345-369 Massam B (1989) Private communications on information dissemination in northern India, using satellite-based tv broadcast Moore A, Jovanis P (1988) Development of a framework for analyzing business travel and telecommunications choices. Transport Res A 22A:4 Papageorgio Y, Pines D (1988) The impact of transportation cost uncertainty on urban form. Reg Sci Urban Econ 18 (2):247-260 Pye R (1977) Office location and the costs of maintaining contact. Environ Planning A 9 (2):149-168 Salomon I (1986) Telecommunications and travel relationships: a review. Transport Res A 20A (3):223 - 238 Salomon I (1988) Geographical variations in telecommunications systems: the implications for location of activities. Transportation 14:311-327 Salomon I, Schofer J (1988) Forecasting telecommunications - Travel interactions: the transportation manager's perspective. Transport Res A 22A (3):219-229 Salomon I, Schneider N, Schofer J (1989) Are telecommunications costs cheaper than travel: an examination of business interactions. 5th World Conference on Transport Research, Yokohama, Japan Saxenian A (1985) Review 3: "Let them eat chips ". Environ Planning D, Soc Space 3:121 - 127 Scott A, Storper M (1987) High-technology industry and regional development: a theoretical critique and reconstruction. Int soc Sci J 112:215-233 Shefer D (1986) The effect of advanced means of communication on the operation and location of high-technology industries in Israel. The Technion, Neeman Institute, Haifa Webber MM (1984) No development without connectivity. Rejoinders. Telecommun Policy 8 (1):67-68 Wellenius B (1984) On the role of telecommunications in development. Telecommun Policy 8 (1):59-66 Weinhaus C, Oettinger A (1988) Beyond the telephone debates. Ablex, Norwood, NJ

You might also like