Diffusion of Technology Flashcards
What does the term “diffusion of technological innovation” mean?
“Diffusion of technological innovation” refers to the process by which successful innovations (be they new products, processes or management methods) spread within and across economies, wholly or partly displacing the existing ‘inferior’ varieties. It often follows a sigmoid (S-shaped) path but different technologies might have different diffusion patterns.
What is the difference between the terms “adoption” and “diffusion”?
The difference between adoption and diffusion relies on the different character of the related studies. Adoption studies are concerned with reasons for adoption at one point in time, or the reasons for time of adoption of individual users. Diffusion studies focus on the market behavior of the diffusion process over time – i.e. it is the study of a dynamic, aggregative process over continuous time.
What are the key elements of the diffusion process? (Mahajan et al., 1990)
As theorized by Rogers (1983), the key elements of a diffusion process are:
a. The presence of an innovation
b. A channel through which communication flows (mass media and/or interpersonal communications)
c. Time
d. The social system (a set of interrelated units engaged in joint problem solving to accomplish a common goal)
Name four types of theories traditionally set forth to explain the S-shaped nature of diffusion processes.
The 4 main theories explaining the S-shaped nature of diffusion processes are (as mentioned in Botelho and Pinto, 2004):
a. Epidemic: demand-induced. As information about the existence of a new technology spreads across individuals, new people adopt the technology. (Griliches, 1957)
b. Expected profitability of the innovation at firm level, and dissemination of information about its characteristics (Mansfield, 1961)
c. Communication-based model. According to Rogers, communication is “a process in which participants create and share information with one another in order to reach a mutual understanding”. Moreover “diffusion is a very social process that involves interpersonal communication relationship” (Rogers, 1983). According to Rogers, interpersonal channels of communications are stronger than mass media.
d. Network-consumption externality: value of the network for subscriber increase with the number of adopters of the system.
Who are the actors involved in the process of diffusion of a technology? (Rogers, 1995)
Besides the innovators, the other actors involved in the process of diffusion of a technology are the adopters. Depending on the timing of adoption of the technology, these can be divided into early adopters, early majority, late majority, and laggards. The distribution of adopters on the base of innovativeness follows a normal distribution.
What influences the dynamics of adoption of an innovation from a sociologist’s perspective? (Rogers, 1995)
According to Rogers, the dynamic of adoption is influenced by 5 factors, namely:
a. Relative advantage (subjective perception, e.g. of economic profitability, increased comfort, social prestige, time and effort saved, and immediacy of rewards)
b. Compatibility (consistency with existing values and social norms, past experiences and needs)
c. Complexity (degree to which an innovation is subjectively perceived to be difficult to understand and use)
d. Trialability (possibility for experimentation on a limited basis)
e. Observability (degree to which the results of an innovation are visible to others)
What is the intuition behind epidemic diffusion models, and in particular the Bass Model? (Bass, 1969)
The intuition behind epidemic diffusion model is that diffusion results from the spread of information. In other words, as soon as potential adopters become aware of the existence of an innovation they adopt it. At every point in time, the higher the number of adopters, the faster information is spread. With particular reference to the Bass model, this is considered to be epidemic because the diffusion process is positively affected by the number of adopters who transfer information to imitators via the channel “word of mouth”. A second information channel in the Bass model is represented by mass media, whose action has a positive effect on adoption on the side of innovators. While imitators “learn” from those who have already bought, innovators are not affected by the number of previous buyers. Over time, the importance of innovators decreases.
What is the formula of the Bass model? What is the meaning of the parameters p and q?
Where f(t), F(t) represent the non-cumulative and cumulative proportions of adopters respectively and n(t), N(t) represent the non-cumulative and cumulative number of adopters.
f(t) and n(t) are first-order differential equations. If n(t) is integrated we get the S-shaped curve
represented by N(t).
The assumption at the base of the model is that potential adopters of an innovation are influenced by two means of communication: mass media (p, external influence) and word of mouth (q, internal influence).#
In statistics, f(x) is a probability density function (pdf) describing the probability of a random variable to
take a certain value. In our case, the random variable is the dummy variable for innovation adoption so
the pdf expresses the probability of the innovation to be adopted at time t given that it has not yet occurred.
The probability of adoption depends on the proportion of those who have already adopted
the technology.
So the conditional probability of adoption at time t is positively dependent on the fraction of the population that has already adopted (=adoption depends on imitation and the parameter q reflects this influence).
This means that the Bass model is an epidemic model, where the contagious process start with the spontaneous adoption of consumers who positively react to mass media coverage.
More in particular, the parameter p also represents the probability of an initial purchase at T=0 and reflects
the importance of innovators in the social system. The parameter q represents the coefficient of imitation,
reflecting the word of mouth or social contagion.
Several estimation procedures are available to estimate the parameters p, q and m. Examples of such
procedures are OLS and the Bayesian estimation procedure.
What is the meaning of and distinction between stock, order, rank and epidemic effects of diffusion? (Karshenas and Stoneman, 1993)
The meaning of the stock, order, rank and epidemic effects is as follows:
Stock effect: the benefit to the marginal adopter from acquisition decreases as the number of previous adopters increases. For any given acquisition cost, there is a number of adopters beyond which adoption is not profitable. However, the acquisition costs fall over time in order to make adoption more profitable at any given time. The mechanism behind the stock effect is as follows: as firms acquire new technology, their production costs fall thus affecting industry prices and therefore the profitability of further adoption. This approach is also called game-theoretic.
Firms ’ output decisions are being endogenized
The benefit depends on a firm’s output and firm’s output level depends on other firms’ adoption behavior
Moreover, for decreasing acquisition costs, further adoptions are assumed to take place.
Stock effect:
When a firm adopts a new technology one might expect that this reduces its cost,
which may lead to a price change for its products and thus in its output (and
other firms’ output
Benefits that firms will obtain from new
technology will depend on output level, and
firm’s output will also depend on
adoption behavior of other firms
comp. firm’s profits w or w/o new tech.)
Non using firm’s profits may fall
over time if other firms adopt new
technology
Order effects: return to adoption affected by order of adoption, i.e. high order adopters achieve greater return (because of first-mover advantages). For any given cost of acquisition, it is profitable to adopt an innovation until a certain adoption time. However, as adoption costs go down, this makes adoption more appealing.
Rank effects: potential adopters can be ranked in terms of their returns to adoption, which are a function of firms´ characteristics. There will be a distribution of reservation acquisition costs derived from the benefit distribution. Adoption of new technology happens as long as acquisition costs (falling over time) fall below reservation acquisition costs.
Epidemic effect: Assumption that information (and as a consequence technologies / innovations) spreads like infectious disease. It is a rather ‘naïve’ intuition because it implies that potential adopters decide to adopt as a consequence of personal contacts with earlier adopters (i.e. no active search). The higher the numbers of adopters, the greater the probability that a potential adopter meets an earlier adopter and adopts herself. Over time, number of non-adopters decreases and diffusion process slows down mapping of an S-shaped (sigmoid) curve.
Constant population of potential adopters of a technology
Users and non users mix socially homogeneously and make contact over time
Simplest form: non user becomes a user just by making contact with a user of a
technology
Over time, no. of users increases , so that given a constant mixing of the population the
chance of a non user meeting a user increases
Growth in the no. of users will map out an S shaped curve
Question: What happens when a non user meets a user that makes him/her an adopter
of a technology?
Common interpretation: transfer of information , non user learns about new techno
logy; given the acquired knowledge , non user decides to adopt the technology
Diffusion may take many years weak explanation that only reason is knowledge
According to the stock effect, through which channels does the number of (previous) adopters determine adoption benefits?
According to the stock effect, the channels through which the number of (previous) adopters determines adoption benefits are:
a. Decreasing acquisition cost (positive effect on benefits)
b. Decreasing benefit for marginal adopter as stock of adopters increases (negative effect on benefits).
A decrease in acquisition cost may result from learning effects, economies of scale, lobbying (subsidies, taxes, etc.) and facilitated funding.
Decreasing benefit for marginal adopter as stock of adopters increases mainly results from a decrease in the industry price (Reinganum´s specification of the stock model) or congestion (Schumpeter´s vision).
=> Change
in industry price with increasing use is key to understanding the stock
model
Reinganum‘s
specification of the stock model
Initially, all firms are the same (output, characteristics and cost), and all firms have perfect
information regarding the technology
For simplicity, but without much loss in generality, let us assume firms are myopic (makes
presentation easier), and that the new technology reduces the cost of the firm
Industry price is established at p(t) in time t where supply and demand intersect; when firms adopt new technology and generate lower (marginal) costs, the industry supply curve
will shift downward
Hence the position of the supply curve depends on how many firms use the technology in time t , M t ); as M t ) increases, industry price will fall
According to the order effect, in which ways does the order of adoption affect benefits?
The order of adoption has a positive effect on benefits through first-mover advantages (resulting from e.g. scarcity of suitable locations, decreasing public acceptance and increasing insurance premiums / safety regulation).
However, acting first might also have a negative effect due to e.g. technical improvements, learning effects and spillover effects (‘First-mover dilemma’).
Simple Order Model: NPV of adopting new technology is related to the no. of
other users at the date of adoption
At any moment in time, it will be only profitable for firms down to some point in the
adoption order to acquire new technology
In many ways similar to stock and rank models
Early adoption by one firm may have an impact on the preferred adoption date of
other firms
Appeal to model technology diffusion in private households : being “first on the
block” yields a utility gain
According to the rank effect, which types of firm characteristics can affect adopter’s benefits (merit order)?
An acquisition rule compares adoption benefits with costs. While adoption costs fall over time due to market prices and learning effects, adoption benefits depend on firms´ characteristics such as location, tax rate, discount rate, operating costs, plant type and size, etc. The effects of firms´ characteristics on adoption benefits might be both positive or negative.
Assume all firms / households know about existence of a technology
Population is of size N and heterogeneous (different members of the population gain
different benefits from adoption)
Gross benefit of adoption B, cost of acquisition at time t, c(t), for simplicity assumed to be the same for all adopters
Without uncertainty, decision rule is adopt if
m(t) is proportion of population for
whom this applies in time t
Hence level of ownership or use of
the technology is
Diffusion path requires that M t ) changes over time
Cost of acquisition c t ) changes (common
(perceived) gross benefits B t ) change
Decreasing cost and/or rising gross benefits create increase in the level of use (declining technologies rarely investigated)
As c t +1) < c t ) then no. of users increases to M t+1 ) = m t+1 N
Shape of diffusion path depends on the benefit distribution, how that distribution changes over time, and the time path of c t
Final level of use N is determined by where benefits and costs stop changing
Note that (
i ) given identical cost of acquisition at time t, earlier users get the greatest gross
benefit; (ii) different technologies, or same technologies in different populations, will show
different diffusion curves if benefit distributions or time path of acquisition costs are different
What is the “takeoff” time of an innovation?
‘‘Takeoff time’’ is defined as “the time in the life-cycle of an innovation when it transitions from the introduction stage to the growth stage” (Laciana et al. 2013). In other words, the takeoff time represents the change from a ‘‘desirable product’’ to a ‘‘popular product’’. Mathematically, this point corresponds to the time when the change in the derivative of the proportion of adopters reaches a maximum. It is, therefore, a useful macro-level indicator of the penetration of an innovation.
Schumpeter
and the stock effect
An entrepreneur sees a profitable opportunity for the use of a new technology and
pursues this opportunity by innovating
Innovation leads entrepreneur to generate excess profits
Profits act as a signal to other potential users
(in the full Schumpeterian story this is embedded in the context of increasing credit availability and macroeconomic spillovers)
As technology is used more widely, and output expands, there is an increasing demand
for (limited) input factors ; hence the price of such inputs increases, reducing the profits
of both users and non users
Increased price of input factors may lead non users to leave the market, or to find the
switch to the new technology more attractive
