Вы находитесь на странице: 1из 21

Transportation Research Part E 40 (2004) 443463

www.elsevier.com/locate/tre

Management emphasis and performance in the airline


industry: an exploratory multilevel analysis
Sveinn Vidar Gudmundsson *

CERMAS Research Centre and Department Strategy, Toulouse Business School,


20 Boulevard Lascrosses, BP 7010, 31068 Toulouse Cedex 7, France

Abstract

Carrying through decisions in organisations is inherently micro-level, while the aggregated performance
implication of decisions is a macro-level phenomena. This paper determines through exploratory factor
analysis, factors associated with airline performance using a two-level bottomup hierarchical approach.
The determinant factors are used to test, through regression analysis, the relationship across lower and
higher level factors and between the higher level factors and performance (distress versus non-distress).
The results showed that airlines having higher relative score on productivity and brand image were less likely
to be under nancial distress, while airlines having higher relative emphasis on market power were more
likely to be under nancial distress.
2004 Published by Elsevier Ltd.

Keywords: Multilevel analysis; Business performance; Business distress; Airlines; Business success; Airline strategy;
Airline performance

*
Tel.: +33 5 61 29 48 43; fax: +33 5 61 29 49 94.
E-mail address: s.gudmundsson@esc-toulouse.fr

1366-5545/$ - see front matter 2004 Published by Elsevier Ltd.


doi:10.1016/j.tre.2004.08.004
444 S.V. Gudmundsson / Transportation Research Part E 40 (2004) 443463

1. Background

1.1. Introduction

Performance indicators are a source of debate as to which one provide the most eective feed-
back on business performance. What is more, the link between variation in the indicators and
company performance has been dicult to establish. Quantitative failure prediction models
(see Altman, 1993; Gudmundsson, 1998, 1999, 2002) based on nancial ratios are a case in point.
These serve their purpose in segregating companies into the failed and non-failed groups, but do
not produce underlying explanatory relationships behind performance variation. Eccles (1991)
and Argenti (1976) argued that alternative methods of performance measurement was needed,
stressing the point that traditional nancial ratios are too simplistic and backward-looking to
serve their purpose.
Many theories have attempted to explain variation in business performance. Leadership theories
are associated with the belief that variance in the rms performance can be associated with the
leadership qualities of top managers. Preisendorfer and Voss (1990) argued, using human capital
theory, that in small organisations a considerable proportion of variance in organisational activ-
ities and outcomes could be associated with individuals. Another study by Westerberg et al.
(1997), using multidimensional performance constructs of market performance and nancial per-
formance, showed that CEO characteristics had considerable impact on performance. However, a
stream of early research (Salancik and Pfeer, 1977) has claimed that organisations are driven by
other more important factors than top managers.
Another school of thought claims that market orientation holds the key to the rms success.
Narver and Slater (1990), for example, state that market orientation is associated with the crea-
tion of superior value for buyers and therefore superior performance of the rm. Appiah-Adu and
Ranchhod (1998) found positive and signicant association between market orientation and three
out of four performance measures examined. Market dominance is, however, segregated in terms
by assuming superior market-share, while superior market-share can be achieved with or without
market orientation. For this reason a non-market orientated rm under distress can have superior
market-share, achieved through deep price cutting or mergers and acquisitions (Buzzel and Gale,
1987). A similar line of thought is the economies of scale factor (Smith, 1955) that is considered by
some as being the essential element in the rms success: a rm should focus on achieving critical
mass and benet from economies of scale as a consequence. Levine (1987) showed that in the air-
line industry there are further economies that have been demonstrated to play an even greater role
than economies of scale: economies of scope, which denotes the reach of a route network; and
economies of density that depicts the combination of small distributed loads from a number of
points of origin into a larger more economical load through a centre to a nal destination. Yet
the advent of so called low-cost, direct service carriers, has demonstrated that superior perform-
ance can be achieved with relatively low density economies (point to point services versus hub and
spoke) and high performance can be maintained with relatively limited scope economies and slow
growth. 1

1
Southwest Airlines is an example of this.
S.V. Gudmundsson / Transportation Research Part E 40 (2004) 443463 445

The organisational structure factor (Leavitt and Bass, 1964; Lawrence and Lorsch, 1967) holds
that the way the rm is organised is essential for success. Factors such as decentralisation and em-
ployee autonomy play role in the organisational structure factor. The rm should be organised in
such a way that all components work in harmony to reach the rms mission. However, as Eilon
(1992) points out there have been a debate as to the impact of centralisation versus decentralisa-
tion often leading to contradictory results when associated to performance. One side argues that
centralisation leads to better utilisation of resources and lower costs, while the other side argues
that decentralisation leads to faster reaction times and greater innovation.
The productivity factor is another important line of thought that attempts to explain variation
in performance from input cost of resources as ratio to output value of goods and services. Most
researchers agree that one-factor productivity per se is too simplistic on its own to explain vari-
ation in performance and suggest multifactor measures to capture the impact on protability
(Heap, 1992; Schefczyk, 1993; Oum and Yu, 1998). Schefczyk (p. 304) argues that even multifac-
tor productivity measures focusing on return to scale do not reect overall performance, especially
in terms of customer, service and competition issues. He used data envelopment analysis (DEA) to
address protability in the airline industry, which is described as a non-parametric approach to
measure productivity, while Oum and Yu (1998) used total factor productivity (TFP) for the same
industry. TFP is dened as the amount of aggregated output produced by a unit of aggregate in-
put. These approaches do provide results as to relationships between various resource inputs and
performance measures and thus heightening the interest in the underlying qualitative management
and strategic causal relationships that are usually outside the scope of such studies.
Finally, we have contingency theories (Wood, 1979; Miller, 1981) stating that there is an inher-
ent complexity and factors explaining variation in performance may be dependent on variables
such as strategy selected or industry specic factors. Support for this view can be sought in recent
research such as Wouters et al. (1999) that used benchmarking approach to determine operational
performance measures and their relationship with nancial performance in the transportation sec-
tor; and Otley (1999) researching management performance from the standpoint of managerial
control, suggesting an inductive framework that goes beyond measurement and towards a holistic
approach to performance management.
Having briey discussed dierent schools of thought, the purpose of the research that follows
was to come up with a bottom-up framework, separating micro- and macro-factors, when analys-
ing performance and apply it to rms in a specic industry segment.

2. Approach

2.1. Multilevel method

A multilevel method 2 allows a bottomup approach: going from the specic to the general.
The segregation of causes from symptoms of company failure was suggested by Argenti (1976).
The proposition has been much cited in the literature dealing with business failure, but has lacked

2
The use of the term multilevel in this research does not imply a stringent application of the concept as described
by Hox (1995) for example.
446 S.V. Gudmundsson / Transportation Research Part E 40 (2004) 443463

empirical investigation. We integrated this idea into a multilevel framework by dividing factors
into two principal levels: performance factors (PF) and performance indicators (PI). Performance
factors are dened as: factors that can be directly altered by management decisions resulting in,
either individually or collectively, performance change of the whole organisation or any of its sub-
components. While performance indicators are dened as: factors that change directly or indi-
rectly in consequence of change in one or more performance factors.
The PFs and the PIs can have two reminiscent relationships with performance: positive and
negative. PFs having higher emphasis by managers of non-distressed companies, are classied
as PF (+); and factors, having higher emphasis by managers of distressed companies, are classied
as PF ( ). For PIs the same applied, e.g. signicant positive dierence of the means, becomes PI
(+), and negative dierence, PI ( ). Furthermore, PIs, unlike PFs that are all low level factors,
could be considered multilevel factors: determined by the directness of the inuence enacted by
a change in a PF. The higher the level of a PI (i.e. further from direct inuence of a particular
PF) the more macro and important it is in terms of association with the overall performance state
of the rm. Therefore, it follows that the interrelationships become more complex as the level of
the PI is higher (macro level), making it more dicult for top management to use it to track the
inuence of alteration in individual PF (micro-level). For example, a change in aircraft passenger
load-factor on routei due to a change in frequency oered on the route would be regarded as low
level PI. The collective implications of changes in this particular PI, is then reected in higher level
PIs, such as nancial ratios.
It is important to emphasize, at this stage, that if a PF ( ) is emphasised by rms, that partic-
ular group of rms is not necessarily more prone to failure than a group of rms not emphasizing
a factor. Rather it means that the rms observed having nancial diculties are more likely to
emphasise this factor than rms not under distress, for the better or the worse. Thus, the inuence
of each PF must be established in terms of this dichotomy. Looking at the PIs a similar pedagogic
applies, although individual PIs can have the peculiar trait of showing positive change due to a
change in a PF that has a proportionally larger adverse impact on another PI. Take for example,
a reduction in ight frequency (#departures) on routej causing positive change in the passenger
load-factor (seat capacity/seat demand) (low level PI) because of reduction in seat capacity on
a route in a stable demand situation. Overall aircraft utilisation, a higher level PI, is, however,
negatively aected, if we assume that the excess aircraft hours caused by reduction in ight fre-
quency cannot be applied elsewhere in the airline route network. Hence, change in frequency
could be at a greater loss than benet if looking at aircraft utilisation opposed to the passenger
load-factor in isolation.
Given the coverage in this section we can postulate that by examining the multilevel relationship
between the PFs, and PIs, the explanatory power for variation in performance would be superior to
a single level empirical model or a traditional failure prediction model based on nancial ratios.

2.2. Sample and data gathering procedures

A questionnaire was used to gather data on 62 component items, composed of PF and PI


items related to airline management. The population was new airlines 3 established after marked

3
These can be termed new-entrant airlines and include value based airlines (VBA).
S.V. Gudmundsson / Transportation Research Part E 40 (2004) 443463 447

Table 1
Response statistics
Q-I Q-II Total
Response rate by airlines 65% (40/26) 42% (60/25)
Performance status
Non-distress 20 25 45
Distress 25 15 40
Geographical dispersion
US 22 20 42
EU 23 16 39
Other 0 4 4
Total 45 40 85

conditions in a domestic market had been liberalised. 4 The questionnaire was administered
twice: once in 1993 to 40 airlines and 242 individuals and again in 1998 to 60 airlines and 282 indi-
viduals. The surveys mailing strategy was designed in such a way that response rates by airlines
would be maximised. This led to an average of six mailings to each airline in 1993 and ve in 1998.
Response rate by airlines (see Table 1) was 65% in 1993 and 42% in 1998. The lower response rate
in 1998 was probably due to a larger number of airlines located in areas of the world not repre-
sented previously. Mean testing was used to assess whether there was dierence between the peri-
ods and signicant dierences were found for 10 items and ve items that were relevant to the
models presented here. These items are identity marked in the tables (see Tables 3 and 4).
Geographical dispersion of responses in the second survey was ranging from 50% US, 40% EU
and 10% other countries, while in the rst survey it was 49% US and 51% EU. In the more recent
survey about 37% of the airlines had had operating losses for two or more years preceding the
survey and were classied as distressed as a result. In the rst survey, this was dierent with a ma-
jor half or 55% of airlines being in distress.

2.3. Measures

The questionnaire asked respondents to indicate on a scale from 0 to 10, the importance placed
on various management factors in their airline. The items in the questionnaire were all developed
on the basis of literature research and comments on a pilot questionnaire. All items were segre-
gated into two groups posterior to the administration of the questionnaire according to the ability
of management to directly inuence their outcomes. The low level items, usually determinants of
airlines strategic characteristics, were labelled as PF items. Higher level items were considered
those that were more a consequence of manipulating PF items.

4
Population was identied through the March 1992, and 1997 issues of Flight International that publishes brief
historical and contact information for the world airlines.
448 S.V. Gudmundsson / Transportation Research Part E 40 (2004) 443463

2.4. Analysis

In order to reduce items to a manageable level, exploratory factor analysis was used to con-
struct descriptive factors for PF items and PI items. Multiple linear regression (MLR) was used
to test hypothesis pertaining to the relationship of the resulting factor PFs and PIs. On the
assumption that higher level PIs are better predictors of overall performance than lower level
PFs, logistic regression (LR) was used to segregate the relationship of the PI factors with the sam-
ple rms dichotomous performance state: distressed or non-distressed.

3. Results

3.1. Descriptive statistics and reliability estimates

When performing the factor analysis, coecient alpha was used to assess scale reliability for
each sub-scale constructed, based on the latent factors. Reliability for sub-scales varied for alpha
from 0.52 to 0.86. Nunnally (1978) suggested that alpha of 0.500.60 being sucient at the outset
of research and all 17 factors extracted met the minimum criteria. One factor had only one item
loading and was therefore, excluded from further analysis. A number of factors had three items
loading, but one item with a low loading (<.40) that was deleted from the analysis based on
Churchill (1979). This deletion raised the number of factors having two items only, from four
to six. Spector (1992) argues that factors of less than three items are not likely to be useful. This
argument was deemed to be valid in terms of the two-item scales needing further development in
future research, but not substantiating the deletion of the factors from further consideration.
Correlation matrixes were constructed to see if any of the items used in the scales had non-sig-
nicant correlation with all of the other items. Nunnally (1978) recommended that items with con-
sistently low correlation (<.30) to be deleted. This led to a number of deletions from the item list.
The KeyserMayerOlkin (KMO) was used to test for sample adequacy. For the PI items the
result was on the better side of middling (0.76). While for the PF items it was mediocre (0.65). As
the KMO is well above the unacceptable level (0.50) it was deemed that the factor analysis should
go ahead with the existing sample size. The statistic for Bartlett test of sphericity posed also nei-
ther problems in the use of the factor model for PI items (p 6 .001) nor PF items (p 6 .001).
Table 2 shows the means, standard deviations, correlation, and reliabilities for the studys var-
iables. The table also shows acceptable interrater reliabilities for the aggregated variables.

3.2. Factor extraction

The maximum-likelihood method was used for factor extraction in both factor models and var-
imax was used for orthogonal rotation, which enhances the interpretability of the factors.
The PF level exploratory factor analysis (see Table 3) was performed on observed items that can
be directly manipulated by managers. The factor analysis resulted in 10 factors.
The rst factor was labelled service features (a = .86) explaining 22.4% of the variance and
consisting of ve items. The name of the factor was grounded on the fact that most of the items
were related with actions important to serve business passengers. An airline that is high on this
Table 2

S.V. Gudmundsson / Transportation Research Part E 40 (2004) 443463


Factor means, standard deviations and Pearson correlationsa,b
Mean SD Correlationsb
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
1. Productivity 7.54 1.46 (.78)
2. Brand image 7.70 1.61 .06 (.67)
3. Decentralisation 6.00 1.67 .13 .08 (.69)
4. External inuencing 6.39 2.05 .06 .03 .04 (.63)
5. Market power 6.20 2.00 .03 .05 .02 .07 (.52)
6. Channel power 7.51 1.80 .00 .14 .03 .01 .05 (.56)
7. Service features 5.87 2.41 .21 .35 .10 .03 .32 .37 (.85)
8. Operations eciency 6.66 2.08 .33 .07 .44 .06 .13 .02 .01 (.67)
9. Cost competitiveness 7.94 1.43 .14 .13 .41 .38 .26 .00 .01 .01 (.69)
10. Performance 5.82 2.49 .12 .19 .05 .14 .12 .08 .03 .01 .00 (.82)
incentives
11. Gearing 6.69 1.86 .44 .07 .14 .09 .07 .02 .04 .06 .02 .01 (.67)
management
12. Integration 6.62 1.63 .07 .30 .13 .17 .31 .11 .00 .01 .01 .01 .03 (.71)
orientation
13. Labour exibility 5.34 1.89 .32 .25 .00 .11 .01 .40 .02 .02 .03 .01 .04 .01 (.62)
14. Marketing 6.48 1.74 .18 .08 .09 .03 .08 .24 .03 .01 .04 .01 .01 .02 .02 (.54)
orientation
15. Cost reduction 7.41 2.03 .04 .10 .02 .12 .21 .03 .05 .05 .02 .01 .04 .01 .02 .03 (.62)
16. External growth 3.71 1.74 .03 .03 .10 .10 .25 .06 .05 .00 .01 .02 .03 .04 .05 .03 .03 (.55)
a
Results for item correlation are available from the author.
b
Reliabilities are in parentheses. For all correlation above .21, p 6 .05, one-tailed tests; for all correlation above .31, p 6 .01, one-tailed tests; for all correlations
above .35, p 6 .001, one-tailed tests.

449
450 S.V. Gudmundsson / Transportation Research Part E 40 (2004) 443463

Table 3
Factor results for performance factors (PF)
Determinant Component Component Eigenvalue Loading Expl.
factors itemsa Mean SD var.

Service features 7.61 22.39


Alpha = .86 Frequent yer programs (+) 5.26 3.41 .77
Business passengers ( ) 6.89 3.18 .75
Yield management system (+) 7.01 2.78 .64
Hub and spoke operations ( ) 5.06 3.48 .46
Feeder airline agreements (+) 4.42 3.17 .46
Operations eciency 3.76 11.07
Alpha = .67 Matching of aircraft size 7.38 2.46 .88
with route requirement
Interlining agreements ( ) 5.76 2.94 .59
Acquisition of airport slots (+) 6.36 3.38 .42
Cost competitiveness 2.36 6.95
Alpha = .69 Increase margins (+)b 8.24 1.79 .86
Cost controlb 8.85 1.42 .62
Competitor analysis (+) 6.74 2.16 .47
Performance incentives 1.95 5.73
Alpha = .82 Employees incentive program (+) 5.52 2.82 .76
Managers incentive program (+) 5.86 2.89 .76
Gearing management 1.92 5.64
Alpha = .66 Debt reduction ( )b 6.11 3.08 .88
Acquisition of new aircraft (+) 7.13 2.77 .50
Forec. adv. e. of the econ. on the airl. (+) 6.42 2.15 .48
Fuel costs (+) 7.06 2.17 .43
Integration orientation 1.73 5.07
Alpha = .71 Logistics systems (+) 6.42 2.21 .72
Control systems (+) 6.88 1.97 .53
Inter departmental communication 6.55 1.99 .41
Labour exibility 1.39 4.09
Alpha = .62 Flexible job descriptionsb 6.39 2.47 .88
Job rotation ( ) 4.05 2.32 .55
Marketing orientation 1.24 3.64
Alpha = .54 Media advertising (+) 6.20 2.30 .90
Market research ( ) 6.37 2.16 .47
Price leadership in served markets ( ) 6.87 2.74 .43
Cost reduction 1.14 3.36
Alpha = .62 Cost reduction (+) 8.14 1.94 .76
Reduction of labour costs (+) 6.69 2.76 .65
External growth 1.03 3.02
Alpha = .55 Merger/acquis. to gain market share ( ) 3.11 2.97 .63
Alliance with the incumbents ( )b 5.34 3.05 .52
Diversication into other industries ( ) 1.50 2.21 .41
a
Sign in parentheses shows dierence of means, if signicance is <.10, then no sign is provided. (+) means a
signicantly higher mean for non-distressed airlines.
b
Signicant dierence between 93 and 98 surveys.
S.V. Gudmundsson / Transportation Research Part E 40 (2004) 443463 451

dimension will tend to believe in the importance of high yield passengers. The factor loading ran-
ged from 0.46 to 0.77.
The second factor was labelled operations eciency (a = .67) and explains 11.1% of the total
variance. This dimension consists of three items with factor loading ranging from 0.42 to 0.88.
One item had a loading <.40 and was deleted. Items loading on this factor seemed to measure
issues associated with setting up and maintaining ecient routes. An airline that scores high on
this factor will tend to believe in the importance of setting up and maintaining ecient route
structures with well connected ights. The third factor labelled cost competitiveness (a = .69) ex-
plained 7.0% of the total variation. This dimension consists of three items with factor loading
ranging from 0.47 to 0.86. One item had a loading <.40 and was deleted from consideration.
An airline scoring high on this factor tends to believe in the importance of relative cost competi-
tiveness, by researching competitors actions and keeping costs in check.
The fourth factor was labelled performance incentives (a = .82) and explained 5.7% of the total
variation. This dimension consisted of two items with approximately the same factor loading of
0.76. One item had a loading <.40 and was deleted from further consideration. An airline scoring
high on this factor will emphasise incentive programs as a tool to motivate employees. The fth
factor labelled gearing management (a = .66) explained 5.6% of the variation. This dimension
consisted of four items with loading ranging from 0.43 to 0.88. An airline scoring high on this
factor tends to have high awareness of the impact of economic cycles and debt in the eet acqui-
sition process.
The sixth factor labelled integration orientation (a = .71) explained 5.1% of the total variation.
It consists of three items with loading ranging from 0.41 to 0.72. An airline that scores high on this
factor tends to believe in the importance of organisational integration as well as control through
information systems.
The seventh factor labelled labour exibility (a = .62) explained 4.1% of the total variation. It
consists of two items with loading of 0.88 and 0.55. An airline that scores high on this factor tends
to believe in the importance of exible workforce and job rotation to motivate employees and
achieve responsiveness to customers needs, e.g. service quality. The eight factor labelled market-
ing (a = .54) explained 3.6% of the total variation. The factor consists of three items with loading
ranging from 0.43 to 0.90. One item had loading <.40 and was therefore deleted from considera-
tion. An airline scoring high on this factor places high importance on stimulating demand though
advertising and price leadership that is well grounded through market research. The ninth factor
labelled cost reduction (a = .62) explained 3.4% of the total variation. Two items loaded on this
factor: 0.76 and 0.65. Airlines scoring high on this factor are in the process of reducing costs op-
posed to keeping costs in check, as a result, high ranking implies restructuring. The tenth factor
labelled as external growth (a = .55) explained 3.0% of the total variation. This factor consists of
three items with loading ranging from 0.41 to 0.63. Airlines scoring high on this factor tend to
seek growth through external means such as mergers and alliances.
The PI level exploratory factor analysis (see Table 4) was performed on observed items that can
normally not be directly manipulated by airline managers.
The rst factor labelled productivity (a = .78) explained 28.6% of the total variation. This fac-
tor consists of six items with loading ranging from 0.40 to 0.73. Airlines scoring high on this factor
tend to believe that high productivity, quality and long-term protability is achieved through
shared employee beliefs, reected in vision and culture. The second factor labelled brand image
452 S.V. Gudmundsson / Transportation Research Part E 40 (2004) 443463

Table 4
Factor results for performance indicators (PI)
Determinant factors Component itemsa Component Eigenvalue Loading Expl.
Mean SD var.

Productivity 5.71 28.56


Alpha = .78 Employees productivity (+) 8.24 1.72 .73
Shared company vision (+) 6.66 2.59 .69
Company culture (+) 7.52 2.02 .67
Aircraft utilisation 8.54 1.41 .52
Service quality (+) 8.13 1.86 .52
Long-term rather than short-term prots (+)b 6.18 2.74 .40
Brand image 2.21 11.05
Alpha = .67 Brand image (+) 7.70 2.19 .73
Favourable attitude of travel agents (+) 7.31 2.39 .60
Decentralisation 1.50 7.52
Alpha = .69 Employees autonomy to take ( ) decisions 6.52 1.69 .69
Decentralised organisation ( ) structure 4.81 2.69 .60
External inuencing 1.38 6.90
Alpha = .64 Managements external contacts (+) 6.33 2.25 .80
Inuencing government policy on aviation (+) 6.45 2.56 .53
Market power 1.22 6.08
Alpha = .52 Market share ( ) 5.91 2.68 .94
Achieving critical mass ( ) 5.67 2.98 .43
Investors attitudes towards the airline ( ) 7.02 2.72 .42
Channel power 1.00 5.01
Alpha = .56 Computer reservation systems ( ) 7.26 2.61 .87
Passenger load factors ( ) 7.58 2.07 .41
a
Sign in parentheses shows dierence of means, if signicance is <.10, then no sign is provided. (+) means a
signicantly higher mean for non-distressed airlines and vice versa.
b
Signicant dierence between 93 and 98 surveys.

(a = .67) explained 11.0% of the total variation. This factor consists of two items with loading of
0.73 and 0.60. A third item had a loading <.40 and was deleted from consideration. Airlines scor-
ing high on this factor tend to emphasise brand image, believing that it has importance in building
customer loyalty and product preference for example in the distribution network: favourable at-
titude of travel agents. The third factor labelled Decentralisation (a = .69) explained 7.5% of the
total variation. It consists of two items with loading ranging from 0.60 to 0.69. Airlines scoring
high on this factor tend to believe in employees autonomy to achieve faster reaction times and
greater innovation. The fourth factor labelled external inuencing (a = .64) explained 6.9% of
the total variation. The factor consists of two items with loading of 0.80 and 0.53. Airlines scoring
high on this factor tend to believe in the importance of facilitating the airlines well-being through
external inuence: employee networking and inuencing air transport policy. The fth factor la-
belled market power (a = .52) explained 6.1% of the total variation. It consists of three items with
loading ranging from 0.42 to 0.94.
Airlines scoring high on this factor tend to believe in the importance of size in order to attract
investment in the airline. The seventh factor channel power (a = .56) explained 5.0% of the total
variation. The factor consists of two items with loading of 0.87 and 0.41. Airlines scoring high on
S.V. Gudmundsson / Transportation Research Part E 40 (2004) 443463 453

this factor tend to believe in the importance of computer reservation systems to boost passenger
load factors. This indicates an acceptance of channel power to facilitate sales of seats, despite high-
er associated costs compared to direct sales: telephone and Internet.

3.3. Hypotheses

A number of hypotheses were formulated based on literature review and case studies in order to
predict relationships between the extracted factors across the two aggregated factor levels: PF and
PI.
The factor productivity is assumed to represent how airlines believe that high productivity, qual-
ity and long-term protability is achieved through shared employees beliefs, reected in the vision
and culture. Therefore, it is assumed that productivity is inuenced by rst level factors that have
positive impact on employees productivity, while factors such as cost reduction will either have
negative or non-signicant relation with productivity. This stems, from the assumption that em-
ployee redundancies or other forms of cost reduction will create tensions (Doherty and Horsted,
1996) associated with reduced productivity in the respondents mind. The assumption is made in
qualitative terms regardless of whether such negative productivity impact occurs or not. What is
clear, however, is that cost reductions and employee redundancies will most likely aect the busi-
ness culture and vision in an adverse way in the short-term. Finally, gearing management can have
major impact on the rms aggregated inputoutput ratio. For example, through poor timing of
aircraft orders with delivery lag, frequently causing deliveries during economic downturns (Liehr
et al., 1998).

H1a: Service features, operations eciency, performance incentives, integration orientation, labour
exibility and gearing management, will be signicantly related to productivity.
H1b: Productivity is positively related to non-distress.

The factor brand image is assumed to represent airlines that believe in the importance of the
brand to generate customer loyalty and preference. For airlines this factor would facilitate pene-
tration in the distribution network through favourable attitude of travel agents (TAs). Good
brand image works as selection advantage in a case of two similar choices. This being, not only,
an advantage with the TAs, but also the passenger depending on, who exercises the decision
power. The creation of brand image in the minds of these two groups is probably somewhat dif-
ferent although no studies exist to support the assumption. It is expected that the factor is signif-
icantly related to service features because service is an important function of brand image.
Another service element is operations eciency, which provides a match with what the TA is look-
ing for on behalf of the customer, such as: good connections (function of interlining agreements
and alliances), convenient departure times (function of airport landing slots) and frequency (func-
tion of matching aircraft size with route requirement). The last item listed can imply two things.
First, more frequency with smaller equipment is more favourable, especially in terms of business
passengers. Second, the airline should be competent in adjusting aircraft size to demand to reduce
the probability of passenger diversion from specic ights (involves more time spent on behalf of
the TA nding a ight). Brand image is also expected to be related to labour exibility, as positive
sta willing to walk that extra mile for the benet of the passenger has a strong impression on the
454 S.V. Gudmundsson / Transportation Research Part E 40 (2004) 443463

customer. This is what Carlson (1987) termed as turning the organisational pyramid upside down,
empowering the front-line sta to take unprecedented decisions for the customers benet at their
specic level. Further, if the concept is expanded one can assume that exible job descriptions in-
crease the responsiveness of the organisation, a potentially important element in facilitating brand
image and proactive service provision. Finally, marketing orientation is expected to be related to
brand image. The factor marketing is composed of items dealing with advertising and market re-
search, both of which shape the ability of the airline to build its brand image. In the past low cost
carriers, with PeoplExpress being the prime example, have had brand image conicts due to
changes in strategy, such as no-frills to frills, projected to the market through advertising.

H2a: Service features, operations eciency, labour exibility, and marketing orientation, will be
signicantly related to brand image.
H2b: Brand image is positively related to non-distress.

Airlines scoring high on the factor decentralisation tend to believe in employees autonomy
(Thomas and Velthouse, 1990) to achieve higher performance through faster reaction times and
better innovation. As such the factor is expected to be signicantly related with labour exibility.
Labour exibility increases the necessity to co-operate. Job rotation, as a function of decentrali-
sation introduces to the employee dierent functions and stimulates proactive behaviour and
cross-departmental knowledge (Spreitzer, 1995). Labour exibility as an element of decentralisa-
tion is also a crucial cost saving tool, by cross-training sta to perform various functions and
make decisions. Hence, decentralisation is expected to be signicantly related to cost
competitiveness.

H3a: Labour exibility and cost competitiveness, will be signicantly related to decentralisation.
H3b: Decentralisation is positively related to non-distress.

Airlines scoring high on the factor external inuencing tend to believe in the importance of facil-
itating the airlines well-being through external inuence. Policy making in air transport is high on
the agenda in most countries, although, one can assume that the nature of airlines inuence on
governments and agencies has changed in liberalised and deregulated markets. For low-cost or
value based airlines, opposed to large incumbents, we can expect some variation in interest, with
two issues high on the agenda: CRS biases (Beuvais, 1992) and lack of landing slots at airports
(Morrell, 1998). Therefore, it is expected that operations strategy will be signicantly related to
external inuencing. The external inuencing inuences costs, and governmental inuence, federal
or local, is often necessary when entering or protecting markets. Hence, the external inuencing is
expected to be positively related with cost competitiveness.

H4a: Operations eciency and cost competitiveness, will be signicantly related to external
inuencing.
H4b: External inuencing is positively related to non-distress.

Airlines scoring high on the factor market power tend to believe in the importance of size in
order to attract investment. As a result, it is expected that factors dealing with the network, com-
S.V. Gudmundsson / Transportation Research Part E 40 (2004) 443463 455

petitiveness, routes, growth, integration and reputation in the nancial community, are impor-
tant. The following factors are expected to have signicant relationship with market power: serv-
ice features, operations eciency, cost competitiveness, integration orientation, marketing
orientation, cost reduction and external inuencing. Service features includes items such as hub
and spoke networks that are important to achieve economies of density and scope (Gimeno
and Woo, 1999; Bania et al., 1998) that are usually associated with market power. Operations e-
ciency includes items such as acquisition of slots, necessary to boost market-share. Cost compet-
itiveness is an important element in achieving market power, allowing the airline to oer
competitive prices in order to build market share. The PIMS (Buzzel and Gale, 1987) program
identied low prices as a vehicle towards market-share, without cost competitiveness and quality,
as being a non-sustainable strategy. Integration orientation is necessary to keep up with growth,
especially if it grows fast, which is a characteristic of market-share driven airlines. Marketing ori-
entation is expected to be related to market power, even though these two can be considered quite
distinctive in terms. Cost reduction, such as labour cost reduction, is especially important to show
nancial control (Flint, 1999) resulting in favourable image within the nancial community, facil-
itating capital access for expansion programs. External growth is the nal factor expected to be
related to market power. This factor was not emphasised highly by most respondents as can be
seen from the item averages. However, alliances are most prominent and especially important
for an airline wishing to achieve critical mass quickly.

H5a: Service features, operations eciency, cost competitiveness, integration orientation, marketing
orientation, cost reduction and external growth, will be signicantly related to PI market
power.
H5b: Market-power is positively related to distress.

Airlines scoring high on the factor channel power tend to believe in the importance of computer
reservation systems to boost passenger load factors. For low-cost or value based airlines the
importance placed on this factor could be diverse as some do not participate in computer reser-
vation systems (CRSs), while those that do, have had strong views on partiality of the CRS owner
airlines (Feldman, 1997). It is expected, that this factor, is signicantly related to service features,
operations eciency and marketing orientation. The factors service features assumes high conven-
ience for passengers through ease of product access. Computer reservation systems are important
to communicate information about products and to facilitate seamlessness in meeting complex
customer requirements. Operations eciency is composed of items pertaining to capacity, connec-
tions, and convenient departure times and routes through availability of slots. As all of these items
are communicated through the CRS as a distribution tool for the product, it is expected that oper-
ations eciency is signicantly related to channel power. The nal expected factor to be signi-
cantly related with channel power is marketing orientation. Advertising, market research, and
price leadership that compose the factor, are all items that are communicated through the
CRS, thus, impacting loads.

H6a: Service features, operations eciency and marketing orientation, will be signicantly related
to channel power.
H6b: Channel power is positively related to non-distress.
456 S.V. Gudmundsson / Transportation Research Part E 40 (2004) 443463

The testing of the a hypotheses was conducted through multiple linear regression analysis to
explore further the strengths of the predetermined relationships reected in the hypotheses estab-
lished in this section, while the b hypotheses were tested by the means of logistic regression.
3.4. Multiple linear regression analysis

Stepwise approach was used for variable entry in the multiple linear regression (MLR) analysis
(see Table 5). Colinearity was examined through tolerance and VIF analysis and no factor posed
Table 5
The relationship between PFs and PIs: MLR results
DR2 b SE Beta t
I. Dependent variable: productivity
Gearing management .195 .409 .080 .45 5.15***
Operations eciency .130 .333 .080 .36 4.18***
Labour exibility .097 .288 .080 .31 3.62***
Service features .037 .185 .081 .20 2.29*
Marketing orientation .033 .164 .078 .18 2.11*
R2 .492 Adj. R2 .455 F = 13.19***
II. Dependent variable: brand image
Service features .124 .319 .090 .35 3.54***
Integration orientation .090 .262 .087 .30 2.99**
Labour exibility .055 .208 .089 .23 2.34*
Performance incentives .040 .173 .087 .20 2.00*
R2 .310 Adj. R2 .270 F = 7.74***
III. Dependent variable: decentralisation
Operations eciency .190 .351 .078 .43 4.51***
Cost competitiveness .163 .327 .077 .40 4.23***
R2 .353 Adj. R2 .334 F = 19.33***
IV. Dependent variable: external inuencing
Cost competitiveness .314 .089 .38 3.52***
R2 .147 Adj. R2 .135 F = 12.40***

V. Dependent variable: market power


Service features .103 .298 .097 .30 3.06**
Integration orientation .097 .292 .094 .30 3.09**
Cost competitiveness .067 .258 .095 .26 2.72**
External growth .052 .240 .099 .24 2.42*
Cost reduction .041 .197 .094 .20 2.08*
R2 .360 Adj. R2 .313 F = 7.65***
VI. Dependent variable: channel power
Labour exibility .159 .373 .094 .39 3.98***
Service features .128 .343 .095 .35 3.60***
Marketing orientation .048 .207 .092 .22 2.25*
R2 .336 Adj. R2 .307 F = 11.79***
*
p 6 0.05.
**
p 6 0.01.
***
p 6 0.001.
S.V. Gudmundsson / Transportation Research Part E 40 (2004) 443463 457

problems in the analysis. The MLR models explanation power of the dependent variables ranged
from 7% to 49%.
The outcomes of productivity. The results of the stepwise MLR analysis supported hypotheses
H1a and the factors in the model explained 49% of the variation in the depended variable. The
results showed that productivity was signicantly related to gearing management (b = .41,
p 6 .001), operations eciency (b = .33, p 6 .001), labour exibility (b = .29, p 6 .01) and service
features (b = .19, p 6 .05). Another unexpected factor that entered the model was marketing ori-
entation (b = .16, p 6 .05). Variable contribution to the model was detected through change in R2
(gearing management, DR2 = .195, p 6 .001; operations eciency, DR2 = .130, p 6 .001; labour
exibility, DR2 = .097, p 6 .01; service features, DR2 = .037, p < .05; and marketing orientation,
DR2 = .033, p 6 .05). Two expected factors performance incentives and integration orientation
did not enter the model.
The outcomes of brand image. The results of the MLR analysis partially supported hypotheses
H2a and the factors in the model explained 31% of the variation in the depended variable. The
results showed that brand image was signicantly related to service features (b = .32, p 6 .01)
and labour exibility (b = .21, p 6 .05). However, two other factors contributed to the explained
variation in brand image, namely integration orientation (b = .26, p 6 .01) and performance
incentives (b = .17, p 6 .1). The change in R2 was examined to detect each factors contribution
to the explained variation in R2 (service features, DR2 = .124, p 6 .01; integration orientation,
DR2 = .090, p 6 .01; labour exibility, DR2 = .055, p 6 .05; performance incentives, DR2 = .040,
p 6 .1). Two expected factors did not enter the model, operations eciency and marketing orien-
tation, while two unexpected factors entered: integration orientation and performance incentives.
The outcomes of decentralisation. The results of the MLR analysis partially supported hypoth-
eses H3a and the factors in the model explained 35% of the variation in the depended variable. The
results showed that decentralisation was signicantly related to operations eciency (b = .35,
p 6 .001) and cost competitiveness (b = .33, p 6 .001). The change in R2 was examined to detect
each factors contribution to the explained variation in R2 (operations eciency, DR2 = .19,
p 6 .001; cost competitiveness, DR2 = .16, p 6 .01). Unexpected factor operations eciency entered
the model instead of the expected labour exibility.
The outcomes of external inuencing. The results of the stepwise MLR analysis supported
hypotheses H4a and the factors in the model explained 15% of the variation in the depended var-
iable. The results showed that external inuencing was signicantly related to cost competitiveness
(b = .31, p 6 .01). No other factors entered the model.
The outcomes of market power. The results of the MLR analysis supported hypotheses H5a and
the factors in the model explained 36% of the variation in the depended variable. The results
showed that market power is signicantly related to service features (b = .30, p 6 .01), integration
orientation (b = .29, p 6 .01), cost competitiveness (b = .26, p 6 .01), external growth (b = .24,
p 6 .05) and cost reduction (b = .20, p 6 .05). Variable contribution in the model was detected
through change in R2 (service features, DR2 = .103, p 6 .01; integration orientation, DR2 = .097,
p 6 .01; cost competitiveness, DR2 = .067, p 6 .05; external growth, DR2 = .052, p 6 .05; cost reduc-
tion, DR2 = .041, p 6 .05). Two expected factors did not enter the model, operations eciency and
marketing orientation.
The outcomes of channel power. The results of the stepwise MLR analysis supported hypotheses
H7a and the factors in the model explained 34% of the variation in the depended variable. The
458 S.V. Gudmundsson / Transportation Research Part E 40 (2004) 443463

Table 6
The relationship of PIs with performance: LR resultsa
Independent variables Coe. (b) S.E. Wald R
*
Productivity .618 .315 3.840 .140
Brand image .841* .378 4.945 .177
Decentralisation .517 .348 2.214 .048
External inuencing .416 .367 1.284 .000
Market power .641* .308 4.326 .157
Channel power .345 .294 1.379 .000
Constant .098 .278 .1236
2 Log Likelihood 77.36
Nagelkerke R2 .293
Goodness-of-t testb .325
*
p 6 0.05.
a
Dependent variable = nancial performance (1 = distress, 0 = non-distress).
b
Hosmer and Lemeshow.

results showed that channel power was signicantly related to labour exibility (b = .37, p 6 .001),
service features (b = .34, p 6 .01) and marketing orientation (b = .21, p 6 .05). Variable contribu-
tion to the model was detected through change in R2 (labour exibility, DR2 = .159, p 6 .001; serv-
ice features, DR2 = .128, p 6 .01; marketing orientation DR2 = .048, p 6 .05). An unexpected factor
entered the model, labour exibility instead of the expected operations eciency.

3.5. Results of logistic regression

The non-stepwise logistic regression model (see Table 6) had signicant model chi-square
(p 6 .001) and good Nagelkerke R2 (0.29). Hypotheses H1b was conrmed,productivity has ex-
pected negative sign, which indicates positive relationship (p 6 .05) with non-distress. Hypotheses
H2b was conrmed, brand image has expected negative sign, which indicates positive relationship
(p 6 .05) with non-distress. Hypotheses H3b was not conrmed, decentralisation has unexpected
positive sign, which indicates positive relationship with distress. The relationship was, however,
non-signicant. Hypotheses H4b was not conrmed, external inuencing has expected negative
sign, which indicates positive relationship with non-distress. However, the relationship was
non-signicant. Hypotheses H5a was conrmed, market power has expected positive sign, which
indicates positive relationship (p 6 .05) with distress. Hypotheses H6a was not conrmed, channel
power has unexpected positive sign, which indicates positive relationship with distress. The rela-
tionship was however non-signicant.

4. Discussion

4.1. Summary of ndings

The major ndings (see Fig. 1) are that a PI factor labelled productivity has the strongest rela-
tionship with PFs labelled gearing management, operations eciency, labour exibility, service fea-
S.V. Gudmundsson / Transportation Research Part E 40 (2004) 443463 459

Service features
-.21
+.33 +.35
Operations efficiency
Productivity

Cost competitiveness
Brand image
+.19 +.41
Performance incentives

+.44 +.38
Decentralization
External influencing Non-distress
Gearing management
+.44
+.25
-.30
Integration orientation Decentralization
+.26 Distress
+.31
+.32 +.32
Labor flexibility

+.18 +.40 Market power


Marketing orientation +.24
+.37
+.21
Cost reduction Channel power

+.25
External growth

Fig. 1. Results of regression analysis: MLR and LRa. a Lines show signicant relationships in the models (p 6 0.05, or
better), broken lines show non-signicant relationships. Direction (+/ ) of relationship is shown with correlation
coecient where appropriate.

tures and marketing orientation. The rst factor gearing management that explains comparatively
most of the variation in productivity, can be related to the necessity to manage gearing in order to
prepare for industry downturn and preserve the overall airline productivity. The second factor
operations eciency deals with the matching of resources with requirements, binding together
the various components of a route to provide optimum service given the resources available.
As such, it is quite logical that operations eciency has positive relationship with productivity.
Labour exibility is another clearly related factor, as has been demonstrated by so many low cost
airlines, such as Southwest Airlines that is able to achieve lower costs through greater sta ex-
ibility. Service features had a negative relation with productivity. A justiable result having in
mind that the factor service features is composed of items related to business passengers that re-
quire more service. The last factor market-demand deals with the generation of demand through
advertising but also through research and price leadership. As such, it has a spill over eect on
how the employees view the airline they work fora well researched marketing campaign rein-
forces the employees sense of purpose. What is more such campaign is essential to generate de-
mand to maintain utilisation of resources, especially at the outset of opening new routes.
There were four PFs related to the PI factor brand image: service features, integration orienta-
tion, labour exibility, and performance incentives. Service features deals with what product the air-
line is oering, therefore, logically associated with brand image. Positive brand image has
460 S.V. Gudmundsson / Transportation Research Part E 40 (2004) 443463

important impact on demand, whose interaction with the airline is in most cases through travel
agents. The pull of brand image is therefore through systems that must be well integrated to rein-
force the image created through an eective service features. Labour exibility is crucial for cre-
ating a good service impression. Flexibility allows employees to react quickly and eectively to
acute situations, creating a good impression with the passenger, e.g. employees ability to react
to special situations will either reinforce or destroy a good brand image during the consumption
stage. Finally, performance incentives contributed to explaining variation in brand image. Here
again the expected relationship is the motivation of employees to provide good impression in
the process of their jobsto walk that extra mile to make a dierence.
Two PFs contributed to the explanation of the variation in the PI factor decentralisation: oper-
ations eciency and cost competitiveness. Operations eciency should have a relationship with
decentralisation assuming that it increases labour exibility. Cost competitiveness should have
an even stronger relationship as we expect that decentralisation will have positive impact on
the airlines overall eectiveness.
The factor external inuencing was related to one PF, cost competitiveness. Here it seems that
managements external contacts and ability to inuence government policy can have cost implica-
tions for the airline. It was expected that there would be dierence between European and US car-
riers on the items composing the factor external inuencing, but a t test showed a non-signicant
dierence between the two groups. As such the factor must, therefore, be viewed as having com-
munality between both regions and constitute an element in achieving cost competitiveness.
Five PFs contributed to explaining the variation in the PI factor market power: service features,
integration orientation through systems, cost competitiveness, external growth and cost reduction.
Market power represents the emphasis the airline levies on size and investors attitudes towards
the airlines. The investors play a crucial role in raising capital necessary for market-share building.
One vehicle towards market power is broad target market through many service features. Integra-
tion orientation becomes increasingly important as the organisation grows quickly in a market
power culture. Thus, it is not surprising to see a relationship between market power (size emphasis)
and integration orientation (organisational integration). Cost competitiveness is another important
vehicle towards market power as it is unlikely that the airline will reach size without cost compet-
itiveness. This is more relevant to low-cost airlines as such airlines emerge from size disadvantage
(subject to the market power of incumbents) and must therefore achieve substantial advantage on
this factor. External growth is an external means to an end, where the airline emphasizes quick
expansion through alliances and mergers. Diversication into other industries one of the items
in the sub-scale was not emphasised much by any of the airlines, but showed nevertheless corre-
lation with the underlying factor. Cost reduction was the fth and nal factor that was related with
the factor market power. Here it was assumed that cost reduction is a vehicle towards positive im-
age in the nancial community to maintain capital inux for further expansion and increased stay-
ing power during industry recession.
Three PFs contributed to the explanation of the variation in the PI factor channel power, labour
exibility, service features and marketing orientation. An obvious theoretical explanation for the
explanatory power of labour exibility to channel power was not found. However, service features
are an important element in the channel strategy as such. This stems from the fact that distribu-
tion system presence is not enough, there has to be a sellable product in order for travel agents to
search the CRS and sell the airline. The third and last item, marketing orientation is related, as it
S.V. Gudmundsson / Transportation Research Part E 40 (2004) 443463 461

deals with targeted advertising of the airlines products, with most of the selling going through the
TAs and the CRS.
The relationship of the identied PIs with performance was as expected with two exceptions
channel power and decentralisation which signs indicated positive relationship with distress, but
both had non-signicant coecients. However, variables with signicant coecients showed rela-
tionships as expected. Thus, the results show that airlines placing relatively high emphasis on pro-
ductivity and brand image are less likely to be distressed, while airlines placing relatively high
emphasis on market power are more likely to be distressed. As market power is composed of items
that imply market-share building and mass, the ndings support Appiah-Adu and Ranchhod
(1998) that distinguished between market orientation and market dominance and found the for-
mer associated with better relative performance. Furthermore, brand image can also be associated
with market orientation, explaining the positive relationship with non-distress. External inuenc-
ing had expected negative sign, but a non-signicant coecient.

4.2. Limitations, strengths and future directions

The results should be viewed in the light of the datas limitations. The population of new-en-
trant airlines necessitates a combination of longitudinal and cross-sectional approach to boost
the number of cases due to both high entry and exit rates, but relatively short life-cycles. To val-
idate the scales with higher degree of certainty still more cases are needed, although, this study
almost reached the recommended minimum of 100 cases (Spector, 1992) and was well above
the minimum criteria of the KMO test for sampling adequacy.
The longitudinal approach necessitates high degree of concurrence between responses over
time. The study showed that similarity in rating strengths is de facto for the two studies as few
items showed statistical dierence of the means, implying consistency in the constructs. For items,
showing signicant dierence the explanation appears to be that the former study was conducted
during an industry recession and the second study during an industry up-turn. Hence, items asso-
ciated with the nancial aspects of airline management were the only ones showing a signicant
dierence between the two studies.
The exploratory approach rather than conrmatory approach was selected due to the novelty
of the approach and lack of related research on the industry under observation. A logical next step
would be to improve the weakness posed by the scales having two items and use a conrmatory
approach to establish further the theoretical relationships.
Although the questionnaire instrument administered to the airline managers may still need
improvement it is important as an initial step to identify useful constructs and underlying factors
that can provide for a standard qualitative scale measuring airline management. What is impor-
tant to note is that the scale takes external approach in order to make the association of actual
variation in strength applied to the items easier to associate with the airlines performance. This
approach was undertaken to reduce the impact of social desirability on the rating strength, e.g.
ratings according to what is generally accepted in the industry or to make one look good exter-
nally. However, it must also be made clear that, as with all factor-analytical, studies there may
be a number of dierent factors possible based on the data. This fact does, however, not under-
mine the value of factor analytical studies, as a factor that is proved to be stable in separate stud-
ies and measure what it allegedly is supposed to measure is valuable to researchers.
462 S.V. Gudmundsson / Transportation Research Part E 40 (2004) 443463

The research has attempted to link together the PFs and the PIs as single level factors. The PIs
could in future research be segregated into several levels on priori. In addition, using the PF con-
cept on purely quantitative data at the PF level and several PI levels, as demonstrated in the exam-
ples, is denitely a viable area for comparative qualitative and quantitative studies.

4.3. Conclusion

The general nding of this study is that airlines under distress are more likely to have pursued
market power tactics in an attempt to achieve fast growth and superior market-share, while non-
distressed carriers have placed greater importance on productivity and brand image. The last two
being, in prior research, associated with sustainable superior performance. As a consequence, this
study supports prior research that has showed similar relationships in other industries pertaining
to market orientation versus market power orientation. This study, however, enhances our under-
standing of the productivity factor in industries with high xed costs and a perishable product.
Productivity emphasis was shown in this study to distinguish between distressed versus non-dis-
tressed airlines in the model.
It is hoped that this attempt at identifying factors for qualitative research into airline manage-
ment using the PF approach will spark interest in further research into more detailed qualitative
and quantitative multilevel performance measurement.

References

Altman, E.I., 1993. Corporate Financial Distress and Bankruptcy: A Complete Guide to Predicting and Avoiding
Distress and Proting from Bankruptcy, second ed. Wiley, New York.
Appiah-Adu, K., Ranchhod, A., 1998. Market orientation and performance in the biotechnology industry: an
exploratory empirical analysis. Technology Analysis & Strategic Management 10 (2), 197210.
Argenti, J., 1976. Corporate Collapse: The Causes and Symptoms. McGraw-Hill, New York.
Bania, N., Bauer, P.W., Zlatoper, T.J., 1998. US air passenger service: a taxonomy of route networks, hub locations
and competition. Transportation Research E 34 (1), 5374.
Beuvais, G.A., 1992. Prepared statement in: Hearing on the Airline Competition Enhancement Act of 1992. US
Government Printing Oce, Washington, DC.
Buzzel, R.D., Gale, B.T., 1987. The PIMS Principles: Linking Strategy to Performance. Free Press, New York.
Carlson, J., 1987. Moments of Truth: New Strategies for Todays Customer-driven Economy. Ballinger, New York.
Churchill, G.A., 1979. A paradigm for developing better measures of marketing constructs. Journal of Marketing
Research 16, 6473.
Doherty, N., Horsted, J., 1996. Re-engineering peoplethe forgotten survivors. Business Re-engineering 3 (1), 3946.
Eccles, R., 1991. The performance manifesto. Harvard Business Review, JanuaryFebruary, pp. 131137.
Eilon, S., 1992. Management Practice and Mispractice. Routledge, London.
Feldman, J.M., 1997. Sticker shock. Air Transport World, April, pp. 6771.
Flint, P., 1999. Saving for the lean years: in the midst of prosperity, United readies for the next recession. Air Transport
World, 4648.
Gimeno, J., Woo, C.Y., 1999. Multimarket contact, economies of scope, and rm performance. Academy of
Management Journal 43 (3), 239259.
Gudmundsson, S.V., 1998. Flying too close to the sun: the success and failure of the low cost airlines. Ashgate,
Aldershot.
S.V. Gudmundsson / Transportation Research Part E 40 (2004) 443463 463

Gudmundsson, S.V., 1999. Airline failure and distress prediction: a comparison of quantitative and qualitative models.
Transportation Research E: Logistics and Transportation Review 35 (3), 155182.
Gudmundsson, S.V., 2002. Airline distress prediction using non-nancial indicators. Journal of Air Transportation 7
(2), 324.
Heap, J., 1992. Productivity Management. Cassell, London.
Hox, J.J., 1995. Applied Multilevel Analysis. TT-Publikaties, Amsterdam.
Lawrence, P.R., Lorsch, J.W., 1967. Organization and Environment. Harvard University Press.
Leavitt, H.J., Bass, B.M., 1964. Organizational psychology. Annual Review of Psychology 15, 371399.
Levine, M.E., 1987. Airline competition in deregulated markets: theory, rm strategy, and public policy. Yale Journal
on Regulation 4, 393494.
Liehr, M., Groler, A., Klein, M., 1998. Understanding business cycles in the airline market. Working paper,
Industrieseminar der Universitat Mannheim.
Miller, D., 1981. Towards a new contingency approach: the search for organizational, Gestalts. Journal of Management
Studies, 126.
Morrell, P., 1998. Air transport liberalization in Europe: the progress so far. Journal of Air Transportation World Wide
3 (1), 4260.
Narver, J., Slater, S., 1990. The eect of market orientation on business protability. Journal of Marketing 54, 2035.
Nunnally, J.C., 1978. Psychometric Theory. McGraw-Hill, New York.
Otley, D., 1999. Performance management: a framework for management control systems research. Management
Accounting Research 10, 363383.
Oum, T.H., Yu, C., 1998. Winning Airlines: Productivity and Cost Competitiveness of the Worlds Major Airlines.
Kluwer, Norwell.
Preisendorfer, P., Voss, T., 1990. Organizational mortality of small rms: the eects of entrepreneurial age and human
capital. Organization Studies 11 (1), 107129.
Salancik, G.R., Pfeer, J., 1977. Who gets power and how they hold on to it: a strategic contingency model of power.
Organizational Dynamics 5, 321.
Schefczyk, M., 1993. Operational performance of airlines: an extension of traditional measurement paradigms.
Strategic Management Journal 14, 301317.
Smith, C.A., 1955. Survey of the empirical evidence on economies of scale: in Universities National Bureau Committee
for Economics Research, Business Concentration and Price Policy. Princeton University Press, pp. 213230.
Spector, P.E., 1992. Summated rating scale construction: an introduction, Sage University paper series. Qualitative
application in the social sciences, series no. 082. Sage, Newbury Park, CA.
Spreitzer, G.M., 1995. Psychological empowerment in the workplace: dimensions, measurement, and validation.
Academy of Management Journal 38, 14421465.
Thomas, K.W., Velthouse, B.A., 1990. Cognitive elements of empowerment: an interpretive model of intrinsic task
motivation. Academy of Management Review 15, 666681.
Westerberg, M., Jagdip, S., Hackner, 1997. Does the CEO matter: an empirical study of small swedish rms operating
in turbulent environments? Scandinavian Journal of Management 13 (3), 251270.
Wood, S., 1979. A reappraisal of the contingency approach to organisation. Journal of Managment Studies, 334354.
Wouters, M., Kokke, K., Theeuwes, J., Donselaar, K.van., 1999. Identication of critical operational performance
measuresa research note on benchmarking study in the transportation and distribution sector. Management
Accounting Research 10, 439452.

Вам также может понравиться