Content uploaded by Tzu-Ming Liu
Author content
All content in this area was uploaded by Tzu-Ming Liu on Dec 09, 2016
Content may be subject to copyright.
Tourism Economics, 2014, 20 (1), 143–156 doi: 10.5367/te.2013.0258
Analysis of the economic impact of
meteorological disasters on tourism:
the case of typhoon Morakot’s impact
on the Maolin National Scenic Area
in Taiwan
T
ZU
-M
ING
L
IU
Graduate Program of Sustainable Tourism and Recreation Management, National TaiChung
University of Education, No 140 Min-Sheng Road, TaiChung 40306, Taiwan.
E-mail: liutm.tw@gmail.com.
This study explores the cost of extreme weather disasters to the
tourism industry, taking the Taiwan Maolin National Scenic Area as
an example. The paper evaluates the economic damage caused by
typhoon Morakot. The author uses long-term tourist trend lines as
the basis for calculating disaster losses and adopts an error correction
model as a measurement method for the estimation. The study finds
that the entire park lost over 700,000 visitors in the year and a half
after the disaster, representing a loss of NT$1.39 billion in tourism
business – a value three times the infrastructure loss. Any delays in
reconstruction will increase the losses to the tourism industry. However,
several tourist spots were gradually losing their attractiveness even
before the disaster, and hastening to rebuild such sites is not
conducive to the recovery of tourism. Rather, efforts should first be
made to understand the reasons why these tourist spots were becoming
less attractive and to gauge tourist demand for them before their
reconstruction.
Keywords: weather disaster economic costs; climate change; tourism
demand analysis; disaster recovery; error correction model; Taiwan
Tourism is deeply influenced by climatic conditions (Gómez-Martín, 2005;
Amelung et al, 2007; Scott et al, 2007; Conrady and Bakan, 2008; Taylor and
Ortiz, 2009; Scott, 2011), and so scholars have actively studied the impact of
climate change on tourism. These studies have explored the influence of the
Many thanks to anonymous reviewers for their helpful comments on an earlier draft, and to the
National Science Council of Taiwan for research grants (NSC 97–2621-M-424–002 and NSC 100–
2621-M-006–001) under which this work was completed.
T
OURISM
E
CONOMICS
144
amount of rainfall, temperature and other meteorological conditions on tourist
numbers and then combined these factors with climate models to predict the
impact of climate change on future tourism markets. The relationship between
climate change and tourist numbers has yielded clear results on the national,
regional and global scales (Hamilton and Tol, 2007; Eugenio-Martin and
Campos-Soria, 2010; Rosselló-Nadal et al, 2011). Many studies have
specifically evaluated the impact of long-term temperature increase on the
tourism market. For example, a severe coral bleaching of Australia’s Great
Barrier Reef caused by sustained high temperatures could decrease tourist visits
by 35% (Prideaux, 2006). Rising temperatures, which can result in the
disappearance of glacial landscapes (Brugman et al, 1997; Hall and Fagre,
2003), changes in vegetation cover, habitat reduction and even the extinction
of organisms (Halpin, 1994; Cumming and Burton, 1996; Scott et al, 2002;
Hall and Fagre, 2003), are projected to decrease the number of visitors by 19%
and 31%, respectively, at Glacier-Waterton International Peace Park and Banff
National Park (UNWTO-UNEP-WMO, 2008).
The aforementioned studies used long-term meteorological data to assess the
impact of long-term temperature increases caused by climate change on tourism.
However, the impact of another characteristic of climate change – extreme
weather conditions – is rarely discussed. In recent years, the global frequency
of extreme weather events has increased substantially, including some of the
worst flooding in half a century in Thailand in 2011 and Australia in 2012
and a serious drought in China during the same period. Climate disasters have
caused serious damage to the tourism industry, and some tourist spots have even
been destroyed; for example, a 2010 flood destroyed Inca ruins in Machu
Picchu, Peru. However, previous analysis of the impact of such extreme
meteorological disasters on tourist visitation is inadequate. Analysis of the
damage caused by extreme meteorological disasters is crucial for planning
climate change strategies in the tourism industry. Such analysis should include
an estimation of the scale of natural disasters, an assessment of the opportunity
cost of reconstruction and a provision of standards for comparison in a recovery.
To compensate for the inadequate prior research on the impact of extreme
meteorological disasters on tourism, this study investigates the impact of
typhoon Morakot on the Maolin National Scenic Area in Taiwan, estimates the
damage caused by this disaster (also called the 88 typhoon disaster in Taiwan)
and the opportunity cost of reconstruction, and assesses the priorities for
allocating reconstruction resources.
Typhoon Morakot hit Taiwan on 7 August 2009, resulting in 673 deaths,
26 missing people and NT$27.94 billion in property damage, of which NT$2.18
billion was damage to tourist facilities (Morakot Post-Disaster Reconstruction
Council, Executive Yuan, 2010). Among the major scenic attractions, Maolin
National Scenic Area had the most serious damage. The Maolin National Scenic
Area was established on 2 October 2001 as one of the tourist spots with the
most potential in southern Taiwan. The Purple Butterfly Valley in this area is
one of the two world’s largest wintering butterfly valleys, with high ecological
importance and tourism value. The heavy rain brought by typhoon Morakot
disrupted the vast majority of local traffic in the Maolin National Scenic Area.
The management headquarters and Liugui Visitor Center collapsed. Landslides
destroyed most hotels, hot springs and scenic spots. Property losses totaled
145Impact of meteorological disasters on tourism
NT$4.6 billion (Morakot Post-Disaster Reconstruction Council, Executive Yuan,
2010).
However, the impact on tourism is not limited solely to tangible facilities.
The invisible decline of tourist activity was another outcome of this disaster.1
Decreased tourist numbers can be used to estimate reduced tourism activity.
The tourist numbers projected for the case in which the disaster did not occur
constitute the benefit that could have been generated (the opportunity cost).
This cost must be counted as one of the losses from the typhoon disaster.
Currently, only project costs are listed in Taiwan’s official typhoon disaster
losses. To correctly account for typhoon Morakot’s impact on the Maolin
National Scenic Area, both the scale and the economic value of tourist activity
loss due to the typhoon disaster should be carefully estimated.
The assessment of tourist visits or tourist demand is the basis of damage
analysis. Tourism demand forecast can provide information on the past status
and future trends of tourism and serve as the basis for decision making during
reconstruction. Thus, to reconstruct effectively the Maolin National Scenic
Area, the Taiwanese government should recognize the status of tourists’
activities and tourism demand in this region, and then define the priorities in
each reconstruction phase.
Tourism demand analysis is rarely used to assess disaster damage. Bonham
et al (2006) and Gut and Jarrell (2007) studied the impact of the September
11 terrorist attack on the tourism industry. Kuo et al (2008) and Huang and
Min (2002) assessed the impact of SARS and avian flu and the 921 earthquake
on the Taiwanese tourism industry. Other tourism disaster assessments include
those for the global financial crisis (Song and Lin, 2010). Bonham et al adopted
an error correction model (ECM) to assess the tourist demand function. This
ECM can be combined with overall economic data to predict changes in tourist
numbers. This model is suitable for the estimation of the opportunity cost of
post-disaster reconstruction and meets the requirements of this study. Therefore,
we followed the method of Bonham et al to estimate tourism demand and
predict tourist trends for the Maolin National Scenic Area. Based on the results,
we calculated the output loss in tourism caused by the damage brought by
typhoon Morakot. Our research targets four major tourist spots in the Maolin
National Scenic Area: the Maolin Scenic Area, the Taiwan Aboriginal Culture
Park, the Wutai Recreation Area and the Baolai and Pu-lao Hot Springs.
The data analysed in this study can be divided into two parts: tourist
numbers and other general economic variables. Tourist numbers were acquired
from the Tourism Bureau (2010) in the form of monthly tourist numbers at
major domestic tourist spots. Maolin National Scenic Area Administration is
officially responsible for determining the tourist numbers and uses a variety of
methods to estimate the number of tourists visiting the park. These methods
include mandatory visitor registrations in the Wutai Recreation Area, an
entrance ticket requirement to the Taiwan Aboriginal Culture Park, and visitor
contacts at visitor stations in Maolin Scenic Area and Boalai and Pu-lao Hot
Springs. The visitor station of Maolin Scenic Area is located at the only entrance
of the scenic area and the park staff count visitors passing the visitor station.
Visitor counts from each site have to be reported within two weeks of the end
of each month. Other general economic variables were sourced from the national
income and economic growth database at the DGBAS (Directorate General of
T
OURISM
E
CONOMICS
146
Budget, Accounting and Statistics) (2010). All data are reported on a monthly
basis, covering the period from January 2001 to December 2010. In addition,
per capita tourist spending data were taken from the 2010 National Tourist
Survey Report (Tourism Bureau, 2011).
The introduction of this paper summarizes the purpose of this study and
research methods. The second section describes the significance of the long-term
tourist trend line and its policy implications. We discovered that using the
tourist trend as a basis can help us to understand the impacts of disasters on
tourism activity and to evaluate the adequacy of reconstruction efforts and their
priorities. The third section uses the error correction model to predict long-
term tourist trend lines at the major sightseeing sites in the Maolin National
Scenic Area and estimate the influence of the typhoon disaster on tourism
activity changes. Such changes can be represented as numbers of tourists or
alternatively represented by currency after monetization for comparison with
other losses from the typhoon disaster. Our research conclusions are presented
in the final section.
The economic view of the impact on tourist numbers
It is generally believed that when tourist numbers are restored to pre-disaster
levels, the tourism industry2 is considered to have recovered from a disaster.
However, this view does not consider the influence of overall economic changes
during the reconstruction period and thus misjudges the damage to the tourism
industry from environmental disasters. This error becomes exacerbated as the
reconstruction schedule is extended. To correctly measure typhoon Morakot’s
impact on Maolin tourism, we must consider both the length of the reconstruction
schedule and the original tourism activity without the disaster.
Tourist numbers at the Maolin National Scenic Area are influenced by the
overall environment including population and economic growth. By only
examining the total number of tourists at each time point without carefully
studying each variable’s influence on the tourist numbers, we would find that
the number of tourists demonstrates a certain trend, as illustrated by the trend
line in Figure 1. If the number of tourists before the typhoon is represented
by point A following the impact of the typhoon disaster the number of tourists
would drastically fall to point B; subsequently, the number of tourists may rise
to point C, recovering to the pre-disaster level. If reaching point C is considered
the full recovery of tourism activities, then the impact on the tourism industry
is represented by ∆ABC. This view assumes that the external environment
remains unchanged or that the external environment is irrelevant to the number
of tourists visiting the scenic area. Such assumptions are clearly inconsistent
with the characteristics of the tourism industry.
The inadequacy of using pre-disaster tourist numbers as the benchmark for
recovery can be easily demonstrated by an example at time T. Suppose the
disaster did not happen; under normal circumstances, the number of tourists
is D at time T. Owing to the typhoon disaster, the actual number of tourists
at time T is E. Therefore, DE is the number of tourists lost due to the disaster,
indicating the impact on tourism (the total impact of the disaster is ∆ABG,
a far greater impact than ∆ABC). However, if A is used as the basis point,
147Impact of meteorological disasters on tourism
Figure 1. Change in the number of tourists after the disaster.
because the number of tourists at time T is already higher than point A, it
could be considered that the tourism industry has already recovered and
experienced growth EF.
Using a time trend as the basis can demonstrate the overall influence of the
external environment on tourism activity during the reconstruction period and
thus reflect the opportunity cost of the disaster. Using a pre-disaster value as
the basis, however, ignores the original features of the scenic area during the
reconstruction period. Comparing the two approaches, we find that the latter
would misjudge the impact of the disaster. Therefore, a time trend should be
used as the basis for typhoon disaster assessment.
The impact of typhoon Morakot on the tourist numbers at the Maolin
National Scenic Area is illustrated in Figure 2. The line of asterisks represents
the time trend before the typhoon disaster, representing the monthly numbers
of tourists from January 2001 to June 2009. The thin solid line represents the
time trend after the disaster as the monthly number of tourists after the disaster
until December 2010. July 2009 is excluded because of possible errors in that
month’s data. July 2009 was one month before typhoon Morakot, when tourism
activity had not yet been affected by the typhoon. However, the visitor statistics
of July 2009 were incomplete because of the disaster. Tourist numbers should
be consistent with the time trend, but the recorded number is much smaller
than it should be, approaching zero. This anomaly likely represents an error
in the statistics rather than a sharp decline in tourist numbers. To avoid
propagating this error, data from July 2009 were removed from our study.
According to the historical data on tourist statistics, the numbers of visitors
at the four major attractions at the Maolin National Scenic Area display a high
degree of inconsistency. In terms of the total numbers of visitors, the Baolai
and Pu-lao Hot Springs are the most popular, followed by the Taiwan
Aboriginal Culture Park. The average number of total monthly visitors at these
four attractions can reach as high as 200,000 but can also be near zero. These
data show not only that tourism activity at the Maolin National Scenic Area
is influenced by seasonality but that the popularity of the different attractions
also varies. From the perspective of time trends, only the tourist numbers at
A
B
C
D
E
F
G
T
Tourist
numbers
Time
Disaster
occurred
Disaster
recovery
Long-term
trend line
for tourist
numbers
T
OURISM
E
CONOMICS
148
Figure 2. Comparison of the pre-disaster and post-disaster tourist numbers at the Maolin National Scenic Area.
Maolin Scenic Area
Taiwan Aboriginal Culture Park
2000M04
2000M12
2001M08
2002M04
2002M12
2003M08
2004M04
2004M12
2005M08
2006M04
2006M12
2007M08
2008M04
2008M12
2009M08
2010M04
2010M12
2003M04
2003M08
2003M12
2004M04
2004M08
2004M12
2005M04
2005M08
2005M12
2006M04
2006M08
2006M12
2007M04
2007M08
2007M12
2008M04
2008M08
2008M12
2009M04
2009M08
2009M12
2010M04
2010M08
2010M12
Wutai Recreation Park
Baolai and Pu-lao Hot Springs
2005M04
2005M08
2005M12
2006M04
2006M08
2006M12
2007M04
2007M08
2007M12
2008M04
2008M08
2008M12
2009M04
2009M08
2009M12
2010M04
2010M08
2010M12
0
2000M04
2000M12
2001M08
2002M04
2002M12
2003M08
2004M04
2004M12
2005M08
2006M04
2006M12
2007M08
2008M04
2008M12
2009M08
2010M04
2010M12
Number of tourists Pre-disaster linear trend Post-disaster linear trend
×
*
140,000
120,000
100,000
80,000
60,000
40,000
20,000
0
–20,000
80,000
70,000
60,000
50,000
40,000
30,000
20,000
10,000
0
–10,000
–20,000
–30,000
25,000
20,000
15,000
10,000
5,000
0
200,000
150,000
100,000
50,000
0
149Impact of meteorological disasters on tourism
the Baolai and Pu-lao Hot Springs show positive growth, whereas the tourist
numbers at the other three parks are gradually decreasing, especially that of
the Maolin Scenic Area.
The effects of the disaster on the four attractions were also inconsistent. The
Baolai and Pu-lao Hot Springs were the most popular attractions before the
typhoon disaster and were thus influenced the most. After examining the post-
disaster tourist numbers, we found that the post-disaster tourist trend was still
far from an upward trend line before the disaster, showing that tourism
activities at Baolai and Pu-lao Hot Springs have not recovered yet. Although
the post-disaster tourist numbers at the Wutai Recreation Area are gradually
trending towards pre-disaster numbers, this increase is likely not because the
post-disaster trend is upward but rather because pre-disaster tourist numbers
were already decreasing. Tourism activities recovered quickly at the Taiwan
Aboriginal Culture Park, almost keeping up with the pre-disaster trend. Tourist
numbers at the Maolin Scenic Area are also recovering gradually.
Comparing the tourist trends before and after the disaster, we discovered that
in rebuilding tourism activity at the Maolin National Scenic Area, the
primary focus should be placed on the reconstruction of the Baolai and Pu-lao
Hot Springs. The Baolai and Pu-lao Hot Springs were the most important
attractions before the disaster and have failed to show any signs of recovery,
suggesting the necessity of reinforcing reconstruction activities at these sites.
The Maolin Scenic Area and the Taiwan Aboriginal Culture Park show more
positive responses to reconstruction. Although tourism activity at both sites was
decreasing before the disaster, they have recovered quickly after the disaster.
Tourist numbers at Wutai Recreation Area were sparse before the disaster and
have remained limited after the disaster.
Studying historical data for tourist statistics can help us to understand the
adequacy of post-disaster reconstruction efforts. Despite some increases, the
tourist numbers at Baolai and Pu-lao Hot Springs remain far from what they
should be, suggesting inadequate reconstruction efforts in the area. Although
the attractiveness of the Maolin Scenic Area and the Taiwan Aboriginal Culture
Park was declining before the disaster, it recovered quickly after the disaster.
Therefore, by leveraging post-Morakot reconstruction, it is possible to resurrect
the Maolin Scenic Area and Aboriginal Culture Park as domestic tourist
hotspots. The recovery strategies for these two areas should not be limited to
considering the recovery of their original tourist numbers but should instead
first focus on understanding the reasons for the decline of their attractiveness
to tourists and identifying these two areas’ positions in the domestic tourism
industry in the future before undertaking infrastructure reconstruction. The
Wutai Recreation Area was a popular recreational area (with monthly numbers
of visitors exceeding 20,000), but its number of visitors had been gradually
decreasing. The downward trend continued after the disaster. Therefore,
rebuilding Wutai Recreation Area should be a lower priority in the reconstruction
planning at the Maolin National Scenic Area.
In the previous discussion, we discovered that using a tourist trend is helpful
for understanding the impact of a disaster on tourism activity. It also helps to
compare the trends before and after the disaster in each area and evaluate the
adequacy of reconstruction tactics and priorities. However, a graphical analysis
can only grasp general conditions and does not take full advantage of the
T
OURISM
E
CONOMICS
150
information available in statistics. To fully capture the impact and evaluate the
effects of typhoon Morakot on tourism activity in the Maolin National Scenic
Area, we used an econometric tourism model to estimate the long-term tourist
trend lines at each of the major sightseeing sites. Based on the trend lines, we
calculated the loss of tourism output at Maolin.
Quantitative analysis of the number of visitors
In this study, we used an econometric tourism model to evaluate the major
factors that influence the long-term tourist trend lines at each major tourist
site in the Maolin National Scenic Area. The results were then used as bases
for calculating tourism output loss. Because autocorrelation and unit roots (Song
et al, 2008) can exist in time series data in the tourism industry and affect the
statistical nature of linear regression estimates, we must first determine the
variables and then choose a suitable econometric model according to the
statistical characteristics of the data.
The existence of a unit root reflects that the data reflect stationary time series
variables. Traditional quantitative analysis methods assume that all variables are
stationary time series. Regression analysis using non-stationary time series data
may produce spurious regression phenomena (Granger and Newbold, 1974;
Engle and Granger, 1987; Granger, 2007). In this situation, least squares
estimates will not be consistent, and the variance of the residual will approach
infinity as time increases. In addition, external shocks (such as a typhoon
disaster) occurring in a time series only produce temporary influences on
stationary time series but have permanent influence on non-stationary time
series. Therefore, econometric models cannot be applied directly to non-station-
ary time series. Differential calculations must be applied to non-stationary time
series data to convert them into stationary time series. If a non-stationary time
series becomes stationary after a time difference operation d, such a time series
is called integrated in order d, denoted as Xt~I(d).
We applied the augmented Dickey–Fuller (ADF) test (Engle and Granger,
1987) to tourist numbers for the major tourist spots at Maolin National Scenic
Area to examine unit roots and found that unit roots exist in the data of all
four parks. Therefore, a regression analysis (Table 1) cannot be directly applied.
Because first-order differential operations removed all the unit roots, all of the
sequences at the Maolin National Scenic Area are therefore I(1) sequences.
Table 1. Unit root tests for tourist numbers.
Original values of First-order differences
of tourist numbers in tourist numbers
Maolin Scenic Area –4.45 –18.90***
Taiwan Aboriginal Culture Park –2.83 –17.24***
Wutai Recreation Area –3.08 –12.91***
Baolai and Pu-lao Hot Springs –3.07 –15.60***
Note:***p < 0.01.
151Impact of meteorological disasters on tourism
Traditional econometric model cannot be used for the analysis of non-
stationary time series data in the tourism industry. To address this problem,
Song, Witt and Li (2008) proposed an error correction model, following models
first proposed in the 1960s. According to Engle and Granger (1987), if a
co-integration relationship exists between two variables, the relationship can be
represented by an error correction model. This concept explains short-term
changes in relationships among sequences and the process of adjusting from a
short-term imbalance to a long-term equilibrium using the imbalance in the
long-term co-integration relationship to adjust and fix short-term dynamics.
Therefore, changes in the current variables are influenced by factors including
previous error correction items, previously changed auto-correlated items, and
previous changes in other variables. In summary, error correction consists of
corrections and adjustments in the current time cycle to compensate for the
deviation of residuals from the previous time cycle.
The advantages of the error correction model compared with other econometric
models include the following. (a) The error correction model simultaneously
demonstrates short-term and long-term equilibriums. (b) The error correction
model avoids false correlation issues among variables (such as spurious
regressions). Because the majority of time series data in tourism are non-
stationary data, this method can solve false correlation issues. (c) The error
correction model can reduce regression collinearity problems. (d) The error
correction model avoids the controversy of data mining. (e) The error correction
model does not presume relationships among parameters but is rather a
generalized model (Song et al, 2008).
The error correction model can be expressed as follows:
∆yt =
β
0∆xt – (1 –
φ
i)[yt–1 – k0 – k1xt–1] +
ε
t.
The Engle–Granger two-stage estimation method is often used for such esti-
mations. However, this method does not yield long-term equilibrium estimates
and produces errors in the analysis of small samples. Another estimation
technique is the Wickens–Breusch one-step estimation method (WB1S), ex-
pressed as follows:
∆yt =
α
+
β
0∆xt +
φ
i∆yt–1 +
λ
1yt–1 +
λ
2xt–1 + ut. (1)
The parameters determined using the WB1S method are consistent, valid and
unbiased, regardless of whether long-term or short-term relationships are
analysed. Therefore, in this study, we used WB1S method to estimate the model
parameters.
The dependent variable in the tourist demand model (the tourism long-term
trend line) is the number of tourists (T). There are many possible options for
the independent variables. Based on the results of recent comparative analyses
of tourist needs (Crouch, 1995; Song and Li, 2008), we chose Taiwan’s total
population (POP), per capita national income (GDP) and consumer price index
(PI) as the independent variables for the model. To estimate the parameters,
we first require the differentials of each variable (DPOP, DGDP, DPI and DLT)
and one-step-backwards values for each variable (LPOP, LGDP, LPI and LT).
Then, plugging the results into Equation (1), we obtain the values listed in
Table 2.
T
OURISM
E
CONOMICS
152
Table 2. ECM estimation results (based on the Wickens–Breusch one-step approach).
Intercept DPOP DGDP DPI LPOP LGDP LPI DLT LT
Maolin –312.79 1045.57*–3.18 –0.10 21.94 –3.70** –0.08** 0.01 –0.83***
(299.63) (573.12) 2.25 0.07 18.74 (1.70) (0.04) 0.12 (0.15)
Culture Park –326.91 672.08 0.23 –0.10 19.77 0.46 –0.07 –0.18 –0.64***
(420.78) (499.46) (2.06) (0.08) (25.46) (1.34) (0.05) (0.13) (0.17)
Wutai –1291.83 214.76 –1.93 –0.03 78.34*–1.97 –0.07 –0.06 –0.57***
(799.65) (972.50) (4.61) (0.15) (48.24) (3.45) (0.08) (0.18) (0.19)
Baolai –261.02 505.92 –2.00 –0.04 15.85 –0.06 –0.03 0.00 –0.36***
244.36 462.43 1.75 0.06 515.14 1.25 0.03 0.10 (0.08)
Note:*p < 0.1; **p < 0.5; ***p < 0.01.
Table 3. Differences between actual and estimated tourist numbers (units: visits).
Maolin Scenic Aboriginal Wutai Recreation Baolai, Pu-lao
Area Culture Park Area Hot Springs
2009Q3 –17,197 –14,564 –16,001 –98,468
2009Q4 –12,634 –25,078 –14,481 –103,206
2010Q1 2,007 22,730 –8,836 –86,974
2010Q2 3,252 –13,922 –13,158 –98,925
2009Q3 –251 –13,628 –13,016 –98,521
2009Q4 –5,801 8,058 –11,368 –91,399
As shown in Table 2, by inputting the total population, per capita national
income and consumer price index into the model, we estimated the expected
numbers of tourists per month if the typhoon had not occurred. After subtract-
ing the actual monthly tourist numbers from the estimated numbers of tourists,
the residual numbers represent the impact of typhoon Morakot on tourism. The
quarterly aggregated differences in the numbers of visitors are listed in Table
3. Taking the Maolin Scenic Area as an example, the actual number of tourists
in the third quarter of 2009 is 17,000 below the expected number. This
difference is due to the typhoon and can therefore be considered the impact
of the disaster on tourism. From immediately after the disaster until the end
of 2010, the total number of tourists in the entire Maolin National Scenic Area
was reduced by 721,381 visitors, suggesting that the Maolin National Scenic
Area not only suffered serious property damage to buildings, roads and tourist
sites but also suffered a significant loss in the number of tourist visits.
In addition to the number of tourist visits, the invisible tourism damage can
also be measured by the output of the tourism market. To monetize such a loss,
we must estimate the value of each tourist visitor or the average value of all
visitors. Because the tourism database estimates only national average spending
on tourism based on an annual tourism survey and does not include tourists’
spending at each tourist spot, we can estimate only the output loss as the
153Impact of meteorological disasters on tourism
Table 4. Tourism output fluctuations due to changes in tourist numbers (units: thousand
NT$).
Maolin Scenic Aboriginal Wutai Recreation Baolai, Pu-lao
Area Culture Park Area Hot Springs
2009Q3 –33,035 –27,978 –30,737 –189,156
2009Q4 –24,270 –48,175 –27,818 –198,259
2010Q1 3,856 43,664 –16,974 –167,077
2010Q2 6,246 –26,744 –25,277 –190,035
2009Q3 –482 –26,180 –25,004 –189,258
2009Q4 –11,144 15,480 –21,838 –175,577
spending per person per trip from the survey report. According to the national
tourism survey report in 2010, the annual tourist spending per person per trip
is NT$1,921. Using this spending level and the changes in tourist visits listed
in Table 3, we can calculate the change in tourism output at the Maolin
National Scenic Area; the results are shown in Table 4. As of December 2010,
total tourism output was reduced by NT$1.39 billion after the typhoon disaster.
This loss is three times the loss from property damage, showing that
considering only facility damage seriously underestimates the impact of the
typhoon disaster.
Conclusion
This study estimated the impact of typhoon Morakot on tourism activity in
the Maolin National Scenic Area. The estimation was measured in tourist
numbers and New Taiwan dollars. This study marks the first report of the use
of long-term tourist trend lines supplemented with an error correction model
as a basis for measuring disaster losses in Taiwan. Tourism losses from the
typhoon disaster, long-term tourist trend lines and error correction models are
rarely discussed in the tourism industry literature. However, only by combining
these three methods can we correctly assess the losses caused by this disaster.
Although a series of natural disasters occurred, the nation is still actively
promoting tourism. A correct tourism damage assessment method is urgently
needed to assist in formulating tourism-related policies. This study fills this
research gap in the tourism industry.
The use of long-term tourist trend lines as bases can facilitate a more in-
depth understanding of disaster impacts on tourism activities and help us grasp
the priorities of post-disaster reconstruction by comparing trends before and
after a disaster. The Maolin Scenic Area and the Taiwan Aboriginal Culture Park
have recovered considerably. The Baolai and Pu-lao Hot Springs are currently
the most popular tourist attractions but suffered the most serious impacts.
There remains a considerable distance to full recovery, and more resources
should be invested in the reconstruction of these areas. The number of tourists
who were there was sparse before the typhoon disaster and remained sparse after
the disaster. Therefore, the priority for the Wutai Recreation Area can be
reduced in reconstruction planning for the Maolin National Scenic Area.
T
OURISM
E
CONOMICS
154
The reconstruction of the Maolin Scenic Area and the Taiwan Aboriginal
Culture Park has yielded considerable results. However, reconstruction should
not be performed hastily, particularly just restoring the area to its original state.
The long-term tourist trend line shows that the attractiveness of the two regions
to tourists was decreasing before the disaster. Simply restoring damaged
facilities may not reverse the declining trend in tourist numbers even after
significant reconstruction, which would drastically reduce the effectiveness of
the allocated reconstruction funds. In contrast, if we first investigate the reasons
for the decline in tourist attractiveness before the disaster and consider the
positions of these two regions in the future domestic tourism market and
employ suitable tactics, we may be able to restore the attractiveness of the
Maolin Scenic Area and the Taiwan Aboriginal Culture Park and generate the
benefits of increased national tourism.
A unit root test revealed that tourist numbers are I(1) sequences, so this error
correction model was used to estimate long-term tourist trend lines for the four
major parks in the Maolin National Scenic Area. Based on these trend lines,
we calculated the damage to tourism caused by typhoon Morakot. The entire
park lost 700,000 tourist visits during the 18-month period after the disaster,
equivalent to at least NT$1.39 billion in tourism output, a value three times
that of the loss of tourist facilities. This loss will increase as the schedule of
the reconstruction period is extended.
In the past, estimates of typhoon disaster loss included only the costs of
buildings, facilities, roads and environmental remediation but not the scale of
decreased tourist activity. This study found that this decrease is rather
significant. The exclusion of the opportunity cost of tourism activities seriously
underestimates the disaster’s impact on tourist spots, preventing them from
acquiring adequate disaster prevention resources. Delays in reconstruction will
aggravate the loss of tourism. However, our study also found that several tourist
spots were already losing attractiveness before the disaster. The reconstruction
of these tourist spots should be suspended and restarted only after proper long-
term planning to avoid the consumption of resources that could be used for
the reconstruction of other spots and prevent the reconstruction of facilities that
may not meet the needs of the tourism market. The most urgent task for the
authorities is to find a solution that shortens the reconstruction schedule and
thus reduces tourism losses while taking into account the long-term needs of
tourist activities. The results of this study can be used as a reference for the
allocation of post-Morakot reconstruction resources.
Endnotes
1. Because the public infrastructure, particularly the roads, bridges, water supplies, power facilities
and telecommunication were severely damaged, even residents living within the park area had
to be rehoused elsewhere for months. Tourism also totally ceased for a couple of months.
2. Accommodation, food and beverage services, passenger rail and bus transport, passenger air
transport, vehicle rental, travel agency services, and recreation and entertainment are listed as
tourism industries in Taiwan tourism satellite account.
References
Amelung, B., Nicholls, S., and Viner, D. (2007), ‘Implications of global climate change for tourism
flows and seasonality’, Journal of Travel Research, Vol 45, No 3, pp 285–296.
155Impact of meteorological disasters on tourism
Bonham, C., Edmonds, C., and Mak, J. (2006), ‘The impact of 9/11 and other terrible global events
on tourism in the United States and Hawaii’, Journal of Travel Research, Vol 45, No 1, pp 99–
110.
Brugman, M., Raistrick, P., and Pietroniro, A. (1997), ‘Glacier-related impacts of doubling
atmospheric carbon dioxide concentrations on British Columbia and Yukon’, in Taylor, E., and
Taylor, B., eds, Canada Country Study: Climate Impacts and Adaptation – British Columbia and Yukon,
Environment Canada, Ottawa, pp 6-1–6-9.
Conrady, R., and Bakan, S. (2008), ‘Climate change and its impact on the tourism industry trends
and issues in global tourism 2008’, in Conrady, R., and Buck, M., eds, Springer, Heidelberg,
pp 27–40.
Crouch, G.I. (1995), ‘A meta-analysis of tourism demand’, Annals of Tourism Research, Vol 22, No 1,
pp 103–118.
Cumming, S.G., and Burton, P.J. (1996), ‘Phenology-mediated effects of climatic change on some
simulated British Columbia forests’, Climatic Change, Vol 34, No 2, pp 213–222.
Directorate General of Budget, Accounting and Statistics (2010), National Accounts, Directorate
General of Budget, Accounting and Statistics, Executive Yuan, ROC.
Engle, R.F., and Granger, C.W.J. (1987), ‘Co-integration and error correction: representation,
estimation and testing’, Econometrica, Vol 55, No 2, pp 251–276.
Eugenio-Martin, J.L., and Campos-Soria, J.A. (2010), ‘Climate in the region of origin and
destination choice in outbound tourism demand’, Tourism Management, Vol 31, No 6, pp 744–
753.
Gómez-Martín, M.B. (2005), ‘Weather, climate and tourism a geographical perspective’, Annals of
Tourism Research, Vol 32, No 3, pp 571–591.
Granger, C.W.J. (2007), ‘Spurious regressions in econometrics’, in Baltagi, B.H., ed., A Companion
to Theoretical Econometrics, Blackwell, London, pp 557–561.
Granger, C.W.J., and Newbold, P. (1974), ‘Spurious regressions in econometrics’, Journal of Econometrics,
Vol 2, No 2, pp 111–120.
Gut, P., and Jarrell, S. (2007), ‘Silver lining on a dark cloud: the impact of 9/11 on a regional tourist
destination’, Journal of Travel Research, Vol 46, No 2, pp 147–153.
Hall, M.H.P., and Fagre, D.B. (2003), ‘Modeled climate-induced glacier change in Glacier national
park, 1850–2100’, BioScience, Vol 53, No 2, pp 131–140.
Halpin, P.N. (1994), ‘Latitudinal variation in montane ecosystem response to potential climatic
change’, in Beniston, M., ed, Mountain Ecosystems in Changing Climates, Routledge, London,
pp 180–203.
Hamilton, J., and Tol, R. (2007), ‘The impact of climate change on tourism in Germany, the UK
and Ireland: a simulation study’, Regional Environmental Change, Vol 7, No 3, pp 161–172.
Huang, J.-H., and Min, J.C.H. (2002), ‘Earthquake devastation and recovery in tourism: the Taiwan
case’, Tourism Management, Vol 23, No 2, pp 145–154.
Kuo, H.-I., Chen, C.-C., Tseng, W.-C., Ju, L.-F., and Huang, B.-W. (2008), ‘Assessing impacts of
SARS and Avian flu on international tourism demand to Asia’, Tourism Management, Vol 29, No 5,
pp 917–928.
Morakot Post-Disaster Reconstruction Council, Executive Yuan (2010), Morakot Post-Disaster
Reconstruction Projects, Morakot Post-Disaster Reconstruction Council, Executive Yuan,
Taipei.
Prideaux, B. (2006), ‘The use of scenarios to project the impact of global warming on future
visitation to the Great Barrier Reef’, in ATLAS Asia Pacific 2006 Conference – Tourism After Oil,
Dunedin, Department of Tourism, University of Otago.
Rosselló-Nadal, J., Riera-Font, A., and Cárdenas, V. (2011), ‘The impact of weather variability on
British outbound flows’, Climatic Change, Vol 105, No 1, pp 281–292.
Scott, D. (2011), ‘Why sustainable tourism must address climate change’, Journal of Sustainable
Tourism, Vol 19, No 1, pp 17–34.
Scott, D., Jones, B., and Konopek, J. (2007), ‘Implications of climate and environmental change
for nature-based tourism in the Canadian Rocky Mountains: a case study of Waterton Lakes
National Park’, Tourism Management, Vol 28, No 2, pp 570–579.
Scott, D., Malcolm, J.R., and Lemieux, C. (2002), ‘Climate change and modelled biome represen-
tation in Canada’s national park system: implications for system planning and park mandates’,
Global Ecology and Biogeography, Vol 11, No 6, pp 475–484.
Song, H., and Li, G. (2008), ‘Tourism demand modelling and forecasting – a review of recent
research’, Tourism Management, Vol 29, No 2, pp 203–220.
T
OURISM
E
CONOMICS
156
Song, H., and Lin, S. (2010), ‘Impacts of the financial and economic crisis on tourism in Asia’,
Journal of Travel Research, Vol 49, No 1, pp 16–30.
Song, H., Witt, S.F., and Li, G. (2008), The Advanced Econometrics of Tourism Demand, Routledge,
New York.
Taylor, T., and Ortiz, R.A. (2009), ‘Impacts of climate change on domestic tourism in the UK:
a panel data estimation’, Tourism Economics, Vol 15, No 4, pp 803–812.
Tourism Bureau (2010), Annual Statistical Report on Tourism, Tourism Bureau, MOTC, ROC.
Tourism Bureau (2011), 2010 Survey of Travel by R.O.C Citizens, Tourism Bureau, MOTC, ROC.
UNWTO-UNEP-WMO (2008), Climate Change and Tourism – Responding to Global Challenges, World
Tourism Organization, Madrid.
