The Demand For Internet Usage

The first area to be examined in this analysis is the components of demand for internet usage. In this essay the focus of the analysis will be on the components of demand for internet usage. Demand for usage refers to a users consumption of internet services. Thus the first issue which needs to be highlighted is the difference between demand for internet access, and demand for usage. This is an important distinction since "the important characteristic of Internet demand for access is that it is binary. An end user either has access or he does not." (Villasis (1996:61)). By contrast, internet usage refers to an individuals utlisation of internet resources once access has been obtained. In this essay the rate of data/traffic transferred in a given period will be used as the measure of internet usage. Other methods of measuring usage could be total amounts of data transferred or hours of time connected to a service.

Villasis (1996:70-76) discusses five factors which influence demand for internet usage. These elements are the price of usage, cross-price effects, income effects, user preferences and lastly the network effect. The network effect is a particularly important component of the demand for internet usage and will therefore be considered separately.

Firstly, the price of usage. In the context used by Villasis (1996:70) this can also refer to the type of pricing scheme. That is, whether a flat access fee is charged - in which case the usage price is zero, or whether a usage sensitive pricing scheme is used.

Secondly, cross-price effects.¹ Villasis (1996:71-72) recognises the impact of both substitutes and compliments on demand for usage. In relation to internet usage it is important to recognise that the substitutes tend not to be substitutes for internet usage per se, but rather substitutes for different components or uses of the internet. (Villasis (1996:71)) For example the substitutes for email might include telephone and fax calls as well as postal mail. The substitute for purchasing CD's online is buying them through mail order, over the phone or via the home shopping television channels

If we therefore assume that substitutes tend to be for components of internet usage, rather than internet usage as a whole, the cross-price elasticity for internet usage will then be related to the proportion of total usage made up by the substitutable component. For example, the cross-price effect on internet usage of a rise in fax or postal mail prices will be influenced by the proportion of total internet usage the substituted component (e.g. email) makes up. From this it may be possible to argue that the size of the cross-price elasticity between a particular substitute and total internet usage is likely to be low, although this would need to be confirmed through empirical study.

The price of compliments also has an impact on the demand for internet usage. Villasis (1996:72) argues that the compliments to internet usage are not just the costs of computer equipment required, but also the cost of the users time. An increase in time saving computer applications will increase a users demand for usage since for any given amount of time they will be able to get more valuable usage from that time.

The Income Effects. Income also affects demand for internet usage. Villasis (1996:72-73) raises two points of particular interest in relation to this point. It is suggested that internet usage is a normal good. That is, it rises with income. But if the rise in income is associated with a rise in the users wage rate or hours of employment this increases the value placed on time by the user. They may therefore be more prepared to pay for a substitute (such as a conventional telephone call), than wait for the level of congestion on the internet to decline to a level where the user can utilise real time audio applications without degradations in quality. Thus under some circumstances internet usage may be considered an inferior good.

User Preferences and other miscellaneous factors. Villasis (1996:74) uses consumer preferences to account for differences not attributable to the above factors or to the network effect. Changing demographics is one factor cited by Villasis (1996:74) as possibly impacting on the demand for usage. However it is possible to imagine a wide range of other factors which could alter consumer preferences and have an impact on the demand for usage. For example, changes in the weather may induce users to stay indoors and use the internet rather than undertake activities outdoors, whilst a popular television show may reduce the demand for internet usage at that time.²

The final component of demand is the network effect. The network effect refers to the situation where "the value of the good increases with the expected number of units to be sold." (Economides (1996:8)) Network effects can be either direct or indirect.³ It is the direct network effect which has an impact on the demand for internet usage. Direct network effects arise as the result of the effect a new user has on the benefits received by existing network users. In this section the positive network effect will be focussed on, although the impact of negative network effects will also be considered.

How then does the direct network effect arise? The operation of the network effect is best explained by the simple example of the telephone or fax networks.⁴ In the extreme case where there are zero users of the network, there is no benefit in being connected since there is no-one to talk to! However as the network grows, the benefits to an individual of becoming connected increase since there are more people to communicate with.⁵ As the network size increases, so does the benefits to participants since "a new user joining the network increases the range of choice open to other members." (Villasis (1996:60))

The same may be argued to exist in relation to use of the internet. The more WWW sites , or users on the Internet Relay Chat (IRC) network, the greater the benefit a user can expects to receive. In relation to internet use in universities and by researchers ;

Figure 1 below is a graphical representation of a possible relationship between network size, Marginal Private Benefit (MPB) and Marginal Social Benefit (MSB). In Figure 1, Marginal Benefit (MB) is drawn on the vertical access and network size on the horizontal. In the context of the internet, network size could refer to the number of users connected to the internet or the number of WWW sites accessible by internet users, the choice depending on the particular problem one wishes to analyse. It is important then to remember, that in this analysis network size does not refer to the physical capacity of the network. The model to be developed in this essay also assumes that increases in network size do not result in changes in the physical capacity of the network.⁶

For all network sizes up to N* the marginal benefit of increased size is positive. This reflects the earlier explanation of larger network size creating more 'goods'. Furthermore, the MSB curve is drawn above the MPB for all network sizes less than N* due to the presence of the external benefits which existing network users receive from new users joining the network.⁷

Examining the shape of the marginal benefit curves it can be seen that initially the marginal benefit is positive and increasing due to the increasing benefits from expanding the network. Eventually however the marginal benefit reaches a maximum after which the marginal benefit of increased size begins to diminish. The diminishing marginal benefit observed in Figure 1 may be due to a number of influences.

Liebowitz and Margolis (1994:140) argue that although some activities may require a certain critical mass to exist, increased size beyond this level may not contribute significant gains.⁸ Internet shopping is a possible example. Although a certain number of internet users may be necessary for on-line shopping to become viable the gains once this critical point has been reached may begin to fall.

Increased network size may also lead to higher levels of 'spamming'⁹ and other anti-social behaviour, which may reduce the marginal benefit of increased size. Similarly, increased network size can produce the internet equivalent of highway traffic jams, again acting to reduce the benefits of increased size. The problem of congestion is significant enough that it will be examined in more detail later in this essay. Eventually the effects of these negative factors may actually change the network externality from being a positive to a negative.¹⁰ 

We turn now to examine more fully the relationship between network size and the demand for internet usage. In developing this new explanation it will be necessary to use the simplifying assumption used in Economides (1995:5-6) relating to the relationship between the network effect and demand. The assumption made is that all users value the network effect equally and therefore that the increased willingness to pay due to the presence of the network effect "is the same for each unit sold, irrespective of its position on the demand curve." (Economides (1995:6)). This is significant since it means that the network effect shifts the demand curve without altering its slope.

Since the demand for internet usage is based on the benefits received, where the increase in network size produces a positive marginal benefit this will result in an increase in demand. When the increase in size produces a negative marginal benefit then this will result in a decrease in demand. Figure 2 illustrates this relationship.

At network size N+, the corresponding demand curve is DN+. At a price of P* the quantity demanded is Q+ megabytes per hour. As the network size increases to N*, the demand curve shifts outwards to the right and is represented in Figure 2 by the demand curve DN*. At price P* a greater quantity, Q* megabytes per hour is demanded. However as the network grows past N* and moves to N- , the marginal benefit becomes negative due to the previously mentioned influences. The demand curve thus shifts inwards to the left. At price P*, quantity Q- megabytes per hour is demanded, with Q- < Q*. Demand for internet usage is therefore maximised at the network size where the marginal benefit of increased size is zero.

Now that a relationship has been established between network size and changes in quantity demanded it is necessary to develop a convenient way to describe this relationship. A possible way of doing so is the concept of a network elasticity. Here we will use the notation 'Ne' for network elasticity. Thus:

The network elasticity is potentially useful in a number of ways. For example if Ne < 0 , then ceteris paribus the marginal benefit of increased network size is negative, or conversely, if Ne > 0 then the marginal benefit of increased network is still positive. The network elasticity could also be used to succinctly indicate the strength of the increase or decrease in demand from a change in network size. For example, if | Ne | > 1 , then the percentage change in demand will be greater than the percentage change in network size.

In this section of the essay we have introduced the concept of demand for internet usage. The demand for internet usage is influenced by the price of usage, cross-price effects, income effects and the network effect. The network effect acts to shift the demand curve for internet usage without altering its slope. Where the marginal benefit of an increase in network size is positive, this will result in an increase in demand. Where the increase in size results in a negative marginal benefit, this will result in a decrease in demand. The direction and strength of the relationship between network size and demand can be summarised by the use of the network elasticity.

Endnotes 

1. Cross price effects refers to the "impact of a change in the price of one good on the quantity demanded of another good." (Katz & Rosen (1994:63)). In general terms for two goods x₁ , x₂ with prices p₁ , p₂, good 1 is a substitute for good 2 if ((Dx₁ / x₁)/(Dp₂/ p₂) ) > 0. For compliments ((Dx₁ / x₁)/(Dp₂/ p₂) ) < 0 .