J[kt"l C (il -� CANADIAN TRANSPORTATION RESEARCH FOR UM )l., ::- LE GROUPE DE RECHERCHES SUR LES TRANSPORTS AU CANADA
PROCEEDINGS OF
SEVENTEENI'H ANNUAL MEEI'ING
CANADIAN TRANSPORTATION RESEARCH FORUM
Volume 1
MONTREAL, WEB&:: MAY 26, 27 & 28, 1982 Compiled by: R. Lande & K. Tansey
.... VI-22 • VI-2 1 Introduction
Weather and population are frequently cited as reasons why U.S.-style airline deregulation is not a feasible policy for Canada. The argument is that Canada's more severe winter weather and its small and dispersed population result in unique oper- ating problems that require the continuation of the extensive regulation that has 1 characterized Canadian air transportation since 1938. At the same time, while these differences between Canada and the United States are relied upon to oppose the imple- WEATHER, POPULATION AND CANADIAN AIRLINE REGULATORY POLICY mentation of deregulation, no explanation is given as to why these differences did not serve to prevent Canada's adoption of airline regulation back in 1938 when regulation was also adopted by the United States.2 This raises the question of why is the argument asymmetrical? Why should weather and population constitute reasons for differentiating between Canadian and U.S. policies since 1978 while not providing a basis for doing so between 1938 and 1978 when regulation was the common policy of both countries? If weather and population have important differentiating effects on William A. Jordan airline per- Professor of Economics formance, it seems reasonable to expect that those concerned with airline operations Faculty of Administrative Studies and regulation would have studies that York University sponsored or been aware of measured the effects of those factors so that specific adjustments could have been made to reflect their impact. Yet, nowhere in their statements regarding weather and population have the opponents of deregulation referred to research findings that support their assertions. This apparent lack of formal evidence deserves to be rectified. In a recently completed research project jointly sponsored by Consumer and Corporate Affairs Canada and the Economic Council of Canada (in collaboration with Transport Canada), the question of the effects of weather and population differences was studied within the context of whether or not regulation has an appreciable effect on airline performance.3 While major performance differences were found to be asso- ciated with differences in regulation, both direct and indirect evidence failed to identify appreciable effects of either weather or population on operating expenses per RTM or on employee productivity. This paper will summarize these research findings. In the process, it will demonstrate that substantial similarities existed in airline operations in the two countries under common regulatory environments, thereby implying that airline performance in the two countries would also be similar under deregulation.
Methodology and Data Sources
The research project was based on comparisons of the performance of the Canadian CTRF Annual Meeting May 26-28, 1982 airlines operating large turbine Montreal, Quebec -powered aircraft with that of selected U.S. airlines operating the same aircraft types. All of the Canadian airlines were regulated by the
-2- VI-.23 VI-24
Transport Commission [CTC(A)], and they con- Air Transport Committee of the Canadian The data analyzed in the research project emphasized the four-year period from carriers), plus Eastern Provincial, sisted of Air Canada and CP Air (the mainline 1975 through 1978. Four years were analyzed in order to avoid anomalies occurring in Transair (the regional carriers). Seven of Nordair, Pacific Western, Quebecair and a single year. The comparison ends with 1978 because the major changes in U.S. air- federally-regulated, but by the Civil Aeronautics the U.S. airlines studied were also line regulation resulting from the passage of the Airline Deregulation Act (ADA) on Northwest and Trans World (trunk carriers Board (CAB). They consisted of Delta, October 24, 1978, make it inappropriate to use U.S. data for subsequent years during carriers) plus Allegheny, Frontier, North Central and similar to the Canadian mainline which the transition from regulation to deregulation has been in progress.5 Southern (local service carriers similar to Canadian regional carriers). Except for Delta and Southern, they were selected because their systemwide geographic operating Weather and Airline Performance areas were most similar to those of the Canadian carriers. Delta and Southern, in Weather affects productivity, and therefore costs, in all industries where signif- contrast, were selected because their operating areas were located largely in the icant proportions of total production must be undertaken outdoors. Examples include southern United States and both were headquartered in Atlanta, Georgia. Thus, if agriculture, construction and all transport modes. It happens that airlines have an adverse weather has a significant impact on operations, their performance should be advantage over surface modes in being less affected by snow, ice, rain and fog during that of the Canadian and the more northern U.S. airlines. superior to or enroute operations. Especially adverse weather can often be avoided by flying over four U.S. airlines included in the study were not regulated by a An additional to around it at relatively low additional cost, while surface carriers generally have federal commission. These were the four major intrastate carriers -- Air California, plough through adverse conditions. Airlines are, however, disadvantaged relative to Air Florida, PSA and Southwest -- that, prior to the end of 1978, were regulated only surface carriers in terms of weather in and around terminals. Airplanes can only stop by state commissions. Since these intrastate carriers served city pairs within (land) at airports, and this can be done only if local weather conditions are not too Texas that were also served by CAB-regulated airlines, they California, Florida and or severe. Also, an airplane cannot depart from an airport if it is closed by weather operated within regulatory duopolies comprised, on the one hand, of a group of air- if the destination and alternative airports are forecast to be closed at the estimated the other hand, by intrastate carriers regulated lines regulated by the CAB and, on are time of arrival. Clearly, the critical weather conditions of airline operations was in sharp contrast to the regulatory monopolies that by a state commission. This effects those experienced at airports rather than enroute. Thus, a comparison of the of the U.S. under the CAB, and in Canada under the existed throughout the remainder conditions of weather on airline performance can concentrate on the relative weather CTC(A). Not only did the regulatory duopolies allow the entry of new airlines into at the airports served by each carrier. Finally, it should be recognized that all the intense rivalry that developed among the air- the U.S. industry, but they allowed degree airlines experience adverse weather conditions. Therefore, the question is the lines to be expressed both through lower fares and through differentiations in service to which some airlines experience relatively more adverse weather than other airlines, is in contrast to the regulatory monopolies that virtually prohibited quality. This conditions. not whether some airlines enjoy good weather conditions while others have bad the entry of new airlines and rarely allowed general fares to deviate from commission- approved fare formulas. As a result, rivalry solely between federally-regulated air- Direct Evidence Regarding Weather lines was restricted mainly to improving service quality through scheduling more all) airports served by the airlines flights with newer aircraft and by providing higher levels of inflight service. Not Historical weather data for many (but not U.S. Department of Commerce.6 From surprisingly, these differences in regulatory environments resulted in the federally- have been published by Environment Canada and the weather factors were recorded for every regulated airlines providing levels of service that were generally superior to those these sources, the data pertaining to four in 1978 by each of the 18 Canadian and U.S. of the intrastate carriers, which, in turn, adopted fares that were as much as 50 available North American airport served resulting "sample" ranged from a 93 percent percent below the levels set by the CTC(A)/CAB fare formulas.4 Overall, since they airlines included in this study. The percent for Air California. The four weather allowed both entry and price competition, the regulatory environments within California, coverage for Air Canada down to 45 snowfall; percentage of times during regularly Florida and Texas approached the environment associated with deregulation and, there- factors studied were: mean annual (taken throughout the day) that the ceiling and/or fore, provide evidence on how airlines will perform after the full effects of deregu- scheduled weather observations Category I minimums of 200 foot ceiling and/or one-half mile lation have had time to work their way through the airline industry. visibility fell below the
-3- -4- VI-25 VI-26 Table 1 visibility, thereby preventing landings and takeoffs at most major airports; mean minimum temperatures during December through March; and the mean maximum temperatures Weather Factors,a System Operating Expenses per RTM and Passenger Trip Lengths Canadian during June through August. Mainline, Regional and Selected U.S. Carriers The simple averages for the four weather factors for each carrier are presented in Mean % of Obs. Table 1. These averages show that Canadian carriers face heavier snowfalls 1975-78 Average System within Carrier Annual Belay Mean Temp.( °F) Op. Exp./RTM Pax. North America than do the federally-regulated U.S. carriers, and that the U.S. intra- Snowfall Category I Dec-Mar Jun-Aug SCan. & U.S. Ol Trip (inches) Minimumsb Minimum Maximum state carriers experience very little snow. Associated with this is the expected Actual Trend Lengthd finding that Canadian carriers have lower average minimum winter temperatures than U.S. Mainline Air Canada 75.2 2.2 17.7 74.5 71.5¢ carriers (except for North Central), and they also experience lower summer temperatures. 68.60 1,064 CP Air 59.2 1.1 17.9 72.2 61.9 60.9 1,789 Again, the intrastate carriers are well off having the highest minimum winter temper- Trunk atures, but Air Florida and Southwest also have the hottest summer temperatures. All Trans World 22.4 0.8 29.9 84.5 68.6 64.1 1,390 this, of course, simply verifies the obvious -- winters are more severe and summers Northwest 31-7 1.0 25.6 80.4 54.8 67.8 1,110 Delta 18.5 1.1 32.7 are cooler in Canada than in the U.S.; and the Canadian carriers with transborder 85.3 74.3 81.4 636 routes do not serve enough U.S. airports to offset the effects of Canadian weather Intrastate Air Calif. 0.02 1.3 45.1 80.4 when calculating simple weather averages for North American operations. 67.5: 108.9 341 Air Florida 0.0 1.1 51.4 89.2 112.7 125.1 268 The clear dichotomy between the carriers of the two countries does not extend to PSA 0.01 1.6 44.8 81.2 77.2 113.2 318 Southwest 3.1 1.0 39.5 92.2 57.0 Category I ceiling/visibility minimums. Table 1 shows that below minimum conditions 120.8 284 exist for the Canadian carriers from 1.1 to 3.0 percent of the time, compared with Regional East. Prov. 136.1 3.0 10.0 69.3 110.0 0.8 to 1.6 percent of the time for the U.S. carriers. CF Air, Pacific Western and 95.1 444 Nordair 84.3 1.9 0.5 66.7 ,70.5 68.5 1,067 Quebecair, however, all fall within the U.S. range, while Air California, PSA and Pac. Western 54.8 1.2 5.6 69.2 83.1 102.4 383 Quebecair 129.8 1.5 2.4 69.8 76.7 75.8 Southern lie above the lower boundary of the Canadian range. Furthermore, it is obvious 771 Transair 69.2 2.1 -10.4 63.1 85.0 78.2 707 that the U.S. intrastate carriers are not favoured by this weather factor. Indeed, Local Service Air California and PSA have the highest percentages of the U.S. carriers. Allegheny 41.9 1.1 24.0 82.0 110.5 112.2 323 Given that the Canadian carriers generally face more adverse weather than U.S. Frontier 28.8 0.8 23.8 86.8 101.4 99.4 406 N. Central 47.3 1.1 16.7 80.5 131.9 129.5 carriers, the next question is whether or not this makes an appreciable difference in 253 Southern 13.3 1.2 34.4 86.8 114.9 111.1 329 their operating expenses per unit of output. Regressing actual operating expenses per aFor North American airports including Honolulu. RTM for the federally-regulated airlines (also from Table 1) against each of the four weather factors yields statistically insignificant R2 for the best-fit regressions bLess than 200 ft. ceiling and/or mile visibility. ranging from .006 (Category I minimums) to .065 (maximum temperatures).7 Furthermore, cCalculated from data for the federally-regulated airlines using the comparisons of individual carriers also indicate no relationship exists between equation: Y = 49.582 +20,227.564/X, R? = .866.
weather and operating expenses per RTM. For example, Table 1 shows that Allegheny and dTotal system scheduled plus charter RPM divided by total system passengers Eastern Provincial both had 1975-78 average operating expenses of about $1.10 per RTM yields distance in statute miles.
(in their respective currencies), yet, as can be seen from the Table, Eastern Provincial e Partially estimated. Scheduled passengm RTM for 1975-76 assumed to be operated with 3k times more snow, below Category I minimums 2.7 times more frequently, 98 percent of total RD( (based on 1977 experience). and average winter temperatures 14 degrees Fahrenheit colder than Allegheny. Only in Year f ended July 31, 1978. RTH estimated by averaging data for calendar summer temperatures did it have an advantage over Allegheny. years 1977 and 1978.
Sources: Jordan (a), supra note 3, Tables 5 and 33.
-5- -6- VI-27 VI-28
Category f minimums), with only the latter R2 being significant at the five percent Figure 1 shows that operating expenses per RTM decrease with distance for the 14 9 level. federally-regulated airlines. The best-fit trend line between these two factors yields Similar results were obtained when regressing the federally-regulated airlines' . a statistically significant R2 of .866, which means that 86.6 percent of the variations deviations of actual RTM per employee (a measure of employee productivity) from their in operating expenses per RTH are associated with distance.8 Therefore, the possible distance-related trend line values with the weather factors in Table 1. In these effects of weather on operating expenses per RTM may have been obscured by this regressions, the R2 ranged from .050 (snowfall) to .298 (below Category I minimums), important inverse relationship. To determine if this was the case, the deviations of with, again, only the latter R2 being significant at the five percent level." Clearly, it is desirable to remove the distance factor before investigating the possible relationship between weather and airline performance. Having done this, we below Category I minimums had Operating lispeneee per RTM in Relation to find that of the four weather factors analyzed, only 130- Total Svatee Trip Loniths. 1975-78 Average 114.e statistically significant relationships with operating expenses per RTM and employee 120 productivity. These findings would indicate that the much lower operating expenses 110- !'t per RTM of the U.S. intrastate carriers could be due in part to less adverse weather 100- were it not for the fact that these carriers did not operate under appreciably favour- able below Category I conditions. Indeed, Table 1 shows that these carriers exper- 0- ienced below Category I minimums from 1.0 to 1.6 percent of the time, which is quite $0•• similar to the experience of the trunk and local service carriers in the U.S., plus 70- Pacific Western and Quebecair in Canada. Therefore, with regards to all four E CF Air, direct evidence indicates that the substantially superior per- 60." weather factors, this formance of the intrastate carriers operating under regulatory duopolies was not due I 30- to favourable weather. Furthermore, the evidence implies that Canadian airlines in 3 40... general are not adversely affected by bad weather. Some of the Canadian carriers, may experience slightly higher ; Key such as Eastern Provincial, Air Canada and Transair, • • Canadian Mainline and Regional operating expenses per RTM due to operating where below Category I minimums are more 2 20- IC U.S. Trunk and Local Service A other Canadian carriers do not suffer from that weather factor any • U.S. Intrastate prevalent, but the more than the U.S. carriers.
200 400 600 800 1.000 1.200 1.400 1.600 1.800 Average System Trip Length (Statute Miles) Indirect Evidence -- Employment Source. Jordan (a). ..,r. not. 3. Figur. 4.
Indirect evidence can also be used to investigate whether or not Canada's more severe weather increases airline costs appreciably. For example, if Canadian carriers are more affected by adverse weather than U.S. carriers, one would expect this to be actual operating expenses per RTM from the distance-related trend line (Figure 1 and reflected in their employment practices. Specifically, Canadian carriers should have Table 1) were regressed against each of the four weather factors given in Table 1. If to hire more employees in order to produce a given amount of output, and most of these adverse weather appreciably affects operating expenses, it follows that federally- added employees should be assigned to jobs that are exposed to weather. Thus, in regulated airlines operating under more adverse weather conditions should have their relation to U.S. airlines, Canadian carriers should have lower employee productivity actual operating expenses per RTM above the trend values, while those enjoying less and should employ proportionally more pilots and copilots, other flight personnel adverse weather should have actual values below the trend values. The results of (cabin attendants) and terminal/ramp personnel, while having proportionally fewer these regressions were R2 ranging from .020 (maximum temperatures) to .299 (below employees working indoors in positions not affected by weather. -7- -8- VI-29 VI-30
already been shown that there was no systematic difference between Canadian It has the Canadian mainline and regional carriers should employ relatively more of these general due to weather, after adjusting for the distance factor. and U.S. airlines in types of employees than U.S. carriers since their work is directly affected by distribution of employees between indoor and outdoor jobs? Even But what about the weather conditions. It happens, however, that the opposite relationship actually available only for the six major categories listed in though Canadian employee data are proves to be the case for the Canadian mainline carriers. Table 2 shows Air Canada possible to investigate whether or not adverse weather results in Table 2, it is still and CP Air had pilot and copilot shares of 7.2 and 7.6 percent which are less than the proportionally more employees in outdoor jobs. One of these Canadian carriers utilizing 10.0 to 13.0 percent shares for the U.S. trunk and intrastate carriers.11 Similarly, covers pilots and copilots, and a second covers other flight personnel. six categories for the other flight personnel category, the Canadian mainline carriers' shares were explanation of higher Canadian airline operating expenses, Under the adverse weather 13.4 and 11.9 percent compared with 14.5 to 18.4 percent for the U.S. trunk and intrastate carriers. Table 2 The differences in pilot and copilot shares proved to be negligible for the Canadian regional and the U.S. local service carriers. The five regional carriers' shares Distribution of Number of Employees by Category Percentage ranged from 10.5 to 15.0 percent, compared with 11.8 to 13.7 percent for the four local Canadian Mainline, Regional and Selected U.S. Carriers 1975-78 Average Values service carriers (and 10.5 to 13.0 percent for the intrastate carriers). Even these similarities, however, are inconsistent with the reasoning that weather adversely• Percentage of Total System Employees (1975-78 Average) affects employee productivity more in Canada than in the U.S. The percentages for other Aircraft flight personnel for the five regional carriers provide the only case that is partially Carrier Pilots & Other Flt. Maint. & Traffic General Other Copilots Personnel Labour Servicing Mgt. Employees consistent with the weather explanation of lower Canadian labour productivity. These percentages ranged from 11.9 to 15.5 for the regional carriers in comparison with only Mainline Air Canada 7.2% 13.4% 15.2% 36.4% 0.6% 27.2% 10.4 to 11.4 percent for the local service carriers. However, the intrastate carriers' Air 7.6 11.9 15.9 26.2 1.1 37.3 CP range was from 16.5 to 17.0 percent, which is not consistent with the weather expla- Trunk nation. One possible reason for the regional carriers' larger percentages over the Trans World 10.0 14.5 17.4 33.8 0.1 24.2 local service carriers is their large charter operations. The cabin attendant require- Northwest 12.6 18.4 10.9 40.9 0.4 16.8 Delta 10.3 15.0 12.1 51.1 0.2 11.2 ment for long-haul charter services may be greater than the requirements for short-haul local service operations.12 Intrastate Air Calif. n.a. n.a. n.a. n.a. n.a. n.a. The obverse comparison, utilizing the general management and other employees cate- Air Florida 11.0 16.5 10.0 41.5 9.0 12.0 gories that generally work indoors, is consistent with the conclusion derived from the PSA 13.0 17.0 15.0 39.9 3.0 12.1 Southwest 10.5 16.6 9.4 51.3 4.5 7.7 comparisons of flight personnel. Since there appears to be some difference of opinion among the Canadian carriers on how to allocate personnel among the general management Regional East. Prov. 10.5 15.5 29.8 29.5 3.3 11.4 and other employees categories, it seems desirable to aggregate these two categories Nordair 14.7 14.8 19.6 25.3 2.9 22.7 and analyze the resulting combined percentages.13 Instead of Air Canada and CP Air Pac. Western 12.0 11.9 21.2 36.8 17.3 0.8 Quebecair 15.0 16.4 16.7 27.9 17.2 6.8 having relatively small percentages for these combined categories, their 27.8 and 38.4 Transair 13.1 13.2 22.2 20.1 14.2 17.2 percent shares were both larger than the 11.4 to 24.3 percent shares for the trunk Local Service carriers and the 12.2 to 21.0 percent shares for the intrastate carriers. Similarly, Allegheny 11.8 11.2 14.3 46.9 0.5 15.3 the five regional carriers' combined shares ranged from 14.7 to 31.4 percent, with Frontier 13.7 10.4 17.1 39.8 1.1 17.9 N. Central 13.6 11.4 14.4 43.5 0.4 16.7 three of these carriers exceeding the local .service carriers' range of 15.8 to 23.0 9.3 44.4 0.9 22.1 Southern 12.6 10.7 percent. Source: Jordan (a), supra note 3, Table 34. Overall, the indirect evidence from the employment data does not support the expla- nation that adverse weather is an -9- important reason for the lower output per employee of -10- VI-31 VI-32
the majority of Canadian carriers. This, of course, is consistent with the conclusion availability of alternative means of transportation (the highway system, railroads, of the previous section that was based on direct evidence regarding the effects of and water transport); an area's isolation relative to other population centres, or its weather. proximity to another city having superior/inferior airline service; and the economic characteristics of the area (institutional, marketing, balanced or industrial Indirect Evidence -- Profits economies) 17
If relatively adverse weather serves to increase costs appreciably and if airlines charge the same fares per mile (as is the case for the federally-regulated airlines), Total Population one would expect carriers operating mainly in the north to have lower profits than those The simplest approach to investigating whether or not differences in operating costs operating primarily in the south. This would be especially true for the CAB-regulated between Canadian and U.S. carriers are affected by differences in population is to com- airlines since they calculated their fares from the same distance-related fare formulas pare these carriers in terms of the aggregate 1970/71 populations surrounding the air- throughout this period.14 ports served by each carrier, that is, the population pool from which most traffic It happens that Delta (a predominantly southern airlines) did indeed have relatively originates.18 These population totals for each carrier are given in Table 3, broken the total low operating ratios (high profits), averaging 91.9 during 1975-78; but so did North- down between areas located in Canada and those in the U.S. It can be seen that west, at 91.0, despite the fact its primary routes extended along the northern-most edge populations of Air Canada and CP Air were heavily influenced by the populations of the of the U.S. and on to Alaska and across the North Pacific.15 At the same time, even metropolitan areas they served in the U.S. Even with the U.S. populations added to though it mainly served the U.S. south, Southern was a relatively low-profit airline Canada, however, the 1970/71 population pool available to Air Canada (51,634,000) was (average operating ratios of 96.5) compared with Allegheny, Frontier and North Central just two-thirds of Trans World's pool (76,097,000), while CP Air's population pool (average ratios of 91.6 to 95.7 during 1975-78) whose routes were mostly in the northern (19,539,000) was only 30 percent of Northwest's (66,083,000). Similarly, the population U.S. with extensions on into Canada.16 Clearly, these profit performances are also pools of the Canadian regional carriers were all very much smaller than those of the inconsistent with the hypothesis that northern climates serve to increase the operating U.S. local service carriers. expenses of airlines over those operating in more temperate climes. If the Canadian carriers had consistently higher operating expenses than comparable U.S. carriers (after adjusting for the effects of distance), these relative population data would support the assertion that small population pools are associated with high operating costs. However, Figure 1 and Table 1 show that this was not the case. There Both the direct and the indirect evidence presented in this section challenge the was no consistent pattern among the federally-regulated airlines along national lines. statement that weather is an important reason for Canadian carriers to have higher Population data for the U.S. intrastate carriers provide an even more serious operating costs than U.S. carriers in general and the intrastate carriers in particular. challenge to the assertion that small population pools are associated with high operating Actually, similarities, rather than differences, characterized the cost performances of costs. The intrastate carriers' 1970 population pools ranged from 4,905,000 (kir the federally-regulated airlines of the two countries during 1975-78. Clearly, one Florida) to 15,523,000 (PSA). These totals were considerably smaller than the totals must look for reasons other than weather to explain why the performance of the federally- for Air Canada and CP Air, and they encompassed the mainline carriers' domestic (Canada regulated Canadian (and U.S.) carriers differed so markedly from that of the low-cost only) population pools. Yet, as shown in Table 1, the intrastate carriers had much U.S. intrastate carriers. lower operating expenses per RTM than any of the other carriers in relation to the line. If total population were a major factor affecting airline Population and Airline Performance distance-related trend operating costs, the intrastate carriers would not have been so superior in this measure. Population is a proxy for traffic demand. Airlines serving areas of low population There were similar anomalies within each carrier group. CP Air's population pool are generally expected to have lower demand for their services than carriers serving was only 38 percent as large as Air Canada's, but relative to the distance-related trend heavily populated areas. Of course, population is just one factor affecting overall lines its operating expenses per RTM (Table 1) were slightly superior to Air Canada's. demand for airline services. Others include price; per capita income; the -12- -11-
C
,
,0 ,0
Z Z
•••••• ••••••
44
v4 v4
44
0
o o
05
'44/3
▪ ▪
00
C.)
000
0
0
4403
0
CO CO
P
#441)
003
W0
-
CI CI
co co
SC
-4 -4
8
0
r4
.11
E.
C.)
44
4-4
0
F4
E.
E-
Cr)
0
03
44
0
4.)
03
cci
co
•
•
4
r4 r4
v4 v4
..: ..:
0 0
C)-. C)-.
w w
.
---SI ---SI
. .
..., ...,
•0 •0
.4 .4
--.. --..
(-4 (-4 .-I .-I 00 00
c) c)
0. 0.
-,..... -,.....
In In
C) C)
.7 .7
.4.-4 .4.-4
rA rA
00' 00'
,T ,T
. .
N.0 N.0 ,-4C' ,-4C'
-aml -aml CV CV
0-1 0-1
....? ....?
r4 r4
CI CI
r4 r4
CV CV
4-4 4-4 C' C'
C..) C..) CV CV
In In
r4 r4
Cr) Cr)
en en 0 0
Ce1 Ce1
r4 r4
0 0 v
CV CV
0 0
0 0
r4 r4
.
...4 ...4
CO CO
, ,
------
-...
co co
el el
CV CV r1 r1
t t
en en
cv cv
.../. .../.
in in
d d
CV CV
o o
VD VD
0, 0,
. .
4-4 4-4
v4 v4 Co Co
.--1 .--1
4-1 4-1 0-4 0-4
Cn Cn
VD VD
N. N.
en en
a': a':
C4 C4
r-4 r-4
CD CD
Ch Ch
Cn Cn
W W
......
E
0044r-4 0044r-4
WHZCZ WHZCZ
4
,
'v. 'v.
,
.1 .1
,I
.-4 .-4
---. ---.
00,0 00,0
el el
•ri •ri
r.4 r.4
...1
r-4 r-4
on on
03 03
r-4 r-4 r
CD CD
DON- DON-
N. N.
CI CI
I
k4'.; k4'.; ul ul
C) C)
r. r. Ch Ch
Ch Ch
0 0
N. N.
N. N.
,..
-I -I
o o
44 44
CO CO
1 1
.
r) r)
40 40
.
.0 .0
......
.0 .0
-...... ,
../. ../.
,
.4 .4
------
el el
-...... -......
'.0 '.0
+.1
1.-- 1.--
CD CD
,0 ,0
Cil Cil
cv cv
co co
CV CV
03 03
...1
N-,-4 N-,-4 CD CD
r-4 r-4
CD CD
C) C) In In
Ce1 Ce1
VI VI
v) v)
CI CI CD CD
4.
N. N.
VI VI
4-
V, V,
0 0
§. §.
co co
,
.
.
4
.0 .0
••••• •••••
03 03
03 03
---. ---. .1
I, I,
Cp Cp
-4. -4.
...1
en en
In In
0' 0'
N. N.
CV CV
r-.4 r-.4 00 00
../. ../.
03 03 r4 r4
CV CV 01 01
N. N.
C4 C4
4) 4)
C Ni Ni
N. N. a
en en
r4 r4
CV CV
(1) (1)
N. N.
03 03
.
,
.
.,'
I•4 I•4
4.15,-1 4.15,-1
4J 4J
44 44
0.-cl.cn 0.-cl.cn
444414•00 444414•00
cl cl 4) 4)
cle--4 cle--4
CO CO
......
0 0
0000 0000
...4 ...4
--...... --......
r4 r4
'.00' '.00'
r.-4 r.-4
r r
r4 r4
44 44
0 r-4 r-4
-1-4 -1-4
Cs4 Cs4
v-4 v-4
N. N.
C-) C-)
N-'0 N-'0
CV CV
en en
Ch Ch -13- -13-
C0 C0
CD CD
en en
VD VD
4 4
,
V V
• •
.-4 .-4
••••. ••••.
rA rA
r4 r4
•.0 •.0
r4 r4
r-4 r-4
r-4 r-4
CV CV
CI CI
it, it,
144 144
v4 v4
vo vo
In In
v-4 v-4
In In
-.../ -.../
C) C)
CM CM
v.l
CD CD
0 0
4 4
44 44
I I
co co
.
......
,-4 ,-4
--...... --......
..3
en en
03 03
en en
.1 .1
CI CI
r4 r4 4-4 4-4
00 00
111 111
Ch Ch
Cr) Cr)
04 04
CV CV
en en
4/1 4/1 el el
r-4 r-4
r-4 r-4
VI VI
VI VI
1 1
.
••••• •••••
.0 .0
.4 .4
el el
r4 r4
cm cm
-4 -4
CI CI
r4
CV CV
In In
CT CT
03 03
03 03
P. P.
eV eV
r
,
eV eV
VD VD
NI) NI)
0 0
&a &a
4.
> >
CO
W W
4 4
.-I .-I
v-4544 v-4544
C4 C4
4000 4000
o o
CO CO
C C
wWZ wWZ
-... -...
-4 -4
op op
-4- -4-
en en
-.....----...,...,. -.....----...,...,.
r4 r4
03 03
r4 r4
N. N.
CO CO
en en
03 03
03 03
01 01
v-4 v-4
r4 r4 cn cn
C..W C..W
en en
r-4 r-4
C•1 C•1 CD CD
4
,
> >
o o
I I
I I
W W
• •
• •
" "
,
--.. --..
...... '0'O '0'O
------
01 01
.1
tnc0,1. tnc0,1.
r4 r4
CV CV
,T ,T
,-.1 ,-.1
N. N.
r4 r4
r-4 r-4
In In
VD VD
,T ,T
,? ,?
r-4 r-4
,I
r4 r4
...7 ...7
ev ev
CD CD
C) C)
v. v. V) V)
N. N.
r-4 r-4 rA rA
0 0
c3 c3
.
.
,
-.. -..
,
......
---. ---.
01 01
%.1
Ch Ch
N. N.
r4 r4
.4 .4
r4 r4
Ch Ch
rA rA
1 cn cn
VD VD
r1 r1
-.1
C) C)
..,
VD VD
In In
CV CV
D D
r4 r4 CV CV
VD VD
C•4 C•4
CV
5 5
.
44 44
.,))C .,))C
T T
.
---.. ---..
en en
C4 C4
oD oD
r4 r4
03 03
0-1 0-1
r-4 r-4
CO CO
03 03 VD VD
'.4 '.4 r4 r4
CD CD
VD VD
Cel Cel
c7 c7
en en
v0 v0
CO CO
CD CD
CD CD
0 0
I I
I I
4)03 4)03
i. i.
a) a)
U0 U0
------
0 0
C C
el el
r4 r4
CV CV
r4 r4
C) C)
.1
-I -I
Co Co
C•4 C•4
r4 r4
CI CI
,t ,t
r4 r4
N. N.
C
r4 r4
en en
CD CD
N. N.
E
v-4 v-4
'4 '4 I I
-
9
co co
W
-. -.
.
1 1
•-4 •-4
v4 v4
h
V) V)
0,40 0,40
> >
0)4)0)4444 0)4)0)4444
u u
W0W140 W0W140
w w
i
0 0
- -
--...... --,-- --...... --,--
-...... -......
Ch Ch 00 00
----...... ----...... In In
r4 r4 a) a)
CT CT
cs, cs,
VD VD
In In
v) v)
,r ,r
0 0
,') ,') Le) Le)
el el CD CD
C*1 C*1
N--4 N--4
.= .=
.1
-...... -......
cn cn
In In
N. N.
eV eV
r4 r4
In In
f`•• f`••
N.CV N.CV
cn cn
r-4 r-4
VI VI
......
>. >.
00 00
4) 4)
.
03 03
03 03
0-1 0-1
en en cv cv
In In
eV eV a) a)
CV CV
CO CO
Co Co
en en
r... r...
CV CV
0 0 vD vD
•-•4 •-•4
...1 ...1
eV eV c) c)
eV eV
In In
eV eV
In In
In In
VD VD
r.cn r.cn
crl crl
CNI CNI
v-400 v-400
0 0
Li Li
44 44
0 0
-... -...
co co
•••1• •••1•
In In
en en
CT, CT,
r
c
, 4-4 4-4 N. N. .
In In
N- N-
CV CV
eV eV
N. N.
rl rl Crl Crl
In'.4 In'.4
'.0 '.0
CV CV rq rq
07 07
...7 ...7
00 00 cn cn
.-4 .-4 In In
a) a)
4.41 4.41
CV CV
el el
N. N.
4 4
C.) C.)
o o
41 41
In In ...1
CM CM
cv cv
CO CO
In In
0 0
cm cm
CM CM
N. N.
ul ul
CV CV
VD VD
r4 r4
.0 .0
r. r.
rA rA
In In
In In
r4 r4
r4 r4
N. N.
CD CD
CD CD
I I
4.4 4.4
1 1
.
o o
V V
'OS
.0 .0
.0 .0
.-1 .-1
,A ,A
..-4
r4 r4
.0 .0
CO CO
r-4 r-4 ✓ ...4 ...4
0 0
4 0 0
✓
0 0
4
0 0
0 0
.0 .0
,
.44 .44
P P 4
$4c. $4c.
0. 0.
4.1 4.1
14 0 0
.003 .003
0 0
,
>W >W
0 0
•r4 •r4 o o
4, 4, 5,-4 5,-4
0 0
0) 0)
44 44 $.4 $.4
,
0 0
co co
44 44
u u
w w 41 41
JZ JZ
-4 -4
O O W W
44 44 O
VI VI
50.,-4 v4 v4
E-4 E-4
$.. $..
0 0
4) 4)
4 4
5 5
0 0
0 0
4 4 co co
O
CC
i. i. CI CI
14 14
C.) C.)
w w
wm wm
.
.0 .0
4 4
'CI 'CI
44 44
•r4 •r4
a a
0 0 4) 4)
II II
0 0
•-4
0 0
0 0
.c
0
4.3
s s
0.4
3 3
44 44
.= .=
P P 1
0
i-4 i-4
41 41
0.0)
o 0) 0) 44 44
0 0
514 01 01
01
CU CU 4) 4)
CO CO
vz, vz,
0 0
5 5
v4 v4
0 0
v4
v4 v4
4)0)
.
4.) 4.)
o.4.,
0 0
030
• •
0 0
0 0
0.44 4.1 4.1
S'0
sa)
404)
4) 4)
o o
0) 0)
$4 $4
4) 4)
C.) C.)
M M
C) C)
.
-33
4 4
.0 .0
..0
44 44
44
4-4 4-4
4-4 4-4
A
4J 4J
0 0
.0 .0 4.1
.0
0
P
0
z
C
0J
o
4) 4)
0 0
W
.-4 o o
co
W W CO
g
I W
>, >,
44
CO
w
0 0
v4
4-4
g g
co co
03 03
03 • • :30)
0 0
0 0
4J 4J
44
›N ›N
> 0) 0)
4) 4)
>-
-e
03 03
i., i.,
CCI w w
co
CU
•
,
o
r4 r4
1.
,
•
dependent dependent
relationship relationship
sions sions
airlines:
carriers, carriers,
also also
percent percent
"airport "airport
Source: Source:
countries. countries.
Northwest Northwest
Population Population
as as The The
significant.
operating operating
these these
a a
total total
operating operating
pool pool carrier carrier
and and
total total
much much
given given
few few
a a
R
if if
1. 1.
Finally, Finally,
2. 2.
had had
3. 3.
2
significant significant
availabe availabe
It It
larger larger
bThe bThe
Again, Again,
for for
airports. airports.
airports airports
number number
in in
population population
for for
operating operating
a. a.
as as
Deviations Deviations
Total Total
b. b.
trend trend
a. a.
Deviations Deviations
line line b. b.
a. a.
b. b.
can can
by by
the the
Jordan Jordan
carriers." carriers."
Table Table
variables) variables)
the the
Southwest, Southwest,
had had
costs. costs.
first first
costs costs
Domestic Domestic
large large
North North
Domestic Domestic
North North
Domestic Domestic
these these
North North
per per
the the
20 20
Data Data
be be
population population
Air Air
between between
lowest lowest
and
of of
operating operating following following
line line
regression regression
the the
to to
argued argued
Airport Airport
might might
4, 4,
total total
(a), (a),
Thus, Thus,
airline airline
should should
Canada Canada
number number
American American as as
American American
American American
six six
a a factor factor
are are
costs costs
Every Every
per per
lowest lowest
One One
and and
and
in in
in in
carrier, carrier,
PSA's PSA's
population population
population population
population population
population population
and and
measures measures
had had
regressions regressions
supra
have have
for for
Totals Totals
number number
North North
total total
total total
all all
carrier carrier that that
were were
and and carrier carrier
were were
be be
pool pool
of of
six six
expenses expenses
influencing influencing
airports. airports.
analysis analysis
population population
the the
operating operating
1978.
population population
population population
pool.
of of
population population
the the
similar. similar.
each each
note note
population population
calculated calculated
American American
CP CP
operating operating
pairs pairs
RIM RIM
is is
of of
closely closely
through through
sum sum
second second
the the
of of
same same
having having
Air Air
pool
pooll
pool
the the pool. pool.
airports airports
is is
pair pair
per per
3, 3,
all all
operating operating
per per
ranged ranged
above above
of of
provides provides
were were
counted counted
19
19
relevant relevant
expenses expenses
population population
Table Table
The The
9
employee employee
pool pool
smallest smallest
airline airline
Population Population
airport, airport,
related related
proportionally proportionally
carriers carriers
is is
data data
RIM RIM
pool
pool
pool
a a
expenses expenses
Furthermore, Furthermore,
per per
by by
evidence evidence
-14-
relatively relatively
appreciably appreciably
second second
the the
from from
located located
data data
and
dividing dividing
36.
were were
airport, airport,
once once
expenses expenses
useful useful
number number
operating operating
of of
population population
to to
.000 .000
from from
population population
with with
operating operating
(as (as
per per
is is
calculated calculated
per per
the the
per per
population population
at at
questions questions
in in
the the
independent independent
evidence evidence
the the
to to small small
airport
each each
the the
each each
of of
RTM RTM
rather rather
airport airport
more more the the
the the
three three
below below
and and
.169 .169
number number
Canadian Canadian
distance
costs.
1970/71 1970/71
from from
factor factor
most most
Canadian Canadian
relevant relevant
at at
airport airport
employee employee
airports. airports.
pool pool
population population
Trans Trans
U.S. U.S.
for for
as as
and and
than than
the the
per per
the the
regarding regarding
data data
the the
efficient efficient
of of
-- --
another another
-related -related
variables). variables).
the the
affecting affecting
none none airport, airport,
importance importance
population population
trunk trunk
airports airports
carriers carriers
the the
World World
served, served,
U.S. U.S.
distance
for for
areas areas
carrier carrier
one one
productivity productivity
federally
Were Were
were were
total total
pool pool
each each
that that
carriers, carriers,
carriers carriers
carrier carrier
the the
and and
during during
of of
trend
then then
in in
serving serving
yielding
-related
served served
airline
statistically
now now
this this
lack lack
carrier carrier
of of
was was population
pools pools Northwest Northwest
the the
both
-regulated
The The
both both
being being
population
V V
serving serving
1978.
the the
only only
intrastate
serving
I I
(as
of
yet yet
- -
regres regres
for for
through
the
3 3
case,
carriers'
are
4
44
about
it
each
in
- a VI-35 VI-36
55 percent of its comparable U.S. carriers. Also, with one exception, the Canadian regional carriers had lower populations per airport than the U.S. local service
u 4.4 00 al ,I r4 CN al CM 00 oh r4 oO 00 CD CD C) CM carriers. Given these similar relationships, it follows that the inconsistencies 4.4 La • • • • •O 0 O 0 OD r, .1 VD .1 v0 CI CV CV rA CV CV V1 C4 el, 17 0.1.4 co between total population pool and operating expenses per RTM should also hold for popu- 1.4 14 i4 14 •e4 , < 0 > lation per airport among the federally-regulated carriers. At the same time, the out- s., $.4 w • w a, • Z 0, c4 r, cl Cm U101 OD CM CM VD CI Ch standing performance of the U.S. intrastate carriers continues to challenge the assertion 0.1/) • • • , 99 , • • . • 0 0 co, OD es .1 40 VI ,0 el ,r 0 '7 NI rn ...7 to 0,0 .-4 i.. that population per airport affects operating costs. The total populations per airport CO w 140 -,-. ,4 CC I... of the intrastate carriers were roughly equal to, or less than, those of the Canadian ,ii-, m 0 .-4 el VD 00 00 I's i4 < /4 C:1 I • • I I • • • • • • • . 1 C CV cl M 0CC ..3. .1 CV C,I v-4 ,C1 C4 VD V:, ul U mainline carriers and, in two cases, were similar to the mainline carriers' domestic 0 u) populations per Canadian airport. Yet the intrastate carriers had much lower operating w m C) CD CD CD CD CD CD CD CD CD 0 CD CD CD CD CD CD CD .-I expenses per RTM after adjusting for distance. CD CD 000 0000 CD CD CD CD CD 0000 CC CD CD 000 0000 CD CD CD CD CD 0000 0 .Z Finally, the best-fit regressions relating the three dependent variables with each 0 -.70 en r. r4 cl .1 ..7. el r4 CD CD C4 el 0 el r4 C4 4J E. r4 Co r4 '.400 el c. ,. CI CD .0 ,D CD 00 in CD c4 r, CV rA Cg eg r4 e4 r4 r4 r-4 v-4 c4 c4 r4 C4 v-4 •-4 carrier's total North American and domestic populations per airport (as two separate v-4 CC 00 000 0000 CD CD 0000 independent variables), yield R2 ranging from .035 to .128.21 Again, none of the R2 CD CD 000 0000 CD CD 0000 44 CD CD 000 0000 CD C) 0000 o are significant at the five percent level. Thus, the evidence continues to question '--c03 Nt el 0 44 c,70 0 ,r rn cv C•1 ,C, 4,1 r-4 C,1 CO el ,4 -. e, el ••••T .7 -.7 0 r-4 .1... g CV CV CV CV eq 4-4 '-4'-4 F--4 04 v-4 CV v-4 m the importance of population as a factor influencing operating costs. 0 .0 CD Cp C) CD CD CD CD CD CD 000 aJ CD CD CD CD CD CD CD CD CD 000 CD CD CD CD CD C) CD CD CD CD CD CD Population per Airport Carrier 4>, 03 '7 CD CD en 0000 r4 in ,D c4 en uP c4 e47CD NJ .1. op F. CD c4 ,T CD 00 Co Co Co .13 ,-• r-4 c, t-4 v-4 •-4 C,I v-4 0 A third possibility is that an inverse relationship between population and airline '0 ,4 > operating costs can be found by using average population per carrier for all large -,-4 CCI CD CD 000 0000 C) CD CD CD C) 0000 13 • CD CD CD CD CD CD CD CD CD CD CD CD CD CD CD CD CD CD 00 carriers serving all airports within the relevant population pool. For example, Table W4-4 CD CD 000 0000 CD C) CD CD CD 0000 I.,, r O W co cm en CD .0 ,4- op rl ,40 .4. CV 01 ,7 CN ,I. 1.4 eA .1 CO CD v.4 v-.4 3 shows that Air Canada's North American system covered approximately 11,524,000 people W 0 m ul .1 Fl CV p. -70' CV ,4 c4 CD CD ,4 - ,I. ,r CD in W -'.45-4 CD 00 00 in c4 OD ,T ,4 .1 c4 ,?, 1 4 c4 4-4 r4 e4 r, 00 14 00 —7 M 0 located in areas served by 33 Canadian airports, and 40,110,000 people served by 17 '4 4 0 U.S. airports, for a total population of 51,634,000. Including Air Canada, 90 large CD CD CD CD CD CD CD CD CD CD CD CD CD CD CD u ip CD CD 000 0000 CD CD 0000 0 0 CD CD 000 CD CD CD CD CD CD 0000 0 cn carriers (69 Canadian and 21 U.S.) provided service in 1978 at the 33 Canadian airports, 0 • CC .44 V) Co C, V0 ,10 NJ C..),0 ...7 eV 0 v-4 01 00 s7 CV C,1 0 14 0 in C•1 in e.1 C,1 ,r .7 01 eV 0 e4 v-4 e, 0,.1 o 0 /4 and 151 large carriers (12 Canadian and 139 U.S.) served the 17 U.S. airports, making a m co co of) r-4 U) ..1. .-4 in C,1 r-4 .. ••.7 .7 CV VZ. 00 44 CO C,1 •-4 v-i .-4 •-4 v-4 0-4 r-I a.. total of 241 airport carriers serving all areas on Air Canada's North American system. ""4 O 0 0 O 0 0,4 o o o o o o o o o o o o co 0. > In essence, population per airport carrier indicates the average distribution of popu- c) CD CD CD CD CD CD CD CD 000 4J 0 4-4 41 O o 00 o 0000 000 1.. C r-4 N- O 0 0 lation over a carrier's system adjusted by the degree of competition/rivalry the carrier - C'. ,0 00'.4 0000 r, C) el ..7 r4 ,r CP Ch 0 as .94 $4 00 -70 ,7 %ID 4-4 01, I 0 .1-. ,r 4,1 co 1.4 a., F- cC C400 CI C'l ,4 1 CA C,1 ‘, 1 e.• .1 0 •*4 CO CD 0 faces from other airlines. It is a rough indicator of the traffic available to a o'0 • 0. 0 ,4 = ....7 r4 e4 = ,J 4.4 O. carrier at each airport on its system. 0 0 o 00 .0 • •,:,0 • W CJ $.4 .0 Table 4 also gives both the population per airport carrier and the average number o r-4 > (11 ,4 v-I 4/ v-I 4J 17 /.. ,4 0 .44 •r-4 IJ 0 ,.. 14 > ›N 0 .0 CI ,-4 CI 0 0 4.1 ,4345 W 01,4 14 W0WW0 a ,.., of carriers serving each airport located within the geographic areas associated with the cu 0 5C) 0 ri o 0 r-4 CLoWW0,4 WWW.WW 0 0 0 W U 4.00,1 U 0 ,•4 300 Cn.C,4 00 Z E-• ,/ ,,... ,4 0 ,4 (0 ..0 0 CO 0 44 .0 0 • 0 4/ LO 00 4, 4).0 CV .0 14 CU indicated carrier's population pool. It can be seen that in 1978 the average numbers v-4 .4 .st 0 4.) a.i ea 4J 0 ,4 121 • .0 0 r-4 CU 0 C.) 4.a 0 CU O ii 0 R114,1 W/41447 +4 5 /4 000 0.-10 V 0. 14 -el ,4 CI. Z W 0 a/ ,4 +4 •r4 co0 00 0 0 CO 7 14 uv-4 14 • 0 W Z of carriers were generally two to three times higher at U.S. airports than at Canadian. ..,CU, 14E...Z0 0<<0..1 a) g4.1 Z a. 0* E. 0 < 4. Z cn "4 0 E. 0.4 04 ....1 0 m Therefore, it is not surprising to find the total 1970/71 populations per airport carrier were almost the same among the Canadian mainline and U.S. trunk carriers,
-15- -16- VI-37 VI-38 ranging from populations of 190,000 to 217,000 per airport carrier, and with Air Canada if adverse weather serves to increase airline costs it should be evident in the distri- and Trans World being almost identical. The U.S. intrastate carriers were all lower bution of employees between outdoor and indoor jobs. Yet, no indication of such effects than the mainline and trunk carriers, ranging from populations of 75,000 per airport could be found. To the contrary, the distribution of employees between outdoor and carrier (Air Florida) to 174,000 (PSA). The Canadian regional carriers tended to be indoor jobs for Air Canada and CP Air relative to the U.S. trunk carriers was just somewhat lower than the intrastate carriers in this measure, with populations per air- opposite to the expected distribution, and the distribution for the regional carriers port carrier ranging from 60,000 (Pacific Western) to 160,000 (Nordair). Finally, the did not differ appreciably from that of their local service counterparts. Third, no U.S. local service carriers were divided. into two subgroups. Allegheny (250,000) and statistically significant relationships were found between three measures of population North Central (211,000) were similar to the mainline and trunk carriers, while Frontier and the operating expenses and employee productivity measures. Furthermore, inconsis- (102,000) and Southern (172,000) were more like the U.S. intrastate and Canadian tencies between population and operating costs existed within each of the carrier groups, regional carriers. including the Canadian carriers. Turning again to statistical inference, the best-fit regressions relating the three In contrast to these essentially negative findings, again and again the superior dependent variables with each carrier's total North American and domestic populations , performance of the U.S. intrastate carriers operating/ within regulatory duopolies have per airport carrier, yield statistically insignificant R2 ranging from .004 to .107.22 posed a fundamental challenge to the assertion regarding weather and population. Their So once more the evidence fails to support the argument that there is an important percentages of below Category I minimums were among the highest in the U.S. and their inverse relationship between population and airline operating costs. systemwide population characteristics were not favourable. For example, the total population pools available to Air California and PSA were much smaller than those avail- Sunziar able to Air Canada and CP Air, and were roughly comparable to those Canadian carriers' domestic Whether based on formal statistical inference or on individual comparisons between pools. Yet they radically outperformed the mainline carriers in terms of carriers, the evidence fails to support the assertion that, in general, there is an distance-adjusted operating expenses per RTM., Similarly, Southwest's total population important inverse relationship between population and airline operating costs. Further- pool in Texas was smaller than even CP Air's domestic pool, its population per airport more, there is no consistent indication that the operating costs of Canadian airlines was about the same, and its population per airport carrier was again smaller. The differ from those of comparable U.S. carriers due to population differences. To the metropolitan areas it served were even comparable in population to those in Canada.23 Yet, even contrary, the evidence strongly indicates that population is not an important deter- so, Southwest achieved much lower operating expenses per RTM than CP Air (as well as the U.S. trunk minant of airline operating costs. carriers). The large differences between the performance of the U.S. intrastate carriers and the essentially similar performances of the federally- Conclusions regulated Canadian and U.S. airlines implies that regulatory differences have major effects on airline performance in contrast to small, or nonexistent, effects of weather This paper has investigated the arguments proposed by spokesmen for Canadian air- and population. lines that adverse weather and smaller population are two reasons why Canadian mainline Overall, the evidence presented in this paper casts doubt on the arguments that and regional carriers have higher operating costs than their U.S. counterparts. First bad weather and small population impose higher operating costs, on Canadian carriers. of all, it was pointed out that the cost performances of the federally-regulated air- This finding may explain why these arguments were not proposed during 1938-78 when lines in the two countries were characterized by substantial similarities rather than by regulation was the common policy of both countries. Since Canadian airlines support differences. Second, weather was found to have little impact on airline performance. regulation, there was no need to present reasons for differentiating between Canadian Out of four measures, only below Category I minimums proved to have a statistically and U.S. policy during those years. The need to differentiate has developed since significant relationship with distance-adjusted operating expenses per RTM and RTM per 1978, however, and weather and population have provided useful bases for such argu- employee. While four Canadian carriers were high in this factor, three other fell well ments. But it not seems incumbent on any who continue to make these arguments to within the range encompassing the U.S. airlines, including the intrastate carriers. present comprehensive evidence supporting their assertions. Until such evidence is Further, indirect evidence concerning employment is particularly persuasive. Surely, forthcoming the opposite conclusion stands -- adverse weather and small population -17- do not significantly increase the operating costs of large Canadian airlines. -18 VI-39 VI-40 Footnotes
8Based on the lowest mean squared error, the best fit regression for the 1The following quotation from Mr. G. B. Hunnings, Assistant Vice-President, Public federally-regulated airlines was obtained from the following equation: Affairs, CP Air, summarizes this position: Y = 49.582 + 20,227.564/X, R2 = .866. The R2 is significant at the one percent level (see footnote 7). There are some who will claim that the wide differences between fares in Canada and the "efficient" cost of production of United States carriers 9Based on the lowest mean squared error, the best fit regressions for the federally- is not explainable by the fact that factor input prices are higher in regulated airlines were obtained from the following equations: Canada, that Canada has a Federal Sales Tax; more severe, generally Snowfall: = -5.169 + .089X, R2 = .165 speaking, weather conditions, and that the Canadian market is about a Category I Minimums: Y = -10.388 + 7.240X, R2 = .299 tenth the size and much more randomly distributed than U.S. markets. Minimum Temperatures: Y = 1.527 - .092X, R2 = .021 Maximum Temperatures: Y = -11.003 + 833.311X, R2 = .020 From "Regulating Canada's Airlines: Where Do We Co from Here?" paper presented at the National Conference on Airline Regulation, sponsored by the American Enterprise Institute Only the R2 for Category I minimums is significant at the five percent level (see and the Institute for Research on Public Policy, Ottawa (June 27, 1979), pp. 4-5. footnote 7).
2Transport Act, S.C. 1938, c. 53, parts I-II; and the Civil Aeronautics Act of 10Based on the lowest mean squared error, the best fit regressions federally- 1938, 52 Stat. 973. regulated airlines were obtained from the following equations: Snowfall: Y = 2,965.729 - 37.645X, .050 3W. A. Jordan, (a) Performance of Regulated Canadian Airlines in Domestic and Category I Minimums: Y = 8,850.579 - 5,544.321X, Transborder Operations (Ottawa: Consumer and Corporate Affairs Canada, 1982); and (b) Minimum Temperatures: Y = -302.373 + 4,440.765X, .135 Canadian Airline Performance Under Regulation (Ottawa: Economic Council of Canada, Maximum Temperatures: Y = 14,581.717 - 1,045,575.938/X, R2 = .055 Working Paper No. 29, April 1982). The analyses and conclusions of these two publi- cations and this paper are those of the author and do not necessarily reflect the views Only the R2 for Category I minimums is significant at the five percent level (see of any unit of the Canadian government. footnote 7). See Jordan (a), supra note 3, Chapter X, for a more detailed description of this analysis. 4Ibid (a), Table 1. For a detailed comparison of the powers and procedures of the CTC(A) and the CAB from 1938 to 1978, see W. A. Jordan, "Comparisons of American and 11Since these carriers all operate with the same size cockpit crews for any given Canadian Airline Regulation," in G. B. Reschenthaler and B. Roberts, eds., Perspectives aircraft type, these percentages are not influenced by that factor. For example, in on Canadian Airline Regulation (Montreal: Institute for Research on Public Policy, each case two pilots (rather than three) are used to operate two-engine aircraft. 1979), pp. 17-31. 12Collective agreements for Eastern Provincial and Pacific Western specify a com- 5Airline Deregulation Act of 1978, Public Law No. 95-504, 92 Stat. 1705. plement of four, rather than three, flight attendants on charter flights operated with Boeing 727 and 737 aircraft. "Agreement No. 1 Between Eastern Provincial Airways (1963) 6Atmospheric Environment Service, Airport Handbook (Toronto: Environment Canada, Ltd. and the Canadian Air Line Employees' Association (Flight Attendants)," Effective: 1975). Also, National Oceanic and Atmospheric Administration, Airport Climatological November 1, 1975, p. 6; and "Agreement No. 12 Between Pacific Western Airlines Ltd. and Summary, Climatography of the United States No. 90 (1965-1974) LAsheville, N.C.: U.S. the Canadian Air Lines Flight Attendants Association," Effective October 1, 1978, p. 58. Department of -Commerce, various dates]. 13The apparent differences of opinion can be seen by comparing the percentages for 7Based on the lowest mean squared error, the best fit regressions were obtained Eastern Provincial and Nordair with those for Pacific Western, Quebecair and Transair; from the following equations: and Air Canada with CP Air.
Snowfall: Y = 82.371 + 165.093/X, R2 = .020 14Jordan (b), supra note 3, Figures 2-1 and 2-2. The Canadian mainline carriers' Category I Minimums: Y = 82.659 + 2.879X, R2 = .006 economy fares per mile were virtually identical to the U.S. trunk carriers' coach fares Minimum Temperatures: Y = 88.825 - 9.523/X, R2 = .047 per mile. Maximum Temperatures: Y = 31.899 + 0.717X, R2 = .065 15Ibid., Table 3-1. For a random sample of 14 pairs, the five percent level of significance is achieved at an R2 of .283, while the one percent level of significance obtains at an R2 of .437. 16Ibid. The 1975-78 average operating ratios for the Canadian carriers were: Air Thus, none of these R2 is significant at the five percent level. S. B. Richmond, Canada = 94.4, CP Air = 95.8, Eastern Provincial = 104.0, Nordair = 92.4, Pacific Principles of Statistical Analysis (New York: The Ronald Press Co., 1956), p. 459. Western = 95.5, Quebecair = 98.3 and Transair = 94.0. The 14 federally-regulated airlines selected for this study do not constitute a random sample. They do, however, comprise more than 50 percent of the 26 federally- 17For example, institutional cities such as Ottawa, Washington and Las Vegas generate regulated airlines that operated large aircraft in scheduled passenger/cargo service more passengers per 1,000 population than do industrial cities such as Detroit and in Canada and in the 48 contiguous states of the U.S. during 1975-78. Montreal. Federal Aviation Agency, Air Traffic Patterns and Community Characteristics (1963), pp. 1-10. -19- -20- VI-41
18 The use of 1970 census data for the U.S. and 1971 census data for Canada was dictated by the need to have accurate figures for the various areas surrounding the airports served by each carrier. Such information is not available for 1978.
19The use of domestic population pools as an independent variable responds to the possible argument that these data reflect more accurately the true demand for each carrier's services. After all, Canadian carriers in the U.S. (and U.S. carriers in Canada) are limited to only transborder operating rights in contrast to their more extensive domestic rights. Also, the transborder community of interest is probably less than the overall community of interest within each country. As can be seen in footnote 20, the use of domestic population pools did not yield appreciably higher R2.
20Based on the lowest mean squared error, the best fit regressions were obtained from the following equations: 1.a. Y = 87.113 - 3,161.324/X, R2 = .000 b. Y = 89.512 - 21,609.266/X, R2 = .015 2.a. Y = 2.188 - .000056X, R2 = .045 b. Y = 2.003 - .000061X, R2 = .053 3.a. Y = -1,633.086 + .066X, R2 = .102 b. Y = -1,889.224 + .085X, R2 = .169 None of the R2 are significant at the five percent level (see footnote 7).
21-Using population per airport in place of population pool, the best fit regressions were obtained from the following equations based on the lowest mean squared error (see p. 14): 1.a. Y = 96.110 - .012X, R2 = .088 b. Y = 91.835 - .007X, R2 = .035 2.a. Y = 1.679 - 611.178/X, R2 = .035 b. Y = 2.812 - 842.663/X, R2 = .120 3.a. Y = -1,423.954 + 3.057X, R2 = .060 b. Y = -1,978.599 + 4.465X, R2 = .128 None of the R2 are significant at the five percent level (see footnote 7).
22Using population per airport carrier in place of population pool, the best fit regressions were obtained from the following equations based on the lowest mean squared error (see p. 14): 1.a. Y = 90.481 - .023X, R2 = .004 b. Y = 81.178 + .038X, R2 = .010 2.a. Y = 3.417 - 461.624/X, R2 = .041 b. Y = 4.905 - 596.231/X, R2 = .107 3.a. Y = -2,002.236 + 17.166X, R2 = .054 b. Y = -3,338.508 + 28.183X, R2 = .064 None of the R2 are significant at the five percent level (see *footnote 7).
23The 1970 populations of Dallas-Ft. Worth and Houston were 2,318,000 and 1,985,000, respectively, compared with 1971 populations for Montreal and Toronto of 2,729,000 and 2,602,000. Southwest's smallest area, Amarillo, had a population of only 144,000. Jordan (a), Table 35.
-21-