V
Model System Estimation Results
In this chapter, the model system that represents the process of adoption of telecommuting is specified and estimated.
MODEL SYSTEM SPECIFICATION
The formulation of the model system that represents the adoption process is presented in this section. The models for the employer's and the employee's decision processes are presented separately.
The Employerâs Model
The modeling framework for the employerâs decision process is presented in Figure 5.1. This framework consists of two parts: the design of the telecommuting program and the decision to offer the designed program to the employees.
â PROGRAM DESIGN
In this first stage, the employer selects, from all possible telecommuting programs, the ones that can address her/his motivations, while satisfying existing constraints. In the framework discussed in Chapter 3 it is assumed that the employer's objective is to maximize profit. In that same chapter, telecommuting is defined as a combination of attributes, each of which can assume various levels.
Figure 5.1 - The Employerâs Decision Process -Modelling Framework
Given these assumptions, when designing a telecommuting program, the employer is searching for the combination of attribute values Ai that maximizes:
where A is the set of all possible combinations of the values of the attributes of a telecommuting program, Aj is the vector of attribute values that characterize program j, P is the price of the organizationâs product, and Ď, f, and c are, respectively, the profit, the production and the cost functions of the organization.
To model the employerâs choice of a program design, a probabilistic approach is adopted to account for the randomness due to lack of information about the organizationâs production function and cost structure, as well as for other unobserved factors. The following model is specified:
Structural Model:
Measurement Model:
where:
vector of attribute values of telecommuting program
i.
vector of interactions between telecommuting attributes and organization characteristics.
vector of structural parameters.
vector of structural parameters.
vector of random disturbance, distributed iid Gumbel (0,1).
set of all the telecommuting programs available to the employer.
vector of dummy variables indicating the chosen telecommuting program.
The subscript relating to the individual respondent is omitted for simplicity. Given this specification, the likelihood of observing the vector of choices dA, for a particular employer, given the set of all possible arrangements A and the organization characteristics O, can be written as:
where
is the set of structural parameters
The estimation of these unknown parameter is performed by maximizing the logarithm of this likelihood function over the whole sample. In practice, the outcome of the employerâs choice could be one or more telecommuting arrangements. For simplicity of the data collection process, it is assumed in this research that only one program is designed.
â PROGRAM OFFERING D...