Sie sind auf Seite 1von 10

1.

Introduction

Air cargo is a critical area of business to most airlines. Despite rising fuel prices (Barber, 2010),
growth rates of nearly 6 per cent are predicted annually up to 2029. This makes the market one of
the fastest growing within the airline industry (Zhang and Zhang, 2002). Passenger figures for air
travel, by contrast, are expected to continue declining (Goll and Rasheed, 2011; Zhang and Zhang,
2002).Although the importance of air cargo to the airline industry cannot be underestimated, the
industry faces considerable challenges. One which has begun to attract the attention of scholars
concerns capacity planning (Amaruchkul and Lorchirachoonkul, 2011; Bish et al., 2004), or more
specifically, capacity "overbooking" (Kasilingam, 1997; Low et al., 2010; Rothstein, 1985). The
practice is popular within the industry because it allows air cargo companies to minimize losses of
revenue from unsold capacity. They achieve this by selling, months in advance, more cargo
capacity (volume and weight) than may be available to freight forwarders, (Amaruchkul et al.,
2011; Popescu et al., 2006). Fluctuating levels of uncertainty over the proportions of "shows" to
"no shows" prior to departure are therefore serious business considerations. Henceforth the
question of forecasting and forecast modelling inside the air cargo industry is critical to operational
efficiency.

This paper presents about a study commissioned by Singapore Airlines Cargo, which is one of the
major operator in Asia Pacific, in the market for air cargo and freight. Presently, the Singapore
Airlines Cargo is ranking 7th positon. Singapore Airlines has long been one of the most highly
recognized airlines, and its cargo division is no different. With dense operations throughout
Southeast Asia, the company has expanded service in North America, Europe and the Middle East.
Singapore Airlines Cargo operates 13 Boeing 747 planes, but no new orders are reportedly on the
horizon. (PIERCE, 2011). During 2002 SIA Cargo was able to reach 4.83 % and it was ranked
four in among global freight carriers. On the other hand, by 2010, SIA Cargo market share dropped
and ended up by 3.57 %. (Smeritsching, 2013) In a market comprising approximately 600 active
airlines worldwide with cargo capacity, Singapore Airlines Cargo's share of the market places the
company into the top quartile. The study sought after to undertake an evaluation of existing
forecasting techniques specific to Singapore Airline’s Cargo business. The study's objective was
to recommend a forecasting method that specifically enhanced precision of forecasting by the SIA
Cargo. Air cargo forecasting is regarded as a key capacity-planning component in air cargo and
freight operations. To achieve this objective, the study employed secondary data, journals, articles,
annual reports and electronic resources. Based on this, the following research question was raised:

RQ1. How can forecasting techniques for air cargo at Singapore Airlines Cargo be developed?

Next this brief outline introduction, as reminder of the paper is structured as follows. Secondly
following section, is about research aims and objectives. While a review of problem statement in
the next paragraph, the next following section, it presents an outline of the research methodology.
This is followed in the penultimate section which reports on the data analysis of the forecasting
methods (in this case World Air Cargo Forecast (WACF) analysis) which are used against given
period of secondary data. The paper concludes in section 6 with a discussion of the implications
of the findings and contribution to the knowledge and society.

2. Case overview

2.1 Singapore Airlines Cargo

It has been over 50 years that Singapore Airlines has been in the cargo business for over 50 years,
since its first flight was with the Airspeed Consuls in 1947. In July 1992, a full-fledged Cargo
division was formed to dedicate attention to the cargo business. As airline grow, in air cargo
evolution, the company corporatized to become an independently managed subsidiary of
Singapore Airlines on 1st July 2001.Singapore Airlines Cargo offers more than 900 flights a week
from its Singapore hub, at Singapore Changi Airport, linking more than 100 cities in over 35
countries across 6 continents. Singapore Airlines Cargo operates a dedicated freighter fleet and
also manages the cargo capacity on passenger flights operated by Singapore Airlines, Scoot,
SilkAir and NokScoot. (Cargo, Singapore Airlines, 2015) SIA currently operates nine freighters,
with eight more on firm order, all to be delivered between 2001 and 2005. In the year of review,
SIA Cargo's overall capacity expanded by 7.7 per cent and revenue reached $2,079 million, an 8.9
per cent increase on the previous year. (SIA Cargo Services, 2006)
3. Forecasting

3.1 Overview

There are two chief tactics to forecasting; this can either be undertaken in a qualitative or
quantitative form (Diebold and Mariano, 2002; Winklhofer et al., 1996; Wright and Rowe, 2011).
Several organizations frequently employ qualitative methods which includes explanations or
judgements (Jaffry and Capon, 2005; Sanders and Ritzman, 2004). In this approach, forecasting is
undertaken based on the knowledge and experience of managers or employees (Lui et al., 2011).
In contrast, the quantitative forecasting approach is more likely to rely on the availability of the
suitable and consistent numerical information, which is more likely to employ statistical methods
in analyzing data patterns (Wallström, 2009). Utilizing the appropriate forecasting models ensures
that management proactively matches supply and demand (Ballou, 2006), thus ensuring minimal
disruption to the supply chain. Perhaps, in decreasing unpredictability between supply and
demand, forecasting guarantees firms have increased capability to operate efficiently.

3.2 Air cargo forecasting

Forecasting is of actual importance in the air cargo industry because of the uncertainty linked with
real vs. expected "shows" and the resulting problems of overbooking and spoilage (Moussawi-
Haidar and Cakanyildirim, 2012; Wang and Kao, 2008). This generates extensive uncertainty in
size demand (Choi et al., 2012; Kalchschmidt et al., 2006). Henceforth forecasting can assist air
cargo operators not only to guess capacity demands, but also forecast operational profits and
income streams. Since time series analysis used in this study is a numerical time series data,
quantitative forecasting approach is adopted.
There are several reasons that, including the demand for revenue, capacity and routing
optimization, precision and accuracy remains a main business concern in air cargo forecasting
(Azadian et al. , 2012; Chiang et al. , 2007), with forecasting specialists being particularly deep to
know the prognostic role in terms of accuracy of models contrary to actual outcomes. Contrary,
such accuracy is very difficult because of the plethora of error measurement toolsets available
(Wallström, 2009). On the other hand, several scholars such as Armstrong (Carbone and
Armstrong, 1982; Armstrong and Fildes, 1995; Yokum and Armstrong, 1995), Hyndman and
Koehler (2006), and Fildes et al. (2009), recommends that the accuracy performance of any
forecasting model will mainly rely on data type and measurement of error. Another bibliographies
shows that (see Armstrong and Fildes, 1995; Yokum and Armstrong, 1995), that there are no
commonly accepted forecasting error measurement techniques or principles that can be used to all
forecasting situations. According to other scholars, like Teunter and Duncan (2009) at the
beginning that the application of only one error measurement is unable to establish accuracy of
forecast, scholars like Wallström, (2009) suggest that using several forecasting measurements in
order to thoroughly evaluate forecasting models. Overall, forecasting error identifies that
differences between the definite value observed and the forecasting or estimating value.

Conducting any study in the field of forecasting especially focused on the airline industry and more
precisely on the cargo market is one of the major challenge for researchers and scholars due to
several reasons. Firstly, there is substantial uncertainty as relates to "shows" (Becker and Dill,
2007). Uncertainty like this likely to impact on demand management (Croxton et al., 2002), and
hence the ability of cargo operators to be able to effectively balance customers' requirements with
the capability of their supply chain.

According to Peng and Chu (2009), the utilization of traditional regression analysis may not
necessarily be suitable for cargo demand forecasting, leading to a number of scholars such as Chou
et al. (2011) using fuzzy forecasting methods. A review of literature (see Fung, 2002; Lin, 2000;
Profillidis, 2000; Veenstra and Haralambides, 2001), has shown that regression has been employed
in an allied industry; the sea container industry. Veenstra and Haralambides (2001) for example
employ multivariate autoregressive analysis in their forecast of seaborne trade flows. Chou and
Liang (2003) applied the traditional fuzzy set regression along with regression analysis to estimate
the export and import cargo volume for ports in Taiwan and, in the process, found that fuzzy time-
series prediction produced more accurate results than seasonal time series. Fung (2002) forecasts
Hong Kong container volume using the error correction model while Chou et al. (2003) employed
the s ARIMA model to forecast container volume at Taiwan's Kaohsiung port. One particular study
of interest is that of Peng and Chu (2009) who sought to inspect the most practical and accurate
model for forecasting the monthly data series for sea containers. Six forecasting models are
employed in their study: classical decomposition, seasonal dummy regression, trigonometric
model, grey forecast, hybrid grey forecast and s ARIMAmodel. The main reason for their choice
of forecasting models to inspect being driven by earlier work of Taylor et al. (2006), who suggest
that univariate forecasting models are more suitable for forecasting where prediction times are
relatively short (approximately six hours or less).

4. Methodology

The aim of the paper focuses on the evaluation of forecasting to enhance the operational efficiency
of an air cargo operators. The objective of the paper to undertake an evaluation of forecasting
methods utilizing error accuracy measures in specific airline freight carrier. The study conducted
within Singapore Airlines Cargo. It included data in the form of historical period between years
from 2013 to 2016.

The challenge faced by Singapore Airlines Cargo related mainly to accurately projecting for
allocated cargo capacity. However, observing increasing focus on customer service within the
industry (Meng et al., 2010), a case study approach was adopted in the study. Case studies
(Eisenhardt, 1989; Eisenhardt and Graebner, 2007) are a popular means of conducting supply chain
management and logistics research (Ellram, 1996; Seuring, 2008).

According to Voss (2002) it is necessary to take note that, even though case studies are popular,
there is limitation on those case studies. Case studies could be useful for enhancing practitioner
insight and expanding better understanding on the certain issue. Nevertheless in order to provide
better understanding, relevant information and current up to date data the paper also included
annual reports, book and electronic resources. Most obviously, where data which is obtained from
several sources, caution should be shown when drawing general lessons for the industry as a whole.
The study was evaluated on the understanding that, strives to propose broader lessons to enhance
best practice in the air cargo industry.

5. Analysis and findings

5.1 Time series models

As it has stated earlier to evaluate cargo forecasting in Singapore Cargo Airline time series analysis
used. It is one of the method that has used in to analyze World Cargo Forecast to provide
comprehensive and up to date data for world air cargo. Time series model plans historical trends
into the future using time as the main independent variable. As time series estimates are one-
variable models, they require only the data for the variable to be forecast. Frequently while
forecasting, the forecaster has observations on only a single data series and has to develop forecasts
without being able to include other explanatory variables. In such a case only the past values of
this single variable are available for modeling and forecasting. In general, time series data can be
described by trends, seasonal effects and cyclical effects. (Boeing World Air Cargo Forecast Team,
2016)

5.2 Model selection

To employ time series model fully the first stage in putting together a time series forecast is to
analyze a time series of historical data for the specific airport (in this case Singapore Airlines Cargo
where it operates from Changi International Airport) or market in order to determine the growth
trend. The easiest process for dividing the trend in a time series is to plot the historical data in
graphic form on an x and y axis. The traffic data is plotted on the vertical (y) axis. Time, the
independent variable, is plotted on the horizontal (x) axis. It is then possible to draw a curve or
trend line which minimizes the total distance of the plotted points. In simple forecasts, it is likely
to extend this line into the future to estimate future traffic. Growth rates, positive or negative, can
be calculated from the slope of the line. Generally, expressing growth in terms of CAGR
(compound average growth rate) is recommended. Growth rates can then be applied to the base
year (or last year in the time series in order to project traffic into the future). Many popular
computer software programs, such as Microsoft Excel, automate the process of developing the
historical trend line. The primary drawback to this technique is that it fails to take into account
how the economic, demographic, and industry factors that existed during the historical period are
likely to be different in the future. (Boeing World Air Cargo Forecast Team, 2016)

5. 3 Observations from World Air Cargo forecast

The Boeing Company concerns the biennial World Air Cargo Forecast (WACF) to provide a
comprehensive up-to date overview of the air cargo industry. The forecast summarizes the world’s
key players in the air trade markets, detects growing trends, and presents forecasts for the future
performance and development of markets as well as for the world freighter airplane fleet.
Afterwards recovering above than 19% in 2010 over the depressed levels of 2009, world air cargo
traffic went badly from mid-2011 to early 2013. This extended period of weak growth can be
qualified to two factors which are: a poor or economic dropdown of world economy and slack of
trade growth. Since 2008 during the period of the global economic crisis, world air cargo traffic
has averaged only 1.7% growth per year till to 2013. However starting from in second quarter of
2013, world air cargo traffic began to grow again. By July 2014, traffic had increased 4.4%
compared with the first seven months of 2013. Forecasts for even better economic and trade growth
should lead to sustained air cargo traffic growth in 2015 and 2016. (Boeing World Air Cargo
Forecast, 2014-2015).
6. Conclusions

Nowadays current business environments, organizations will depend extensively on several


management tools and techniques like forecasting in order to make strategic decisions. Forecasting
is a serious business activity in that it allows the organization to align the pace of resource
distribution to changes which are expected within an ever changing business environment
(Makadok and Walker, 2000; Durand, 2003). At the end the outcome is forecasting enables
organizations manage their fixed costs and enhance efficiencies and performance.

It is important for air cargo that forecasting is important. The main challenge for air cargo operators
is not only competition between rivalries, demand management as well which is with high
uncertainty is also crucial. Perhaps there are about 600 active cargo operators worldwide with
cargo capacity. With this knowledge, the study sought to undertake a commissioned evaluation of
current 3 year forecasting utilized by Singapore Airline Cargo.

As it was stated in earlier paragraphs the primary objective of this paper to evaluate a 3 year cargo
performance set by Singapore Airlines Cargo. At the beginning analysis was to undertake a
preliminary analysis of existing forecasting methods and accuracy measures. Secondly, next
objective analyzed specific forecasting methods (in this case Time Series Model used interpreted
from World Air Cargo Forecast (WACF) and then to test for accuracy against 3 years of historical
data provided by the company. Lastly, there are some recommendations given to Singapore
Airlines Cargo on its forecasting strategy. To undertake the evaluation, demand patterns drawn
from historical data on over a three-year period covering Singapore Airlines Cargo routes was
utilized. A major element of the forecasting operations focused on the enhancement of forecasting
precision based on time series data which are described in charts, as one of the key capacity-
planning component in air cargo and freight operations. To complete this task, one forecasting
method used, the time series method, were utilized to forecast based on the yearly data of the total
network. One of the crucial conclusion of this study is that incorporating dynamic functionalities
into forecasting will enhance the precision of forecasting time series analysis by Singapore Airlines
Cargo.
However question may arise, within the context of demand management, what really does
incorporating dynamic functionalities into forecasting imply? There is way of imply the placement
of forecasting firmly within the realm of intelligence management. In result traditional role of
forecasting should be extend from its traditional base within business intelligence (Gerybadze,
1994; Lemos and Porto, 1998), to an environment where the right organizational conditions are
established to ensure that analyst are able to utilize both explicit historical information (Fischer
and Harvey, 1999) and implicit knowledge (McConnell et al. , 2011) to project the future. Such an
environment encompasses the environment of Intelligence Management of which from a review
of earlier literature (see Maguire et al., 2009; Ojiako et al., 2012), the process of forecasting is
critical. The nature of the forecasting process involves that managers should be careful during
employing either individual forecasting methods, or when considering mixture of different
forecasting methods. Additionally, since including dynamic functionalities into forecasting does
have the potential to enhance accuracy, the need for managers to pay greater attention towards
applying combined forecasting (see Holmström et al., 2002; Nakano, 2009), it means that
improving the performance of supply chains needed to be repeated. Studies conducted by Sanders
and Graman, (2009), suggest that a decreasing errors of forecasting can generate more profit to
the firm and it may increase by up to 30 per cent, such an audit has the possibility of throwing
more light of the cost implications of possible errors in current forecasting being undertaking by
Singapore Airlines Cargo.

Generally, looking at the future of the industry itself and to how researches and studies can help
to support its growth, a dynamic way forward may be to advance methods to feed information on
forecasting sharpness into cargo pricing mechanisms to reduce them more clearly. In other areas
like weather forecasting, a major obstacle is to sharpen forecasts (i.e. to reduce what is variously
termed "uncertainty" or "ignorance") whereas clearly communicating, remaining uncertainty to
the general public (Roulston and Smith, 2002). Likewise, in the sphere of insurance,
communicating levels of uncertainty to customer's matters greatly. According to Cabantous
(2007), levels of uncertainty present in the minds of actuaries when setting premium levels,
particularly when actuaries disagree on levels of uncertainty, can cause premiums to skyrocket.
Henceforth it is common to measure levels of uncertainty within insurance loadings so that
insured's can have more precise knowledge of how their premiums are calculated. If air cargo
operators were to move to more transparent pricing strategies, based on the discovery of levels of
uncertainty found to exist on a month-by-month basis in the form of success rates calculated for
their use of combined predictive tools, the logic of instrumental stakeholder theory would then
apply to the relationship between the Cargo Operator and its most regular high volume customers.

Das könnte Ihnen auch gefallen