A firm might want to quantify the relationship between sales volume, product
price, price
of substitutes and complements, national income, advertising
expenditure,
etc., for several reasons. For example:
- A tobacco company might want estimates
of the effect of an increase in
government
taxes on sales.
- An electricity company may be planning
to build a new power station and
would like to
know how demand for electricity is expected to develop over the
next 10 years.
- A brewery might want an estimate of the
effect of income growth and a
changing
age-structure of the population on the demand for beer.
- An automobile industry association
might want an estimate of the future
increase in
demand for imported automobiles as restrictions on imports are
relaxed.
- A computer firm might want to know the
consequences of lower price and
high national
income growth on the demand for microcomputers.
To derive this
information, it will be necessary to estimate a demand function for
the product concerned.
How to estimate
demand functions
There are six steps involved in estimating
a demand function and using it for forecasting purposes. First, the variables
likely to influence demand for the
product must
be identified. Own price, income and price of substitutes are
obvious
candidates, but there may be other specific factors influencing demand.
The next three
steps are technical in nature. They concern the form of the
demand
function, the statistical estimation techniques used and the derivation of
statistically reliable estimates of the parameters. Thus, in the equation in Box 4.2
product will
be affected by changes in the independent variables. This information
is often
presented in the form of elasticity estimates. Step 5 involves careful
evaluation of
the results, including comparison between the economic consultant’s
estimates and
those of other studies. If the objective of the study is to examine the
sensitivity of demand to a particular
variable, the exercise might end here. However, if the objective is to derive a
demand forecast, the final step –
step 6 – is to
construct a range of scenarios involving prices, income and other
variables such
as advertising and demographic changes, and to estimate the level
of demand
corresponding to each scenario.
Clearly, liaison is required between
management and the consultant in deciding
what variables
are relevant for the estimates of each good, how such variables
should be
measured, what statistics are relevant and for which period. Attention
should be
drawn to any unusual features of the data due to exceptional events
such as
strikes, adverse weather or changes in compilation methods. The statistical
procedure will
then generate estimates of the net effect of each variable on
quantity
demanded. In the case of elasticities, while their derivation is largely
technical,
evaluation will require dialogue and cross-checking with other procedures
such as
consumer surveys, marketing tests and the opinions of management.
At all times
it is important to bear in mind that econometric results are
highly
tentative and sensitive to the specification of the equation. Normally a
range of
estimates should be compiled on the basis of different assumptions
about the
future evolution of income and so on. If the company is the sole seller
in the market,
the market demand for the product will be the same as the
company’s
demand curve. If, however, as is usually the case, the company is one of
a number of
suppliers, additional analysis is required. The effects of a change
in the
company’s price will be influenced not by the nature of the market
demand curve
alone, but also by the reaction of competing firms to any change
in the firm’s
price.
Forecasting demand for a firm’s product is a necessary exercise,
but is fraught
with
uncertainty. The statistical analysis provided by economists helps to identify
the
structure and driving forces of market demand. Knowing these basic determinants
of demand, the manager is in a better position to react to a changing future environment.
But statistical analysis relies on historical market data. It should be supplemented
by qualitative analysis based on customer interviews, market
surveys,
managers’ views, regular monitoring of current sales volumes and trends,
and price
experiments. As an example of the last, Dolan and Simon cite the case of a
German mobile phone manufacturer who kept price constant in one region but
allowed it to vary upwards and downwards in other regions. By studying
the regional
response rates, this experiment enabled the firm to implement a
successful nationwide price
strategy.
No comments:
Post a Comment