The governance of telecommuters: an agency and transaction cost analysis.

Brice, Jeff, Jr. ; Nelson, Millicent ; Gunby, Norris White, Jr. 等

INTRODUCTION

This paper seeks to stimulate research about the managerial challenges of a constantly evolving bureaucratic workforce. Specifically, this work seeks to add to organizational theory by integrating transaction cost theory and agency theory with forecasted trends in the governance of telecommuting employees. Telecommuters, or "teleworkers", are employees that are allowed to perform organizational work from a residence, or other location, instead of reporting to a centralized office location (Siha & Monroe, 2006). While the two managerial theories have been applied in a large number of organizational settings and circumstances, there have been no prior attempts to integrate these theories to forecast the effects of uncertainty on the control of a telecommuting workforce. This paper proposes a managerial framework that incorporates transaction uncertainty (from transaction cost theory) and outcome uncertainty (from agency theory) as determinants of employee control. First, the two theories are described and then adapted to present the hierarchical governance of a telecommuting workforce. After which, propositions are developed about which managerial control mechanisms are to be used under varying degrees of transaction and outcome uncertainty, considering the gravity of the work to be produced and the level of trust in the employee.

LITERATURE REVIEW

Transaction-Cost Theory

Transaction costs are the costs of negotiating, monitoring, and governing exchanges between people (Williamson, 1975). In transaction cost theory, a main purpose of the organization is to reduce the overall costs of exchanging goods and services in the environment and the costs of supervising exchanges within the organization. In-house transaction costs are generally labeled as "bureaucratic" or "hierarchical" costs and are the costs that are germane to this analysis. Internal business exchange structures of interest here are those that concern the governance aspect of distance employees, commonly referred to as "telecommuters".

Determinants of Transaction Costs

The major determinants of transaction costs are identified as transaction uncertainty (Jones, 1987) and performance ambiguity (Ouchi, 1980; Jones, 1987).

Transaction uncertainty is associated with the extent that the employer-employee alliance is not consistent. This is primarily due to a lack of familiarity of the employer with the employee. The lesser the degree of familiarity with the employee, the more confirmation concerning the employee's work methods, attitude, and capability is necessary for the employer. Transaction uncertainty may decline over time, with augmented communication and interactivity, allowing the operational relationship of both parties to become less unpredictable (Jones, 1987).

Performance ambiguity is linked to the accepted hazard that an employee takes on when work is carried out for an organization (Ouchi, 1980). Specifically, the implicit exposure involves the costs that must be endured by the employee when work is delivered that does not have acceptable utility for the employer. The consequences suffered for sub-par work are typically a deficiency in the agreed upon remuneration for such work and/or the economic opportunity cost related with having to rectify the mistakes or recreate the deliverable(s) entirely. Accordingly, performance ambiguity refers to the employee's capacity to appraise work accurateness and to appropriately assess it's usefulness for the employer (Ouchi, 1980). It is a function of task specificity, employer assistance, and the employee's competence in the performance of said task. The higher the levels of task specificity, employer assistance, and employee competence, the lower the level of performance ambiguity (Ouchi, 1980; Jones, 1987).

Agency Theory

The goal of agency theory is effective control of employees using some form of monitoring (Eisenhardt, 1988). Agency theory approaches the problem of employee control as a matter of risk sharing. It implies that observability (employee monitoring) and outcome uncertainty (the probability of incorrect work) are the determinants of control; and, that there is an element of risk with regard to income streams in any control system (Eisenhardt, 1988, 1989; Kren & Kerr, 1993). The more an employee can be observed, the less risk (performance ambiguity) the employee has to assume. Because of monitoring, the employer is more knowledgeable about the work performed and, thus, uncertainty is decreased.

Outcome uncertainty is also decreased by the nature of the work involved (Eisenhardt, 1989). Standardized, routine outputs are negatively related to outcome uncertainty because the employer and employee know exactly what is to be produced and received. The more secure the employer is about the usefulness of the employee's work, the more certain is the employee's compensation. However, in instances where observability and/or product standardization is low, outcome uncertainty is high and the employee suffers the risk for producing the wrong outcomes. Thus, compensation is dependent on the organization's perceived utility of the outcomes produced. The consequences that the employer suffers when an employee produces low-utility output are a function of the import of the project assigned. If incorrect employee output causes a loss of future income for the employer, the economic ramifications (costs) are high.

Transaction Cost and Agency theories are complements of each other because performance ambiguity and outcome uncertainty are near equivalents. In prior research, performance ambiguity was conceptualized as the costs that client organizations incur to monitor and evaluate the performance of other parties in an exchange (Jones, 1987). Telecommuters may be conceptualized much like independent organizations or other separate entities. Traditionally, telecommuters maintain their own working schedules and have little physical interaction with the traditional organizational center (Bailey & Kurland, 2002). In this paper, performance ambiguity is related to outcome uncertainty as the economic opportunity cost (risk) incurred by employees under conditions of low observability and/or non-routine product expectations within the organization. Non-routine output increases the risk of incorrect employee performance and increases the likelihood of lost economic opportunities by the employer.

Outcome uncertainty is related, therefore, to the real or economic costs (risk) suffered by employers associated with telecommuting employees and the non-standard (or routine) character of their work.

The Transactional Nature of Internal Organization

A transaction is a bilateral exchange between two or more parties. In an employment transaction, the conditions for execution are determined and work is carried out according to procedures developed and/or agreed upon (Commons 1950). Williamson (1975) writes that all societal, financial, and political transactions bring about transaction costs. In transaction cost theory, organizations transform, in both capacity and internal alignment, in order to diminish transaction cost difficulties over time (Yarbrough & Yarbrough, 1987).

The origins of transaction costs may be viewed as transmission problems that are introduced into the exchange process of a market transaction (Williamson 1975). Transaction costs are the negotiating, monitoring, and enforcement costs that must be endured to facilitate a transmission between several parties (Klein, Crawford, & Alchain, 1978). The six main causes of transaction complications (Jones & Hill, 1988) are:

1. Bounded Rationality--Rational human behavior is determined within the bounded limits of each individual's knowledge and the cognitive facility to process information.

2. Opportunism--People tend to act according to what is aligned with their own self-interests. Opportunism occurs when one, or both, of the parties to a transaction seek to alter the provisions of an agreement after the fact.

3. Uncertainty / Complexity--There is extensive uncertainty and complexity in the business environment.

4. Small Numbers Trading Relationships--Reliance on a single supplier for any resource will likely result in opportunistic behavior on the part of the supplier. However, these types of trading relationships are commonly found.

5. Asset Specificity--Significant sunk costs in resources (assets) that have limited use outside of an individual business transaction.

6. Information Impactedness--In a typical business exchange, it is common for one party to the transaction to have more knowledge than the other(s).

Each one, or any combination of these sources, may bring about transaction complications and increase costs. All of these factors make business dealings risky and inefficient. For example, behavioral uncertainty on the part of an employee may increase the cost of enforcing contracts due to the possible lack of productivity brought about by agent opportunism. In order to prevent this situation, the employer must incur the increased cost of authoring comprehensive agreements that will restrain employee behavior under a diversity of situational contingencies (Arrow, 1974). Given that the employer is limited by the bounded rationality of their own cognitive experience and cannot plan for every contingency, increased transaction costs due to opportunism may still be apparent because of information impactedness in favor of the employee. In other words, the employee may find weaknesses in the contract and exploit these to their benefit. In addition, the threat of increased costs due to opportunism may be heightened if there is asset specificity or small numbers bargaining conditions, which further restrains the employer from seeking other agency alternatives due to the high switching costs involved (Pisano, 1990).

There is an economic benefit to the organization when the use of hierarchy is selected over the use of the market as a basis for employer/employee transactions (Jones & Hill, 1988). The use of hierarchy is highlighted by internalization of the workforce, which remains the primary work arrangement in business. Internalization is more efficient than the market because the more exacting the approach implemented, the more influence each party has over the other's actions. If an organization depended solely on the market to carry out its chief objective, each work deliverable produced would be the consequence of a convoluted bartering and monitoring routine that would have to be repeated every working day. By internalizing the workforce, the bartering expense is kept to a single employment search and offer situation; and, monitoring is made unproblematic by including it as an element of the employment agreement. Therefore, the costs of negotiation, monitoring, and exchange are decreased.

The most important advantage that an organizational employment agreement has over mutually exclusive market negotiations is that the employee consents to be compensated with fixed wages in exchange for suborning to the right of superiors to direct and monitor the employees work efforts, thereby, minimizing the likelihood for incidences of self-interest agent opportunism. Furthermore, hierarchy creates an atmosphere of trust much more apparently than a market can between negotiating parties with their inherent self-interest biases (Ouchi, 1980).

Agency theory makes it clear that transaction costs are, likely, still apparent even when organizations opt for bureaucracy instead of the market. The likelihood of agent opportunism continues to be a threat on any occasion that authority is entrusted to an agent (Shapiro, 2005). The employment of bureaucratic mechanisms in organizations routinely results in some control insufficiency of the work systems (Jones & Hill, 1988). The distribution of authority to subordinates may encourage them to create agendas that further their own self-interests in preference to organizational goals and objectives (Leibowitz & Tollison, 1980). This "moral hazard" inefficiency is characteristic of contractual business exchanges (Alchain & Woodward, 1987). The moral hazard in telecommuting refers to opportunistic shirking, stealing, and other problems that arise when employee actions cannot be viewed by appropriate supervisory personnel. The challenge, therefore, becomes to develop a control structure that provides the organization with the information necessary to discern, and eliminate, opportunistic activity.

Types of Control

Managerial control within organizations may be defined as the practice of developing standards, observing and assessing performance, judging outcomes, acknowledging accomplishments, and taking necessary action to cure deficiencies in the work performed (Hill & Jones, 1993). Most organizational functions are driven off course because individuals who are in charge choose not to behave in accordance with organizational goals and policies (Hofstde, 1978). Agency theory concedes that employees may maximize personal gain by choosing to freeload, embezzle, or otherwise avoiding executing occupational responsibilities (Shapiro, 2005). An indispensable phase in the planning of effective control systems, then, is to make certain that employee decisions correspond with organizational objectives. Agency theory distinguishes two such control methods, behavior control and output control (Eisenhardt, 1989).

Behavior control means that the employee is compensated with earnings at a fixed rate for displaying appropriate work behaviors and that employee work habits during performance can be scrutinized, generally in person, to insure congruence with organizational standards and policies (Eisenhardt, 1988). When output control is used, a worker's pay is dependent on the level of productivity which can be ascertained without the need for observation of the work being performed (Eisenhardt, 1989). With output control, reliance on the use of commissions or piece-rate compensation for performance is customary. Another type of control, clan control, involves self-regulating groups of workers who naturally express behaviors that are congruent with the socialized acceptance of shared organizational customs and values (Ouchi, 1979). Clan control is applied when behavior and output controls cannot be established. Behavior and output controls are of interest here.

In situations involving unfettered behavior control, the worker has no mandate to deliver specific goods or services. The agreement is that he or she will simply employ the required work methodology without deviation (Eisenhardt, 1988). Compensation and incentives are dependent on the demonstration of mandated behaviors through personal observation or written reports. Therefore, risk is diminished for the employee in relation to his or her income by strict adherence to the mandated behavioral methodology. Alternatively, under complete output control workers have a responsibility to deliver goods and services that meet clearly defined standards of acceptance. In cases of failure, the worker suffers by sacrificing rewards (income). Output control indicates that the worker accepts any and all of the effects for any inadequacy connected to the results produced and, therefore, accepts all of the risk involved (Baiman, 1982; Holstrom, 1979) for mistakes. However, the employer also suffers damage when the output from the worker is valuable and necessary to obtain organizational revenues. Consequently, risk exposure and actual damages are experienced by both sides involved in the transaction.

The most advantageous selection between the two options depends on a compromise concerning the cost of evaluating behaviors and the cost of assessing outputs while passing on the risk of the transaction to the employee. Under conditions of diminished surveillance, management can elect to buy information about the employee's behavior; thereby, utilizing a behavioral control approach by developing a structured notification apparatus that can closely monitor the employee's behavior. This notification apparatus might exist in several forms including the usage of incentive-based whistleblowing schemes, additional layers of management, supplementary reporting requirements, and improved fiscal control procedures. Basically, the development of a more restrictive working environment. These mechanisms may be costly, but the employees do not suffer the possibility that mandated behaviors may not yield expected rewards due to causes not within their control.

On the other hand, the organization can choose to employ output control. Agency theory contends that output controls during periods of environmental uncertainty are risky for employees because outcomes are primarily a consequence of their actions and of the prevailing state of affairs (Baiman, 1982). Therefore, a worker may be penalized for inadequate outcomes that are the result of, at least partially, circumstances outside of their control. Due to the perilous nature of this position, organizations may find that they must compensate workers at the highest levels since a significant portion of the risk in the employment arrangement has shifted to the employee. Consequently, it has been previously demonstrated that it is more cost-effective for firms to mitigate this risk by compensating workers by means of fixed salaries than it is to pay high commissions or piece-work rates that recompense for employee acceptance of the risk described (Basu, Lal, Srinivasan, & Staelin, 1985).

To sum up, it is deemed better to implement behavioral controls through a structured notification mechanism when specifics about worker behavior is deficient and when the transaction cost of assembling these details is less expensive than discerning the actual utility of outputs produced.

Trends in Telecommuting

The concept of telecommuting is but one example of what organizations are doing to fight the high costs associated with doing business. By definition, telecommuting is the practice of performing work outside of the traditional workplace while sustaining communication via computer-based means (Wright, 1993). Specifically, work is carried out in locations that are isolated from central offices or production facilities and where employees have no personal contact with co-workers. Typically, there are three types of work locations where telecommuters perform their duties: 1) Satellite work centers, 2) Neighborhood work centers, and 3) Home-based work centers (Olson, 1982). Satellite work centers house groups of employees from a single business and are located a distance from the employee's homes and conventional workplace. Satellite office space is usually selected in an area that is convenient to the employees and/or prospective customers. Neighborhood work centers are identical to satellite work centers with the exception that they are designed to house employees from several businesses at the same time. The most common form of telecommuting occurs within home-based work centers. A home-based work center exists whenever an employee routinely works from their own home. It is estimated that over 20 million workers in the United States telecommute, with the largest portion of this workforce choosing to work from their own homes (Kirk & Belovics, 2006).

Telecommuting is a growing phenomenon that is forecasted to become more of an operational norm in years to come. In the 1990's, an estimated 42% of companies in the United States used telecommuting to cut employee commutes, free office space, offer family friendly flexibility, and to save the company money (Girard, 1997). A more recent study found that the practice of telecommuting is still growing in that 19% of all federal government employees were engaged in some form of routine telecommuting to do their jobs (Arnold, 2006).

How effective are these home-office employees? The National Academy of Science conducted a test in 1985 that showed telecommuting raised productivity by 15% to 25% (Cote-O'Hara, 1993). Other studies have been known to show normative productivity gains from 10% to 40% (Nie, 1999). These productivity gains are largely the result of the conversion of automobile commuting time that is transferred to work time. Other reasons are that telecommuting employees eliminate typical office distractions during the workday (such as personal conversations) and they usually work longer hours than other types of employees (Mello, 2007). Telecommuters generally have more energy than other employees because they can choose to work during the times of day (or night) when their energy levels are highest (Cote-O'Hara, 1993). Also, absenteeism predictably declines because employees can choose to work part-time on those days that they normally would have missed due to illness (Cote-O'Hara, 1993).

The promise of telecommuting offers societal benefits that may ultimately have far greater consequences than increased business productivity. As the population of telecommuters grows, this should result in an inverse decrease in the overall number of traditional automobile commuters. A decrease in the number of conventional commuters should also result in less pollution, less traffic, fewer accidents, and decreased travel-related costs for the worker. A more comprehensive interpretation of telecommuting benefits may envision increased independence from fossil fuels, fewer health-related costs, a healthier (greener) environment, and a redistribution of government investment from infrastructure development and maintenance to helping solve the more critical societal ills (poverty, hunger, education, etc). In essence, the societal benefits of telecommuting may contain partial answers to reinvigorating the economic and competitive strength of the populous.

This premise of this paper is one that visualizes the future workplace as one where employers and employees rarely interact physically. While telecommuting does decrease some bureaucratic costs, such as extensive fixed asset investment in real estate, the question of what type of managerial control should be employed, and under what circumstances, with these satellite employees is an issue that must be resolved.

The Governance Structure of Telecommuting Organizations

Transaction cost theory tells us that the costs of economic activities are not technologically determined but are dependent on the form of organization under which the activities are conducted (Yarbrough & Yarbrough, 1988). Therefore, if an organization wants to structure itself so that it can implement a behavioral control information system, it should and it will. In this case, management may surmise that the increased expense of tight behavioral monitoring is worth it to ensure that the required work methodology is being executed without deviation. Likewise, if management wants to concede observability as a lost cause, an output control system will likely be maintained. It is rational to assume, however, that, in these cases, the nature of products or services produced under an output control environment will be such that delayed acceptance until reasonable quality can be achieved will not place the firm in a position of undue stress. Therefore, it is of no consequence that telecommuters may utilize higher levels of technology to remain in the privacy of their own homes, or work in satellite locations away from the corporate center, because transaction cost theory claims that the resource requirement to implement this strategy is only a secondary consideration. More importantly, organizations will employ a telecommuting strategy because the internal structure of the firm will naturally evolve to select a strategy and governance structure that minimizes total bureaucratic costs (Coase, 1937). So, the question naturally becomes, how can managers implement a behavioral control system with a workforce that is located in physically remote locations; and, when is the total economic cost of obtaining behavioral information in the circumstance of telecommuting less costly than exposing the firm to the inherent risks involved in the utilization of output monitoring?

Behavioral Controls and Telecommuting

There are six primary methods that managers commonly utilize to control the behavior of telecommuters (Fairweather, 1999; Festead, Jewson, & Walters, 2003). They are summarized as followed:

1. Use of Technological Monitoring Devices--Computer Based Performance Monitoring (CBPM) systems are software programs that collect and transmit a wide variety of data about the productivity and work patterns of employees from their computer (Fairweather, 1999). CBPM allows managers to easily, and constantly, access and monitor data collected on all of the employees networked into the system. Typical uses include monitoring the number of data entry keystrokes per hour for each employee, the length of telephone calls and time between calls, the time an individual spends away from his or her desk, the time to process claims, and to record total work hours logged per day, among other things.

2. Monitoring of Telephone Communications--The recording and/or monitoring of telephone communications has been an accepted practice for many years in a wide variety of industries. Reasons that telephone conversations may be monitored include quality control, curtailment of personal use, and organizational privacy concerns (Fairweather, 1999) .

3. Electronic Mail Monitoring/Archiving--Firms routinely monitor and archive electronic mail transmissions to deter employee abuse of organizational time and resources by conversing with friends and relatives. For business related use, other reasons may include monitoring for illegal or fraudulent purposes, divulging trade secrets, and inappropriate relations with organizational stakeholders (Fairweather, 1999).

4. Use of Video and Audio Recording Devices--A typical use of video and audio recording devices in telecommuting is to help combat feelings of isolation formed in telecommuting employees by allowing them to correspond with the main office by means of teleconferencing (Fairweather, 1999). The two-way visual and audio interaction of teleconferencing, if performed on a routine basis, allows the telecommuter to retain visibility, develop personal relationships with their peers, and preserve the perception that they are a part of the team. Alternatively, cameras and microphones may also be installed around the entire work area of the employee and used as a means for comprehensive behavioral surveillance. In this fashion, managers may monitor and record every movement and utterance of employees while they go about the performance of their organizational responsibilities. In essence, these devices may allow the manager the highest level of intimate knowledge about the working behavior of employees because the monitoring may be done at any, or all, times work is supposed to be performed and without detection.

5. Home / Satellite Office Visits--Managers use announced, and unexpected, visits to telecommuter work locations for a variety of legitimate business reasons which includes health and safety assessments, performance appraisals, coaching, training, sickness monitoring, feedback, etc. However, personal visits also give managers the opportunity to inspect the demeanor, attitude, and condition of the worker. Many workers (and their families) find unannounced personal visits to be somewhat intrusive but some managers find it imperative to observe employees up close (Festead, Jewson, & Walters, 2003).

6. Trust--Trust exists when telecommuting workers are recognized by management for possessing the capabilities of individual self-discipline, self-motivation, and self-management. It entails a relationship, between selected workers and management, of shared ethical and business commitment in which rewards, productivity, and respect are exchanged without investigation. In practice, trust functions as a form of control and not an alternative to it. When organizations emphasize trust as a component for eligibility to be allowed to telecommute, the employees will be selected because, based on historical performance with different levels of responsibility, management has discerned that there is no need for high levels of observational scrutiny. Generally, higher grade workers who have clear records of proven reliability identify the archetype of a telecommuter selected on the basis of trust (Festead, Jewson, & Walters, 2003).

Each of the aforementioned behavioral monitoring strategies increases the short-term organizational costs of business transactions in terms of resource acquisition (purchase and installation of monitoring devices), labor expense (personnel cost to monitor telecommuters), travel expense (related to site visits), and opportunity costs (when trust is unfounded and the business is exposed to the cost of risk). However, under transaction cost rationale, these costs are justified because they serve to lessen transaction complications and, thereby, decrease the long-term costs of regular exchanges (Hill & Jones, 1993). For example, portable teleconferencing systems (e.g., laptop enabled with web-cam, microphone, and a portable bluetooth-enabled phone) can be taken to meetings, seminars, and other corporate functions and allow managers to listen in and interact with the telecommuting employee in the field so that they can know exactly what occurred and correct any mistakes in real time (decreases uncertainty and information impactedness). Organizations may also choose to conduct comprehensive surveillance with installed video and audio devices in employee work areas in order to make continuous appraisals of employee performance (decreases uncertainty and opportunism). Additionally, there is time-tracking software that can let managers know how much time an employee has put into certain projects (CBPM) and electronic bracelets that can pinpoint the exact location of an employee out of the office (decreases uncertainty and opportunism). When compared to the cost of maintaining office buildings (decreases asset specificity) and hiring managers to supervise employees on location (decreases opportunism), these behavioral monitoring devices are cost-efficient and readily available.

How do managers feel about telecommuting? Many managers oppose the arrangement, dreading the supposed loss of control over, and access to, employees and perceived reduction in their own value to the company (bounded-rationality). In most organizations, telecommuting requires a fairly major change in culture (Cote-O'Hara, 1993).

Output Controls and Telecommuting

Organizations employ output controls to monitor their telecommuting employees mainly because, considering the costs of behavioral monitoring strategies, they are deemed as unobservable (Kurland & Cooper, 2002). In other words, the assumption is that the cost of behavioral monitoring is prohibitive to these firms so observation of the telecommuter is not a viable option for them. For these businesses, the organizational culture has evolved from managing by activity (direct observation of the work in progress) to managing solely by production results. Agency theory argues, however, that output monitoring is more costly than effecting behavioral controls (Basu, Lal, Srinivasan, & Staelin, 1985) since workers must be compensated to take on the risk of possibly producing low utility output. Therefore, the employer must regularly pay a premium for the employee's services. When the employee delivers unacceptable output results on a project that the employer needs to secure income, the employer bears the penalty of missed economic gain. This is the cost (risk) of outcome uncertainty that is of interest here. However, in cases of standardized and routine output risk is negligible because output usefulness is guaranteed. Transaction costs decrease as employee and employer risks decrease. Thus, under the transaction cost theory, output control may not be advantageous under all situations.

Characteristics of Bureaucratic Transactions

In order for the employer-employee bureaucratic agreement to take place, both parties must benefit from it. Reducing the bureaucratic costs increases the mutual rewards each party may gain from the transaction. Therefore, the purpose of investigating the employer-employee interaction from agency and transaction cost standpoints is to discern those dimensions of the exchange that may lead to risk and increased costs. Two transaction characteristics, transaction uncertainty, from the transaction cost perspective (Ouchi, 1980), and outcome uncertainty, from agency theory (Holstrom, 1979), account for the amount of negotiating, monitoring, enforcement and risk-related costs built-in between employers and, in this case, telecommuting employees. Both of these characteristics have an effect on the scale of information that must be managed to conclude such transactions. For this paper, transaction uncertainty represents those elements leading to costs for the organization to monitor the behavior of telecommuters, while, outcome uncertainty represents those costs for the organization related to the nature of employee work (standard or non-routine) and the inherent risks related to producing this work.

An Organizational Control Framework for the Management of Telecommuters

The determinants for organizational control of telecommuting employees are described by varying levels of transaction and outcome uncertainty while considering the utility of the work performed for the employer.

Transaction Uncertainty and Telecommuters

Transaction uncertainty may be interpreted as the degree of familiarity that the employer has with an employee's work habits, reliability, and performance (Jones, 1987) and exists to the degree that telecommuting employees are not physically observable. However, the organization must have some evidence that employees are performing those duties which they were hired to perform. If not, workers will remain free from scrutiny and agent opportunism tendencies may surface. To alleviate this risk, the organization should take steps to avoid it

Principally, the organization must make an effort to dissuade telecommuting employees from, among other things, farming out sensitive work, shirking during the workday, and misrepresenting the company to clientele. As the tendency for unsupervised agents is to behave opportunistically (stealing, cheating, selling sensitive information, etc.), organizations must guard themselves against possible security breaches (Shapiro, 2005). Therefore, behavioral monitoring is indicated. This rationale implies, however, that the work output from telecommuters has a fairly high utility value for the employer. In cases where the utility value is not significant, output controls may be established because the risk assumed by the employee is minimal. As risk for the employee decreases, organizational costs for those risks assumed decreases making output control less expensive than behavioral monitoring.

In addition, the level of transaction uncertainty is a function of infrequency. The more repeated the exchange between an employee and an organization, the lower the uncertainty and cost of transactions (Jones, 1987). Initial transaction uncertainty will decrease with extensive behavioral monitoring. Over a period of time, however, extensive monitoring will not be necessary. This is because when exchange is frequent, the parties to transactions become used to dealing with each other and rely more on past experience (Williamson, 1975).

In summary, transaction uncertainty will decrease with behavioral monitoring and is cost efficient under conditions of high work utility and/or high employee assumed risk. The level of behavioral monitoring, however, will decrease with an increase in transaction frequency over time, due to a familiarity of the transacting parties.

Outcome Uncertainty and Telecommuters

Outcome uncertainty refers to the nature of the work performed (standard or non-routine) which affects the amount of risk assumed, by the telecommuting employee, for producing low-utility output (Eisenhardt, 1989). Increased employee risks are positively related to increased bureaucratic costs (Jones, 1987). Outcome uncertainty is heightened by non-standard outputs because they are unique and dissimilar. Examples of non-standard outputs are business plans, consulting services, medical advice, and contract proposals. There is variation in the accomplishment of the work, as well as its usefulness to the employer. If the telecommuter, acting as an agent for the employer, incompetently negotiates a project price for work that will result in a loss for the employer, then it is the employer who will lose profits on the deal because of being legally bound to deliver as per the contract agreement. Behavioral monitoring will reduce this risk of potential lost income. Alternatively, if the type of output expected is of low utility and standardized, one should use pure output monitoring (without paying a premium) because the risks are low for the employer and employee. Therefore, output monitoring may be a less expensive option.

A related workforce drawback forwarded by managers is the impact of telecommuting on both formal and informal communications (Gibson, Blackwell, Dominicis, & Demerath, 2002). The significance of such dialogue to the successful conclusion of assignments is critical. The lack of recurrent advice and criticism (formal and informal) from supervisors and colleagues can cause confusion, needless errors, and duplication of work (Ford, 1991). This lack of communication increases outcome uncertainty because high utility outputs cannot be guaranteed without strict direction. Behavioral monitoring may decrease this uncertainty by increasing communication.

In summary, outcome uncertainty is increased for work that is non-standard in nature due to the risks that the employer might suffer for high utility work not being produced adequately.

PROPOSITION CONSTRUCTION

Based on the review of literature previously, a framework of governance possibilities for daily telecommuters may be constructed by integrating the employer-employee relationship variables transaction uncertainty, outcome uncertainty, and work utility (See Figure 1). In review, transaction uncertainty is defined as the level and degree of familiarity that employers have in their employees (Williamson, 1975; Jones, 1987). Outcome uncertainty concerns the risk that employees and employers suffer in case the outputs delivered are unacceptable. This risk is tied into the nature of the work performed--there is more outcome uncertainty for non-routine (or customized) outputs than for simple, routine, or standardized outputs (Eisenhardt, 1989). Finally, work utility is a function of the importance of the outputs produced to the organization (Eisenhardt, 1989). It may be surmised that it is rational for organizations to ensure the quality of high utility outputs by taking precautions. In essence, more control may be necessary when the firm depends on the delivery of high utility outputs.

In CELL ONE, transaction uncertainty, outcome uncertainty, and work utility all assume low levels. Therefore, there is familiarity with the employee and his or her general work habits, the nature of the output expected is routine, and the importance of the product or service is low. Under these conditions, the firm should not invest in technology to monitor the telecommuter because, since there are no risks assumed by the employer or employee, costs will remain low. The firm should, therefore, institute an output control system and compensate the employee at standard rates for the work performed.

In CELL TWO, transaction uncertainty and work utility remain at low levels but outcome uncertainty is high. Specifically, the employer is familiar with the employee's general work habits but the outputs required are non-routine (specialized). Under these circumstances, the firm should implement behavioral control utilizing technology as an alternative to implementing output control and compensating the employee for the risks assumed due to the probability of producing unacceptable work. In other words, it will be less expensive over the long-term for the employer to interact with the employee and provide direction than it will be to continually pay a premium for work under an output control setting.

In CELL THREE, transaction and outcome uncertainty is low but work utility is high. The employer is familiar with the telecommuting employee and his or her work habits; and, the nature of the work produced is routine. However, work utility is high meaning that the work is important for future company income. Therefore, the choice between control mechanisms rests on the trust that the employer has in the employee. The firm should institute behavioral monitoring using technology if there is any question about employee competence, honesty, or timeliness. However, if agent opportunism is not an issue (based on the historical reliability of the worker) an output control mechanism (no premium) will be less expensive.

In CELL FOUR, transaction uncertainty is low but outcome uncertainty and work utility is high. In other words, the worker's habits are well known but the product or service is valuable to the firm and it is of a customized nature. This condition will require the use of behavioral monitoring so that the firm can be assured that the work is being performed correctly and that it will be done on time.

In CELL FIVE, transaction uncertainty is high, outcome uncertainty is low, and work utility is low. This cell describes a situation where the employer is not familiar with the employee's work habits or abilities; but, the output expected is routine and of little importance. This condition requires the implementation of an output control strategy because, since the output is routine, there are no assumed risks for which the employer has to pay a premium. In the case of employee failure, it is more cost effective, under these circumstances, to cancel the employer-employee agreement and to hire someone else rather than invest in technology that would allow the employer to interact with the telecommuter.

In CELL SIX, the employer is unfamiliar with the employee's work habits and abilities, the work performed is of little importance, and the work is non-standard. Since it has been demonstrated that behavioral monitoring may be less expensive than paying output control premiums for employee assumed risk (Basu, Lal, Srinivasan, & Staelin, 1985), behavioral monitoring using technology is indicated. However, if the cost of behavioral monitoring is more than the cumulated cost of paying higher premiums, output controls will be established.

In CELLS SEVEN AND EIGHT, behavioral monitoring will be implemented because of the importance of the work and the unfamiliarity of the employer with the employee. Outcome uncertainty is not relevant, in these cases, because there is no reason to trust an unfamiliar remote employee to work on vitally important projects without some direction from management, regardless of the nature of the work (standard or non-routine). In this state of affairs, the risk of forgone income to the employer is best mediated by behavioral monitoring.

Propositions

From the framework above, the following propositions are theorized:

Proposition 1: Low transaction and outcome uncertainty is positively related to output control of telecommuting employees for low utility work.

Proposition 2: Low transaction uncertainty and high outcome uncertainty is positively related to behavioral control of telecommuting employees regardless of the utility of the work.

Proposition 3: Low transaction and outcome uncertainty is positively related to behavioral control of telecommuting employees in cases of high work utility and a lack of trust for the employees.

Proposition 4: Low transaction and outcome uncertainty is positively related to output control of telecommuting employees in cases of high work utility and high trust for the employees.

Proposition 5: High transaction uncertainty and low outcome uncertainty is positively related to output control of telecommuting employees in cases of low work utility.

Proposition 6: High transaction and outcome uncertainty is positively related to behavioral control of telecommuting employees in cases of low work utility if the cost of monitoring technology is less than the cumulated payment of output control premiums.

Proposition 7: High transaction uncertainty and high work utility is positively related to behavioral control of telecommuting employees regardless of outcome uncertainty.

DISCUSSION & CONCLUSION

This paper contributes to organizational theory and management practice by integrating transaction cost and agency theories to attempt to forecast the impact of telecommuting on the governance practices and structure of organizations. The impending role of the satellite workforce is one that is changing as technology further evolves. Behavioral control has been identified as the most efficient method to insure the quality of outputs and monitor employees. It is currently possible to maintain control of telecommuters by using readily available technological devices. However, the cost of these behavioral controls can be prohibitive, depending on several work characteristics, and in those instances output controls may be most appropriate.

While the list of new technology that makes it possible to observe the behavior of employees and interact with them is expanding, the fear that these devices will be used to abuse the workforce is rising also (Fairweather, 1999). There is a possibility that "micro-managers" will attempt to implement overly intrusive management systems creating, in effect, a 24 hour "electronic bureaucracy" from which there is no escape. It is possible.

Although the scope of this paper was limited to a partial analysis of hierarchical governance choices for a telecommuting workforce, the investigation did not take into consideration other determinants of control such as professionalism, task specificity, task interdependence, incentive programs, organizational culture, specific industry application, etc. The areas for future research and speculation are broad and the implications are numerous.

In summary, based on environmental and business trends, a significant and growing percentage of the future workforce might be one where telecommuters work from their homes, or other satellite offices, and are controlled by organizations, through technology, employing behavioral or output controls. This paper is an attempt to discern under what conditions organizations may favor one control strategy over another.

REFERENCES

Alchian, A., & Woodward, S. (1987). Reflections on the theory of a firm. Journal of Institutional and Theoretical Economics, 143(1): 110-136.

Arnold, J.T., (2006). Making the leap. HR Magazine, 51(5): 80-86.

Arrow, K. J. (1974). The limits of organization. New York: Norton.

Bailey, D., and Kurland, N. (2002). A review of telework research: findings, new directions, and lessons for the study of modern work. Journal of Organizational Behavior, 23(1): 383-400.

Baiman, S. (1982), Agency research in managerial accounting: A survey. Journal of Accounting Literature, 1(1): 124-53.

Basu, A., Lal, R., Srinivasan, V., & Staelin, R. (1985). Sales force compensation plans: An agency theoretic perspective. Marketing Science, 4(1): 267-291.

Coase, R. The nature of the firm. Economica, 4(1): 386-405.

Commons, J.R. (1950) The economics of collective action. Madison, WI: University of Wisconsin Press.

Cote-O Hara, J. (1993). Sending them home to work: Telecommuting. Business Quarterly, 57(3): 104-109.

Eisenhardt, K. (1988). Agency and institutional theory explanations: The case of retail sales compensation. Academy of Management Journal, 31(1): 488-511.

Eisenhardt, K. (1989). Agency theory: An assessment and review. Academy of Management Review, 14(1): 57-74.

Fairweather, N.B., (1999). Surveillance in employment: The case of teleworking. Journal of Business Ethics, 22(1): 3949.

Felstead, A., Jewson, N., & Walters, S., (2003). Managerial control of employees working at home. British Journal of Industrial Relations, 41(2): 241-264.

Ford, R. C., (1991). Is your organization ready for telecommuting. SAM Advanced Management Journal, 56(4): 19-23.

Gibson, J.W., Blackwell, C.W., Dominicis, P., & Demerath, N., (2002). Telecommuting in the 21st century: Benefits, issues, and a leadership model which will work. The Journal of Leadership Studies, 8(4): 75-86.

Girard, K., (1997). Ditching the commute. Computerworld, 31(35): 1-3.

Hill, C., & Jones, G. (1993). Strategic management (2nd ed.). Boston, MA: Houghton Mifflin.

Hofstede, G. (1978). The poverty of management control philosophy. Academy of Management Review, 2(1): 450-461.

Holmstrom, B. (1979). Moral hazard and observability. Bell Journal of Economics, 10(1): 74-91.

Jones, G.R. Organization -client transactions and organizational governance structures. Academy of Management Journal. 30(2): 197-218.

Jones, G.R. and Hill, C., (1988). Transaction cost analysis of strategy-structure choice. Strategic Management Journal. 9(2): 159-172.

Kirk, J., & Belovics, R. (2006). Making e-working work. Journal of Employment Counseling, 43(1): 39-46.

Klein, B., Crawford, R., and Alchain, A., (1978). Vertical integration, appropriable rents, and the competitive contracting process. Journal of Law and Economics. 21(1): 297-326.

Kren, L., & Kerr, J. (1993). The effect of behavior monitoring and uncertainty on the use of performance contingent compensation. Accounting & Business Research, 23(1): 159-168.

Kurland, N., Cooper, C. D., (2002). Manager control and employee isolation in telecommuting environments. Journal of High Technology Management Research, 13(1): 107-126.

Leibowitz, A. and Tollison, R., (1980). Free riding, shirking, and team production in legal partnerships. Economic Inquiry, 18(1): 380-394.

Mello, J.A., (2007). Managing telework programs effectively. Employee Responsibilities & Rights Journal, 19(4): 247261.

Nie, N., (1990). Tracking our techno-future. American Demographics, 21(7): 50-53.

Olson, M. (1982). New information technology and organizational culture. Management Information Systems Quarterly, 6(4): 71-92.

Ouchi, W.Z., (1979). A conceptual framework for the design of organizational control mechanisms. Management Science, 25(4): 833-47.

Ouchi, W.Z., (1980). Markets, bureaucracies and clans. Administrative Science Quarterly, 25 (1): 129-41.

Pisano, G. (1990). The R&D boundaries of the firm: An empirical analysis. Administrative Science Quarterly, 35(1): 153-176.

Shapiro, S. P., (2005). Agency theory. Annual Review of Sociology, 31(1): 263-284.

Siha, S., and Monroe, R., (2006). Telecommuting's past and future: a literature review and research agenda. Business Process Management Journal, 12(4): 455-482.

Williamson, O. (1975). Markets and hierarchies. New York: Free Press.

Wright, P.C. (1993). Telecommuting and employee effectiveness: Career and managerial Issues. International Journal of Career Management, 5(1): 4-9.

Yarbrough, B.V. and Yarbrough, R.M., (1988). The transactional structure of the firm: A comparative survey. Journal of Economic Behavior & Organization, 10(1): 1-28.

Jeff Brice, Jr., Texas Southern University

Millicent Nelson, Middle Tennessee State University

Norris White Gunby, Jr., Elon University

Figure 1: Framework for Organizational Control of Telecommuting
Employees

                            Work Utility

                  Low          High          Low          High

Outcome        Output       Output or     Output       Behavioral
Uncertainty                 Behavioral
(Low)          Cell 1       Cell3         Cell 5       Cell 7

Outcome        Behavioral   Behavioral    Output or    Behavioral
Uncertainty                               Behavioral
(High)         Cell 2       Cell 4        Cell 6       Cell 8

                Transaction Uncertainty    Transaction Uncertainty
                         (Low)                    (High)