Airline safety margins, maintenance expenditures, and myopic behavior: an empirical investigation.

Deppe, Larry A. ; Hansen, Don R. ; Swearingen, James G. 等

INTRODUCTION

The effects of deregulation of the airline industry have been profound. Eastern Airlines no longer exists, Continental Airlines and TWA filed for bankruptcy, and Pan Am has been politely carved up by the remaining gargantuan airlines that survive. The three healthiest domestic U.S. airlines, United, American, and Delta, controlled nearly half the market and all three were in the process of consuming the dismembered parts of disintegrating airlines (Dempsey, 1991).

Proponents of deregulation have cited more competition among airlines, lower ticket prices, and better service as advantages resulting from the 1978 Airline Deregulation Act. Opponents of deregulation have cited less competition among airlines, higher ticket prices, deteriorating service, and an erosion of safety as the effects of the 1978 Act. The effects of deregulation on safety are particularly important in view of the threat an erosion of safety could pose to the travelling public.

Proponents of deregulation argue that the Airline Deregulation Act of 1978 had no negative impact on the steadily improving safety record of U.S. air carriers. Bruggnick (1991) reports that U.S. air carriers experienced thirty fatal accidents during the period 1970-1979 while only 20 fatal accidents occurred during the period 1980-1989 (a decrease in fatal accidents of 33%). Aircraft occupant deaths for the 1970-1979 period were 2,088 while 1,438 aircraft occupant deaths occurred during the 1980-1989 period. The hours flown between fatal accidents increased from 1,930,000 during the 1970-1979 period to 3,970,000 during the 1980-1989 time period (an increase of 105%).

Opponents of deregulation have been concerned that increased competition would tempt airlines to reduce their commitment to safety by reducing time, money, and effort devoted to matters of safety.Henry A. Duffy (1986), a former president of the Air Line Pilots' Association, argues that "airline managers are pressed to cut costs to the bone in order to compete. They are being forced to decide between lower operating costs and maintaining their airline's safety status quo... The net effect has been a slow but steady erosion of the overall safety margins in the industry." Bruggnick (1991) supports this assertion by pointing out that while the best performance record of the industry occurred during the first five years of deregulation, this level of performance did not persist. Only five fatal accidents occurred during the period 1980-1984 and the hours flown between fatal accidents rose to 6,600,000 hours. He suggests that this commendable record occurred not as a result of but in spite of deregulation largely as a result of safety initiatives taken in the 1970s. Only one accident-free year (1986) occurred in the 1985-1989 period. The twenty fatal accidents of this period occurred evenly throughout the other four years (Bruggnick,1991). The number of hours flown between fatal accidents during this period declined from the 6,600,000 between 1980-1984 to 3,100,000 hours during the 1985-1989 period. There were five fatal accidents in 1985 alone involving aircraft with more than 30 seats (i.e., excluding most commuter airlines) (Gesell, 1990).

The concerns expressed by opponents of deregulation are not without some factual basis. USAir supervisors at two airports reportedly falsified records to cover plane repairs that were not done (Salt Lake Tribune, 1993). A USAir maintenance supervisor in Charlotte, NC, acknowledged that he allowed a jet to fly with a defective warning system--to save the airline money. The now-defunct Eastern Airlines was fined $3.5 million in 1991 after it was learned that its managers were forcing mechanics to falsify repair records to save money for the ailing company.

Reducing maintenance expenditures and the commitment to safety is an example of what is labeled in the management accounting literature as myopic behavior. Myopic behavior occurs when managers make decisions which, in the short-run, improve financial performance, but which, in the long-run, produce adverse effects. Cutting expenditures which are discretionary in nature is a common example of myopic behavior. Since maintenance expenditures have a large discretionary component, managers of airline firms could delay or decrease these expenditures in order to improve the short-run profitability of their firms. Although such delays or decreases ultimately will result in more breakdowns, lower output, and decreased safety of the operating equipment, the manager is not concerned since he or she anticipates promotion to a higher level thus avoiding the consequences of past decisions. Such myopic behavior can have extremely serious consequences in the airline industry where the delay or reduction of maintenance expenditures can endanger the traveling public as well as airline flight crews.

Very little literature bearing directly on myopic behavior and the erosion of safety exists. Myopic behavior is frequently discussed in the management accounting literature, but no study has presented any empirical evidence that supports or rejects the concept. A few studies have attempted to assess the overall effect of deregulation (e.g., Rose, 1981, Callari and Cooke, 1987, and Gesell, 1990) while one study (an event study by Michel and Shaked, 1984) examined the effect of deregulation on the stock market. Arguments have been made that safety has or has not been maintained in the deregulated period by appealing to accident and fatality statistics that have either remained the same or improved (Rosenfield,1986). Both Bruggnick (1991) and Gesell (1990), argue that safety actually has eroded due to a misinterpretation of accident and fatality statistics.

This study investigates the existence of myopic behavior on the part of managers of airlines as a result of deregulation and the erosion of safety that might result from such behavior. The remainder of the paper is organized into five sections. The next section presents the research hypotheses. The data and data collection are described in section three. Section four discusses the results of the analysis of the data. Section five discusses the implications of the findings. Section six provides a summary and conclusions.

HYPOTHESES

Regulated Environment Hypotheses

Financial Strength

The determinants of income were less volatile in the period prior to the deregulation of the airline industry in 1978. No price or route competition existed and there were significant barriers to entry into the industry. Prosperity of the airline firms was regulated to a large extent. Since price and route competition were not permitted, the airlines competed on other factors such as safety, comfort, and on-time departures. Indeed, such an environment created an incentive to allocate resources to safety. Increases in maintenance expenditures could be used to justify rate increases thus passing on to the consumer the higher average cost of doing business. Even financially distressed airlines had little incentive to reduce the level of discretionary maintenance expenditures. An inverse relationship may have existed between the level of maintenance expenditures and the level of financial distress as a result of the regulated environment.

Accordingly, the following hypothesis is posited:

H1: The level of financial distress of airline firms (as measured by the Z-scores of these firms) in the regulated period was inversely related to the level of discretionary maintenance expenditures of these firms.

The Z-score developed by Altman (1968) is a widely-used measure of financial distress. The Z-score captures the essence of both the bonus and debt variables originally identified by Watts and Zimmerman (1990). A significant problem noted by Watts and Zimmerman (1990) in the extant accounting choice studies is that of specification of the dependent and independent variables. The Z-score is a function of accounting variables that measure both liquidity and profitability. Use of the Z-score in a regression model collapses liquidity and profitability information into one variable thus avoiding problems of multicollinearity. The theoretical sign of the regression coefficient also will be unambiguous.

The accounting firm of Deloitte and Touche (1989) has used a Z-score specifically applicable to service industries such as airlines as part of their client acceptance procedures. The service industry Z-score is as follows:

(6.56 x ((Current Assets - Current Liabilities)/Total Assets))

* (3.26 x (Retained Earnings/Total Assets))

* (6.72 x (Operating Profit/Total Assets))

* (1.05 x Retained Earnings + Net Shareholders' Equity/(Current Liabilities + Noncurrent Liabilities))

More recent financial distress prediction models are available (e.g., Ohlson, 1980). Nevertheless, Hamer (1983) demonstrates that the various models available in the literature do not statistically differ in their ability to predict business failure.

Age of Aircraft

Bullock (1979) asserts that most types of equipment are characterized by wear-out failure, so the failure pattern is somewhat predictable. No failures occur in early time periods, but after some period of operation, failures begin to occur with increasing, and then decreasing frequency.

Such a pattern would seem to hold in a general sense for aircraft. Airplanes are subjected to particular stress during take-off and landing. The effects of these stresses over a number of years would begin to exact a toll upon the various components of the aircraft. Additionally, airlines during the regulated period were concerned more with safety and comfort since prosperity was less uncertain in the regulated environment. The interiors and other amenities of the aircraft operating in the regulated period likely would be changed more frequently in order to maintain the appearance of safety and comfort. Such changes would become necessary as a plane aged and interiors and other amenities became outdated.

Given the above, the following hypothesis is offered:

H2: The level of discretionary maintenance expenditures in the regulated period will vary directly with the age of aircraft.

The variable for age of aircraft was operationalized by using the percent of accumulated depreciation to total cost of flight equipment for each year for each company in the regulated and deregulated periods.

Activity

All productive facilities are susceptible to failure or deterioration due to the effects of use or age (Bullock,1979). Aircraft are no exception to this general statement. The increased use of an airplane would be expected to result in the need for increased maintenance.

A commonly used measure of activity in the airline industry is revenue miles flown. A revenue mile is defined as one mile flown in revenue producing service. Revenue miles flown measure the activity of the airline firm in terms of revenue-producing flights which would include both commercial and charter flights. Revenue miles is a broad measure of activity that should reflect the effect on the aircraft of both in-flight use and take-offs and landings and thus the need for maintenance. (1) Accordingly, the following hypothesis is offered:

H3: The level of discretionary maintenance expenditures in the regulated period will vary directly with the number of revenue miles flown.

Deregulated Environment Hypotheses

The general thrust of the Airline Deregulation Act of 1978 was to improve the lot of the passenger or user of the service (Farris, 1981). Competition and the market mechanisms were to provide wider passenger choice of carriers, routes, and service with progressively less control by the Civil Aeronautics Board (CAB). Specifically, all control over entry to markets was removed as was controlover rates. The Act provided for complete phase-out of CAB authority over domestic route entry, fares, mergers, acquisitions, and charters between 1978 and 1983 (Cavarra, Stover, and Allen, 1981).

The move from a highly-regulated environment similar to that of the public utility industry to a deregulated environment poses a number of problems. Deregulation could, for example, affect the risk perception of the airline industry by the capital markets. An increase in the risk perception could result in an increased cost of capital for airline firms. Airline companies would be required to earn greater returns in order to compete for capital. The need to increase returns could cause airlines to increase prices, reduce expenditures, reduce services, and reduce capital expenditures.

Cavarra, Stover, and Allen (1981) documented that an increase in risk actually occurred. They estimated the beta coefficients for all trunk and local service carriers that had stock prices listed on the CRSP tapes during the period 1975 through early 1979. (2) A comparison of the interval beta coefficients before the enactment, at the date of enactment, and after the enactment of deregulation showed that all but one airline (National) exhibited an increased beta close to the actual enactment of the law in 1978, with the average change being significant. National's beta remained nearly constant, a fact due in part to the attempts made by Pan Am, Eastern, and Texas International in 1978 to merge with National.

The airlines in the Cavarra, et al. (1981), study were divided into three categories: trunk versus local service carriers; large versus small; and "best," "middle," and "worst." The last classification was based on the financial strength of each airline during the period under investigation according to observations of the financial community.

The results of the tests based on the foregoing classifications showed that smaller airlines and airlines considered to be neither weak nor strong with respect to financial health incurred a less substantial increase in risk as a result of the move toward deregulation. Airlines classified as large and as best and worst all experienced substantial increases in risk.

The policy implication of these results for airline management is, according to Cavarra, et al. (1981), that airlines must find a means to increase their rates of return. Increases in rates of return could be realized by increasing fares, reducing services, reducing capital expenditures, or reducing other expenditures (such as maintenance expenditures).

Increasing Fares

Increasing fares in a period of rapid deregulation and increased competition was not a viable option for carriers in the years immediately following 1978. The Wall Street Journal of April 19, 1990, p. B1, reported that in the years after the Civil Aeronautics Board stopped setting fares, carriers slashed prices. New entrants flooded the market offering bargain-basement prices. But most of the discount airlines subsequently vanished becoming victims to the highly competitive environment. The airline industry became more concentrated as a result of numerous mergers and acquisitions. Nonetheless, inflation-adjusted fares in 1988 were still 20% lower than they were in 1978, the last year of regulation. Fares began to increase only after 1988 when the number of airline mergers peaked and the number of competitors and the level of competition decreased. During the intensely competitive period prior to 1988, increasing fares to improve rate of return was difficult if not impossible.

Reducing Services

Rose (1981) states that deregulation brought a decidedly different philosophy regarding the basis of airline competition. Service was the competitive criterion prior to deregulation. Scheduling, seating configurations, food and ground services, type of aircraft, safety, and geographical coverage were emphasized before deregulation.

Under deregulation, price competition outweighed service competition according to Rose (1981). Service impediments to mid-and small-sized cities began to occur after deregulation, although some of this disruption was later ameliorated by growth of commuter airline service to these cities. The numbers of flights to other destinations were drastically reduced and the now familiar inconvenience of overbooking became more common. The level and quality of cabin and in-flight services as well as ground services began to decline. Reduction of services in all forms became one means of dealing with profits eroded by the effects of deregulation in the somewhat hostile general economic climate of the 1980s.

Reducing Capital Expenditures

Dempsey (1991) states that under deregulation, the U.S. commercial fleet has decayed into the oldest in the developed world. Evidence of this fact is shown in Table 1 which presents the average age of airline fleets as of early 1989 (Wall Street Journal, 1989). The aging of airline fleets has been of some concern in recent years due to accidents involving older planes.

The Wall Street Journal of March 31, 1989, p. B1, reports the opinion of some airlines that an airplane can last virtually forever with proper maintenance. But critics disagree. The same report cites one expert who expresses a concern that old planes are more susceptible to corrosion.

Table 1 illustrates the large difference in fleet ages. On average, the planes of TWA were 5.6 years older than those of Delta. While claims may be made that fleet age is not a factor associated with airline safety, the aging of the U.S. fleet is evident. Airlines have met the financial strains of deregulation at least in part by avoiding major capital expenditures for new fleets. The airlines thus were left with older, less fuel-efficient aircraft that likely require more maintenance.

The aging of the domestic airline fleet in the deregulated period suggests an even stronger relationship between age and the level of maintenance costs in the deregulated period. The policy of airlines not to purchase new aircraft suggests that an increased level maintenance expenditures for these older planes would be necessary in the deregulated period. The following hypothesis is therefore offered:

H4: The level of discretionary maintenance expenditures in the deregulated period will vary directly with the age of aircraft.

Reducing Maintenance Expenditures

After deregulation, competitive pressures have squeezed profits for many, if not most, airlines and have driven some of the less efficient airlines out of business. For managers faced with declining profits and possible bankruptcy, the temptation to cut maintenance costs in order to survive the short-run may have become very strong. Indeed, Duffy (1986) argues that many managers are succumbing to this temptation.

Evidence that Duffy's contention may be correct is found in the two newspaper reports mentioned earlier in this paper regarding USAir and Eastern Airlines. The increased emphasis on competition brought about by deregulation also is evidenced in a report-in The Wall Street Journal of September 29, 1989, p. B1. During 1988, major U.S. airlines spent about the same on commissions to travel agents as they did on maintenance. Twelve percent of the airlines' total expenses of $42.49 billion went for agents' commissions, while 13% went toward aircraft maintenance. Furthermore, studies by O'Brian (1987) and Nance (1986) suggest that competitive pressures brought on by deregulation have resulted in an erosion of safety maintenance. Airlines no longer exceed the minimum FAA safety standards to the same level as that prior to deregulation. Gerston, Fraleigh, and Schwab (1988) report that the number of mechanics employed by the major airlines decreased by 2,000 from 1974 to 1984, while the number of planes in operation increased dramatically. Golich (1988) reports that the number of maintenance workers employed may have decreased by as many as 4,000 between 1979 and 1984, while the number of federal safety inspectors fell by 700 during the same period.

A question arises at this point as to the effect of federal regulations regarding performance of specific maintenance procedures. An argument might be made that maintenance procedures could not be omitted due to federal requirements that certain procedures be performed and careful audit by federalinspectors of airline records of the performance of such procedures. This argument was presented to an airline executive responsible for in-flight safety for his company. His response was that there is nothing to prohibit an airline from "penciling in" a procedure on the maintenance records even though the procedure was not performed. Furthermore, the limited number of federal inspectors available to oversee the maintenance functions of the major airlines could make detection of such entries to the records unlikely. The fact that at least one major airline has been fined for falsifying maintenance records would lend support to these assertions (Salt Lake Tribune, 1993).

The extremely competitive environment brought about by deregulation suggests that myopic behavior on the part of airline managers will be observed. The artificial barriers to entry into the industry have been removed and the nature of the airline industry seems to foster the entry of new airlines and the exit of less efficient airlines. Since this implies the continual presence of struggling airlines and continued profit pressure for many others, it can be argued that the overall safety of air travel has declined in the deregulated environment. Moreover, even if myopic behavior is observed in both the regulated and deregulated environments, there would seem to be a greater likelihood of such behavior occurring in the more pressured environment of deregulation resulting in an erosion of safety occurring in the deregulated period. Accordingly, the following hypotheses are posited:

H5 The level of financial distress of airline firms in the deregulated period affected the level of discretionary maintenance expenditures of these firms.

H6 The presence of myopic behavior was greater in the deregulated environment than in the regulated environment, thus resulting in an erosion of safety after deregulation.

Activity

The relationship of activity to maintenance costs in the deregulated period would not be expected to differ from that of the regulated period. Increased use of aircraft would be expected to result in an increased level of maintenance expenditures. The following hypothesis is therefore offered:

H7 The level of discretionary maintenance expenditures in the deregulated period will vary directly with the number of revenue miles flown.

DATA DESCRIPTION

The sample initially consisted of the largest 11 U.S. trunk carriers and 9 local U.S. carriers in operation before and after deregulation. The sample included all the larger domestic airlines in the United States. The time period covered was 1968 through 1987. The year 1987 was chosen (even though data was available for later years) due to the fact that the large number of mergers occurring during the mid-to late-1980s resulted in a very concentrated industry after 1987 in which the competitive pressures of deregulation were becoming less evident. The regulated period was defined as 1968 to 1978 (the Airline Deregulation Act was signed by President Carter in October, 1978). The deregulated period was specified as 1979 through 1987.

Accounting data, fleet size, and other operating statistics were obtained from the Handbook of Airline Statistics of the Federal Aviation Administration and from Moody's Transportation Manual. Since a number of mergers occurred during the period examined, each airline company was defined as the merged entity and pre-merger data for the merging airlines was combined.

Table 2 shows the original 20 airlines, those that were acquired as well as the acquiring company, and the final list of 11 combined airlines used in the analysis.

EMPIRICAL RESULTS

Figure 1 presents the plot of the price-adjusted maintenance cost for the airline industry for the years 1968-1987. Data was obtained from the Handbook of Airline Statistics of the Federal Aviation Administration. Maintenance expenditures for each year were adjusted to 1987 prices using the Consumer Price Index for all Urban Consumers (CPI-U) as suggested by Statement of Financial Accounting Standards No. 89, "Financial Reporting and Changing Prices" (FASB, 1979).

Figure 1 shows a fairly consistent level of expenditures for the period from 1968-1978, the regulated period. Expenditures fell in 1980-1983 (after deregulation) but began to rise slightly beginning in 1984.

[FIGURE 1 OMITTED]

Figure 2 presents the revenue passenger miles for the airline industry for 1968-1987. Contrary to the graph of price-adjusted maintenance costs, the pattern of which is relatively flat, the graph of industry revenue passenger miles shows a definite upward slope. Figures 1 and 2 taken together simply say that more passengers have been flown more miles (by an industry with an aging fleet) while the price-adjusted maintenance cost has remained relatively constant.

[FIGURE 2 OMITTED]

Regression Model

Tests of the hypotheses proposed in this paper were made using the following general model:

* MAINTCOST = [b.sub.0] + [b.sub.1]RVMLS + [b.sub.2]ADCOST + [b.sub.3]Z + e, where:

* MAINTCOST = The price-level adjusted costs for each airline company for each year in

* the regulated period. Maintenance costs were adjusted using the Consumer Price Index for All Urban Consumers (CPI-U).

* RVMLS = Revenue miles flown for each airline company for each year in the regulated period.

* ADCOST = A measure of the effects of aircraft age based on accumulated

* depreciation to total cost of flight equipment for each year for each company in the sample

* Z = The Z-score for each airline company for each year.

We began by estimating the model for the data in the entire database in order to assess the overall effectiveness of the model in explaining the variance in the dependent variable. We predicted that the signs of the RVMLS and ADCOST variables would be positive, but were unsure regarding the sign of the Z-score for the entire data set due to the hypothesized effects of regulation vs. deregulation and financial distress of the companies in the sample. The results of this first regression are shown in Table 3.

This first regression demonstrated the effectiveness of the model in explaining the variance in the inflation-adjusted maintenance cost. Both the revenue miles (RVMLS) and age (ADCOST) variables are highly significant and produce a very high [r.sup.2] (.8347). Estimation of the model omitting the Z-score variable resulted in an [r.sup.2] of .8222, thus demonstrating the appropriateness of the RVMLS and ADCOST variables in explaining the variance in maintenance cost.

Classification of Airlines

The ambiguity of the sign and lack of significance of the Z-score in the model estimated for the entire data set supported our initial belief that the most effective manner in which to test the hypotheses was to classify the airline companies into two groups (strong and weak) according to the degree of financial distress of each firm as indicated by the Z-score for the regulated and deregulated periods. Accordingly, a Z-score was calculated for each airline for each year for the regulated and deregulated periods. The average Z-score was then calculated for each firm for the regulated period and for the deregulated period. The average Z-scores were then ordered and the companies were assigned to the strong or weak groups for the regulated and deregulated periods. Airlines thus were classified as strong or weak in the regulated and deregulated periods respectively. Classification was based on the criteria that a Z-score greater than or equal to 2.6 means that bankruptcy is unlikely while a Z-score less than or equal to 1.1 means that bankruptcy is probable. These classifications are reported in Table 4.

Many of the average Z-scores calculated for the airlines fell into what Altman refers to as the "zone of ignorance," that is, the average Z-score in this range makes it uncertain about how a firm should be classified for purposes of prediction of bankruptcy. We do not view this situation as posing a major problem for this study for two reasons. First, we are not attempting to predict the future bankruptcy of the firms. Our use of the Z-score is solely to classify the firms according to their financial strength. Secondly, the average Z-scores of the strong and weak firms in both the regulated and deregulated periods seem to divide the firms quite naturally according to their financial strength. For example, in the regulated period, the average Z-scores of the weak companies are less than 1.00 for every firm except TWA. The average Z-score for TWA is 1.01. The next highest Z-score is 1.43 for Pan Am and the average Z-scores for the other strong firms are all greater than 1.43. An even stronger natural division of the firms occurs in the deregulated period in which the average Z-score of .21 for TWA is the highest for the weak firms, but is well below the 1.16 of American, which is the lowest average Z-score of the strong firms. Furthermore, information from the general financial press regarding the financial condition of these airline companies for the periods in question seems to suggest that the classifications presented are reasonable.

Regulated Period Results

Hypotheses 1 through 3 for the regulated period were tested by estimating the coefficients in the regression model described above. Table 5 presents the correlation matrices for the companies classified as weak and strong during the regulated period. Table 6 presents the results of the regression models for the regulated period.

The results for the strong companies were exactly as hypothesized. The level of maintenance expenditures is positively and significantly related to the level of activity (revenue miles flown) and to the age of the aircraft fleet. The level of financial distress is negatively and significantly related to the level of maintenance expenditures.

The results for the weak companies were exactly as hypothesized, except for the Z-score variable. The Z-score variable was not significant for the weak companies. This may suggest that the effect of regulation was even stronger than we anticipated as the financial condition of the weak companies seemed to have had no effect on the level of maintenance expenditures. Managers of financially distressed airline firms in the regulated period thus had little or no incentive to reduce the level of discretionary maintenance expenditures due to the protection afforded by the regulated environment.

Discretionary Expenditures

In order to provide a more effective isolation of the effect of discretionary maintenance expenditures, we used the data for the strong and weak companies in the regulated period to estimate a regression model using only the activity variable (RVMLS) and the age variable (ADCOST). We then used the residuals from these two equations as a proxy for the discretionary portion of maintenance expenditures. The residuals then were regressed on the financial distress variable (Z) to provide a better measure of the effect of financial distress on the level of discretionary maintenance expenditures. The results of these regressions are shown in Table 7.

As before, financial distress seemed to be negatively related to the level of maintenance expenditures for the strong companies, while the level of financial distress had no effect on maintenance expenditures of the weak companies. These results further support the hypotheses that the financial condition of an airline company had no effect on the level of maintenance expenditures (discretionary or otherwise) of the company during the regulated period.

Deregulated Period Results

Hypotheses 4 through 7 were tested by estimating the coefficients in the same regression model as was used in the regulated period. Table 8 presents the correlation matrices for companies classified as weak and strong during the deregulated period. Table 9 presents the results using multiple regression to estimate the multivariate model.

The results for the weak companies are exactly as hypothesized. Activity and age of aircraft are positively and significantly related to the level of maintenance expenditures. More importantly, the level of financial distress is positively and significantly related to the level of maintenance expenditures. This result suggests that as the Z-score rises (indicating less financial distress) the level of maintenance expenditures also rises. Conversely, when the Z-score falls (indicating a greater level of financial distress), the level of maintenance expenditures falls. Such a relationship suggests that managers of the weaker airline companies may have behaved myopically by reducing the level of maintenance expenditures in order to improve short-run financial performance. Furthermore, given that maintenance expenditures are a measure of safety, the fact that myopic behavior did not exist in the regulated environment but does exist in the deregulated environment suggests that for the weak companies maintenance expenditures have declined and safety has eroded.

Results for the strong companies in the deregulated period also are shown in Table 9. Activity was positively and significantly related to the level of maintenance expenditures as hypothesized. The level of financial distress, however, is negatively and significantly related to the level of maintenance expenditures and the age variable no longer is significant. This result suggests that the financially strong companies continued to behave in the deregulated period much as they did during the regulated period as regards maintenance expenditures. These companies were financially strong and had relatively new fleets such that age of aircraft was not a factor in explaining maintenance expenditures. Such a result is not altogether surprising. American and Delta emerged as the strongest companies in the industry during the time period covered by this study. These companies were able to continue many of the practices of the regulated period even after deregulation as a result of their financial strength and the demise of many of their significant competitors. Four of the five companies classified as weak in the deregulated period filed for protection under the bankruptcy laws and two of these four companies (Pan American and Eastern) no longer exist. (3)

Discretionary Expenditures

We again attempted to isolate the effect of discretionary expenditures by using the data for the strong and weak companies in the deregulated period to estimate a regression model using only the activity variable (RVMLS) and the age variable (ADCOST). The residuals from these two equations then were used as a proxy for the discretionary portion of maintenance expenditures. The residuals for the strong and weak companies were regressed on the financial distress variable (Z) to provide a better measure of the effect of financial distress on the level of discretionary maintenance expenditures. The results of these regressions are shown in Table 10.

As before, financial distress was positively related to the level of maintenance expenditures for the weak companies and negatively related to maintenance expenditures for the strong companies. These results further support the hypotheses regarding the existence of myopic behavior in the deregulated period among managers of the financially weak airline companies.

Structural Change

Hypothesis 7 regarding the fact that myopic behavior is posited to be greater in the deregulated environment than in the regulated environment is further investigated in this section.

The regression models generated a statistically insignificant coefficient for the Z-score in the regulated period and a statistically significant, positive Z-score in the deregulated period. We present results in this section of a formal test to determine that a statistically significant difference exists among the coefficients in order to determine that structural change in the model occurred as a result of deregulation. We will test in this section only the full model consisting of revenue miles, age, and financial distress rather than the model attempting to isolate the discretionary portion of maintenance expenditures due to the similarity of the results for the two methods.

Greene (1993) describes a test for structural change in coefficients attributed to Chow (1960). For the data in this study, we estimated a regression for all companies classified as strong in both the regulated and deregulated periods and a regression for all companies classified as weak in both the regulated and deregulated periods. The results of these regressions are presented in Tables 11 and 12. The test is based on comparing the residual sums of squares of the equations for all of the strong companies with the residual sum of squares of the strong companies in the regulated period and the strong companies in the deregulated period as shown in Table 11. The F statistic for testing the hypothesis that the coefficients in the two equations for the regulated and deregulated periods are the same is presented in Table 11. A similar approach for the companies classified as weak is shown in Table 12.

F-statistic for testing if coefficients in the two equations are the same:

F[4 93] = [(1,124,515,700,000 - 383,296,078,987 - 73,722,856,572) / 4]/[(383,296,078,987 + 73,722,856,572) / (50 + 51 - 8)]

= 33.9576734479

Tabled F [4,93] for 5 percent significance = 2.47

Hypothesis that coefficient vectors are the same in the regulated and deregulated periods is rejected.

F-statistic for testing if coefficients in the two equations are the same:

F[4 94] = [(424,394,589,775 - 56,918,756,398 - 242,915,587,405) / 4]/[(56,918,756,398 + 242,915,587,405) / (45 + 57 - 8)]

= 9.76261005733

Tabled F [4,94] for 5 percent significance = 2.47

Hypothesis that coefficient vectors are the same in the regulated and deregulated periods is rejected.

The tabled critical values for the F statistic in both Tables 11 and 12 are less than the calculated F statistics for both the strong and weak companies. The hypothesis that the coefficient vectors are the same in the two periods is rejected for both the strong and weak companies. As a result, a structural change in the models has occurred as a result of deregulation.

IMPLICATIONS OF THE STUDY

This study provides empirical data on the effects of regulation and deregulation on a specific industry, viz., the airline industry. The evidence presented in this study suggests that the regulated environment provided no incentive for managers to behave myopically as regards maintenance expenditures. This is true both of financially strong and weak airline companies. These findings do not suggest, however, that regulation is the appropriate policy for the airline industry or for other modes of transportation generally.

Stigler (1971) argues that regulation serves the private ends of those being regulated rather than the public welfare. Gesell (1990) suggests that regulation was the focus of consumer neglect resulting in social costs and misallocation of resources. Gesell asserts that the Civil Aeronautics Board (CAB) may have been created at the behest of the airlines as a result of an overly competitive market. He suggests that the CAB had enormous powers that were used to create an "imperfect cartel" designed to satisfy air carriers at the expense of the travelling public. The result was a collusion of two great forces of the American society: government and industry. Such an environment would explain, at least in part, the findings of this study suggesting that airline companies would continue to spend on maintenance even when financial condition takes a turn for the worse.

The problems associated with regulation generally were underscored by the condition of the railroad systems in the Northeastern United States in the 1970s. The railroad systems of the Northeast were going bankrupt despite the regulatory protection provided them. This fact, coupled with the energy crisis of the 1970s, the perceived fuel efficiency of air transport, and the consumer movement of the 1960s and 1970s led to the push toward deregulation of the airline industry according to Sampson, Farris, and Shrock (1985).

The results of this study suggest that the virtual total deregulation of the airline industry also may pose some significant problems for the consumer of airline services and public welfare generally. This study provides empirical evidence of myopic behavior on the part of the managers of airline companies. The analysis reported in this study suggests that managers of financially weak airlines may have behaved myopically in the deregulated period by reducing maintenance expenditures as the financial condition of their companies worsened. In this regard, Gesell (1990) asserts that government regulation is necessary as the result of the re-emergence of the profit-orientation that occurred when government control was removed from the industry as a result of deregulation. Gesell believes that the consumer has suffered both in terms of service and safety as a result of the deregulation of the airline industry. He suggests that the failure of the deregulation of the airline industry is a valid reason for government intervention in the economic marketplace and proposes reregulation of the industry (though not to the extent that existed prior the enactment of the Airline Deregulation Act of 1978).

The results of the study reported in this paper suggest that further investigation is needed in several areas. The first is the area of corporate ethics and what Gesell (1990) calls the profit-orientation of corporate executives. Is there, as Gesell claims, a latent tendency of private enterprise to deviate from serving the welfare interests of society as a result of the profit motive and corporate greed? Are there elements of corporate ethics that need to be reconsidered? Research on executive behavior should be (at least in part) of an empirical nature similar to this study in which the behavior of corporate executives is related to the outcomes of their decisions. It is not sufficient simply to ask an executive what he or she might do in a contrived laboratory situation. The role of government in the regulation of private enterprise also should be considered. Is government oversight necessary to counteract the profit motive and corporate greed?

Is government oversight necessary to ensure the proper allocation of resources and the minimization of social costs?

SUMMARY AND CONCLUSION

This study has addressed the effect of deregulation on the level of maintenance expenditures and safety of the airline industry. The study was limited to a specific time period and therefore cannot capture every aspect of a very dynamic economic environment. Because the study covers a specific time period, it does not consider current efforts of airline companies to improve their maintenance functions, such as quality improvement programs Nevertheless, we believe that this study provides information and raises questions that should be considered in future research regarding both the behavior of corporate executives and the role of government in regulating economic activity.

REFERENCES

Altman, E. I. (1968). Financial ratios, discriminant analysis, and the prediction of corporate bankruptcy. Journal of Finance (September), 589-609.

Bruggnick, G. (1991). U.S. aviation accidents and deregulation Airline Pilot (March), 20-24.

Bullock, J. (1979). Maintenance Planning and Control.New York: National Association of Accountants.

Callari, J. and J. Cooke (1987).Deregulation impact: The air carriers. Traffic Management (September), 52-55.

Cavarra, L., R. Stover, and B. Allen (1981).The capital market effects of airline deregulation. Transportation Journal, (Spring), 73-78.

Chow, G. (1960). Tests of equality between sets of coefficients in two linear regressions. Econometrica 28, pp. 591- 605.

Deloitte and Touche (1989). Risk management process for audit client, Z-1. New York: Deloitte & Touche.

Dempsey, P. (1991). The disaster of airline deregulation. Wall Street Journal, (9 May), A15.

Duffy, H. (1986). Deregulation is eroding airline safety. National Safety and Health News, (July), 62-63.

Farris, M. T. (1981). The multiple meaning and goals of deregulation: A commentary. Transportation Journal, (Winter), 44-50.

Financial Accounting Standards Board (1979). Statement of Financial Accounting Standards No. 89: Financial Reporting and Changing Prices. Stamford, CT, FASB.

Genston, L. N., C. Fraleigh, and R. Schwab (1988). The Deregulated Society. Pacific Grove, California: Brooks/Cole Publishing Company.

Gesell, L. E. (1990). Airline Re-Regulation. Chandler, Arizona: Coast Aire Publications.

Golich, V. L. (1988). Airline deregulation: Economic boom or safety bust? Transportation Quarterly (April).

Greene, W. H. (1993). Econometric Analysis, Second Edition. New York: Macmillan Publishing Company.

Michel, A. and I. Shaked (1984). Airline performance under deregulation: the shareholders' perspective. Financial Management, (Summer), 5-14.

Nance, J. J. (1986). Blind Trust. New York: William Morrow and Company, Inc.

O'Brian, J. (1987). Quoted in Pilots union says safety hurt by deregulation. Aviation Week and Space Technology (July).

Ohlson, J. (1980). Financial ratios and the probablistic prediction of bankruptcy Journal of Accounting Research 18, (Spring), 109-131.

Rose, W. (1981). Three years after airline passenger deregulation in the United States: a report card on trunkline carriers. Transportation Journal, (Winter), 51-58.

Rosenfield, H. (1986). Special report: is regulation good or bad? National Safety and Health News, (July), 60.

Salt Lake Tribune (1991). Delta cuts a deal with Air Force on Pan Am debt. (22 October), C8.

Salt Lake Tribune (1993). Repairs falsified, admits USAir; probe continues. (25 March), A12.

Sampson, R. J., M. Farris, and D. L. Shrock (1985). Domestic Transportation Practice, Theory, and Policy (6th Ed.). Boston: Houghton Mifflin Company.

Stigler, G. (1971). The theory of economic regulation. Bell Journal of Economics, (Spring).

Wall Street Journal(1989). Which airlines have the oldest fleets. (31 March), 131.

Wall Street Journal(1989). Odds and ends. (29 September), B1.

Wall Street Journal(1990). One sure result of airline deregulation: controversy about its impact on fares. (19 April), B1.

Watts, R. and J. Zimmerman 1990. Positive accounting theory: a ten-year perspective. The Accounting Review, (January), 131-156.

Larry A. Deppe, Weber State University

Don R. Hansen, Oklahoma State University

James G. Swearingen, Weber State University

ENDNOTES

(1.) The size of the fleets of the airline companies also was considered as an independent variable in the model. Fleet size was found to be highly correlated with revenue miles as the larger the airline fleet, the greater the number of revenue miles flown. Revenue-miles-flown was found to be stronger in explaining the variance in the dependent variable resulting in the exclusion of fleet size from the model.

(2.) Pan Am was excluded from the study of Cavarra, et al., since it was primarily an international carrier during the time period investigated and, therefore, would have been largely unaffected by most aspects of deregulation. Pan Am increased its domestic operations substantially through acquisition of National Airlines in January 1980.

(3.) The Civilian Reserve Aircraft Fleet program of the United States Air Force provides the government access to civilian aircraft for emergency military use. As a part of this program, the Air Force spent a total of $561 million dollars preparing aircraft owned by Pan American World Airways for such emergency use (The Salt Lake Tribune, October 22, 1991, p. C8). Most of this money was provided to Pan American during the deregulated period thus raising the issue of whether the results of this study were affected by the Air Force expenditures on Pan American's planes. The answer to this question is that the results do not seem to be affected by these expenditures of the Air Force in the deregulated period.

Table 1: Average Age of Airline Fleet
1989

AIRLINE AVERAGE AGE
 IN YEARS

TWA 14.3
Northwest 14.1
Eastern 13.8
United 13.6
Pan Am 12.8
Continental 11.0
American 9.4
US Air 9.0
Piedmont 9.0
Delta 8.7

Compiled by Airline Economics, Inc. and
reported in The Wall Street Journal, March
31, 1989.

Table 2: Sample of Airline Companies

ORIGINAL SAMPLE ACQUIRED COMPANIES FINAL SAMPLE
 (Date acquired in
 parenthesis)

Aloha Aloha

American American

Continental Texas International Continental--
 (10/82) combined (')

Delta Northeast (8/72) Delta--combined
 Western (12/86)

Eastern Frontier (12/86) (2) Eastern--combined

Frontier -

Hughes Air West -

National -

North Central -

Northwest Republic' (8/86) Northwest--combined
 North Central
 Southern Hughes
 Air West

Ozark

Pan Am National (1/80) Pan Am--combined

Piedmont Piedmont (')

Republic -

Southern -

Texas International -

TWA Ozark (3/86) TWA--combined

United United

US Air (operated as US Air
Allegheny Airlines
until 1979 name
change)

Western -

Notes:

(') Term "combined" signifies that for companies where a merger
occurred, the airline is defined as the merged entity and premerger
data for the airlines involved in the merger is combined.

(2) Frontier was acquired 12/85 by People Express. Eastern acquired
People Express 12/86. Eastern-combined includes both Frontier and
People Express.

(') Republic was created in 1977 by merger of North Central and
Southern. Republic acquired Hughes Air West 10/80. Northwest acquired
Republic 8/86.

(') US Air and Piedmont merged Aug. 5, 1989. This merger is not
reflected in the data since the study covers only the period 1968-1987.

Table 3: Regression Results

Entire Sample 1968-1987

Variable Predicted Coefficients
 Sign [p-values in
 parentheses]

RVMLS + 27.278
 [.0001]

ADCOST + 4.046
 [.0001]

Z 0.755
 [.4512]

F-statistic 336.555
 [.0001]

R-squared .8347

Adjusted .8322
 R-squared

Table 4
Classification of Airline Companies as Weak or Strong

Criteria: Z [greater than or equal to] 2.6 indicates
bankruptcy unlikely Z [less than or equal to] 1.1
indicates bankruptcy probable

Regulated Period 1968-1978:

Weak Companies

Company Name Average Z
 score--Regulated
 Period

 Aloha -1.27
 Continental 0.86
 Eastern 0.63
 Piedmont 0.61
 TWA 1.01
 US Air 0.46

Strong Companies

Company Name Average Z
 score--Regulated
 Period

 American 1.79
 Delta 2.70
 Northwest (NWA) 3.48
 Pan Am 1.43
 United 1.71

Deregulated Period 1979-1987:

Weak Companies

Company Name Average Z
 score-Deregulated
 Period

 Continental -0.79
 Eastern 0.04
 Pan Am -1.32
 TWA 0.21
 United -0.21

Strong Companies

Company Name Average Z
 score-Deregulated
 Period

 Aloha 1.22
 American 1.16
 Delta 2.15

Northwest (NWA) 1.84

Piedmont 1.75

Table 5: Correlation Matrices Independent Variables
Regulated Period 1968-1978

 Weak Companies

 RVMLS ADCOST Z

RVMLS .31267
 .24657
ADCOST 1.00000 -.10949
 1.00000
Z 1.00000

 Strong Companies

 RVMLS ADCOST Z

RVMLS -0.25166
 .32054
ADCOST 1.00000 -0.14630
 1.00000
Z 1.00000

Table 6: Regression Results
Regulated Period 1968-1978

Companies Predicted Coefficients
Classified Sign [p-values in
as Weak parentheses]

RVMLS 47.615
 [.0001]

ADCOST + 1.974
 [.0536]

Z + -0.159
 [.8746]

F-statistic - 939.618
 [.0001]

R-squared .9815

Adjusted .9805
 R-squared

Companies Predicted Coefficients
Classified Sign [p-values in
as Strong parentheses]

RVMLS 4.624
 [.0001]

ADCOST + 2.019
 [.0493]

Z + -4.711
 [.0001]

F-statistic - 25.256
 [.0001]

R-squared .6222

Adjusted .5976
 R-squared

Table 7: Regression Results Isolating the Discretionary Portion
of Maintenance Expenditures Regulated Period 1968-1978

Companies Classified Predicted Coefficients
as Weak

Variable Sign [p-values in
 parentheses]

Nondiscretionary Portion:

RVMLS + 53.577
 [.0001]

ADCOST + 1.773
 [.0813]

F-statistic 1554.740
 [.0001]

R-squared .9811

Adjusted R-squared .9804

Discretionary Portion:

Z - -0.160
 [.8732]

F-statistic .26
 [.8732]

R-squared 0.0005

Adjusted R-squared -0.0177

Companies Classified as Predicted Coefficients
Strong

Variable Sign [p-values in
 parentheses]

Nondiscretionary Portion:

RVMLS + 4.774
 [.0001]

ADCOST + 1.771
 [.0825]

F-statistic 17.467
 [.0001]

R-squared .4018

Adjusted R-squared .3788

Discretionary Portion:

Z - -4.746
 [.0001]

F-statistic 22.526
 [.0001]

R-squared .3194

Adjusted R-squared .3052

Table 8: Correlation Matrices Independent Variables
Deregulated Period 1979-1987

 Weak Companies

 RVMLS ADCOST Z

RVMLS 1.00000 0.10761 0.27224
ADCOST Z 1.00000 0.08339
 1.00000

 Strong Companies

 RVMLS ADCOST Z

RVMLS 1.00000 0.54994 0.15259
ADCOST Z 1.00000 -0.11969
 1.00000

Table 9: Regression Results
Deregulated Period 1979-1987

Companies Predicted Coefficients
Classified Sign [p-values in
as Weak parentheses]

RVMLS 13.139
 [.0001]

ADCOST + 3.998
 [.0003]

Z + 2.008
 [.0512]

F-statistic + 79.140
 [.0001]

R-squared .8527

Adjusted .8420
 R-squared

Companies Predicted Coefficients
Classified Sign [p-values in
as Strong parentheses]

RVMLS 29.639
 [.0001]

ADCOST + -0.265
 [.7924]

Z + -1.847
 [.0711]

F-statistic 438.597
 [.0001]

R-squared .9655

Adjusted .9633
 R-squared

Companies Predicted Coefficients
Classified Sign [p-values in
as Weak parentheses]

RVMLS 13.139
 [.0001]

ADCOST + 3.998
 [.0003]

Z + 2.008
 [.0512]

F-statistic + 79.140
 [.0001]

R-squared .8527

Adjusted .8420
 R-squared

Companies Predicted Coefficients
Classified Sign [p-values in
as Strong parentheses]

RVMLS 29.639
 [.0001]

ADCOST + -0.265
 [.7924]

Z + -1.847
 [.0711]

F-statistic 438.597
 [.0001]

R-squared .9655

Adjusted .9633
 R-squared

Table 10: Regression Results Isolating the Discretionary
Portion of Maintenance Expenditures Deregulated Period
1979-1987

Companies Predicted Coefficients
Classified as Sign [p-values in
Weak Variable parentheses]

Nondiscretionary
Portion:

RVMLS + 13.697
 [.0001]

ADCOST + 3.977
 [.0003]

F-statistic 108.834
 [.0001]

R-squared .8383

Adjusted R-squared .8306

Discretionary
Portion:

Z + 1.968
 [.0555]

F-statistic 3.874
 [.0555]
R-squared .0827

Adjusted R-squared .0613

Companies Predicted Coefficients
Classified as Sign [p-values in
Strong Variable parentheses]

Nondiscretionary
Portion:

RVMLS + 29.425
 [.0001]

ADCOST + 0.181
 [.8572]

F-statistic 624.824
 [.0001]

R-squared .9630

Adjusted R-squared .9615

Discretionary
Portion:

Z - -1.806
 [.0771]

F-statistic 3.261
 [.0771]
R-squared .0624

Adjusted R-squared .0433

Table 11: Test of Structural Change in Regression Equations
Companies Classified as Strong

Coefficients 1968-1987 1968-1978 1979-1987

Constant -0.220 3.524 1.001
 [.8263] [.0010] [.3221]

RVMLS 12.775 4.624 29.639
 [.0001] [.0001] [.0001]

ADCOST 2.541 2.019 -0.265
 [.0126] [.0493] [.7924]

Z -1.586 -4.711 -1.847
 [.1160] [.0001] [.0711]

[R.sup.2] .7348 .6222 .9655

F-statistic 89.569 25.256 438.597
 [.0001] [.0001] [.0001]

Sum of 1,124,515, 383,296, 73,722,
 squared 700,000 78,987 856,572
 residuals

Number of 101 50 51
 observations

Table 12: Test of Structural Change in Regression Equations Companies
Classified as Weak

Coefficients 1968-1987 1968-1978 1979-1987

Constant -3.275 -2.227 -1.978
 [.0015] [.0302] [.0547]

RVMLS 31.317 47.615 13.139
 [.0001] [.0001] [.0001]

ADCOST 4.336 1.974 3.998
 [.0001] [.0536] [.0003]

Z 2.584 -0.159 2.008
 [.0112] [.8746] [.0512]

[R.sup.2] .9255 .9815 .8527

F-statistic 405.587 939.618 79.140
 [.0001] [.0001] [.0001]

Sum of 424,394, 56,918, 242,915,
 squared 589,775 756,398 587,405
 residuals

Number of 102 57 45
 observations