Forecasting foreign tourists in Thailand by economic condition for tourism index

http://www.iaeme.com/IJMET/index.asp 144 editor@iaeme.com

International Journal of Mechanical Engineering and Technology (IJMET)

Volume 10, Issue 03, March 2019, pp. 144-152. Article ID: IJMET_10_03_014

Available online at http://www.iaeme.com/ijmet/issues.asp?JType=IJMET&VType=10&IType=3

ISSN Print: 0976-6340 and ISSN Online: 0976-6359

FORECASTING FOREIGN TOURISTS IN

THAILAND BY ECONOMIC CONDITION FOR

TOURISM INDEX

Bundit Chaivichayachat

Department of Economics, Faculty of Economics, Kasetsart University

Bangkok, Thailand

ABSTRACT

Tourism sector in Thailand has been promoted as a key sector to create economic

expansion since the early of 2010s. The increasing in demand for tourism in Thailand

induces a significant final demand in tourism related sectors. In order to prepare for

the growing in tourism sector, the number of visitors should be forecasted. This paper

set up a new composite index called economic condition for tourism (ECT) index. This

index summarizes the information of the macroeconomic variables which effect the

demand for tourism in Thailand. The methods of composite index are implemented.

There three weighted average methods that were used: simple average, variance

decomposition and factor important. Then, the ECT index can be used as a leading

indicator for number of foreign tourists. The ECT index also act as the important

factor forecast the number of foreign tourists. Tourism price in Thailand, tourism

price in AEC and crime rate are the most important economic factor to determine the

ECT index and for the forecasting the number of foreign tourists. The results emphasis

that to support the tourism promoting, price stability policy and crime reducing policy

should be considered.

Key words: Tourists, Leading Indicator, Transfer Function,

Cite this Article: Bundit Chaivichayachat, Forecasting Foreign Tourists in Thailand

by Economic Condition for Tourism Index, International Journal of Mechanical

Engineering and Technology, 10(3), 2019, pp. 144-152.

http://www.iaeme.com/IJMET/issues.asp?JType=IJMET&VType=10&IType=3

1. INTRODUCTION

In Thailand has been promoted as a key sector to create economic expansion since the early of

2010s. Therefore, tourism promoting policies both in form of qualitative and quantitative

measures have been implemented. As a result, the demand of Thai‟s tourism increases

continuously. Not only the policy to promoting the tourism sector, the demand for Thai

Tourism also response to the macroeconomic variables including exchange rate, inflation,

tourist‟s income, securities and political condition (Kara et al., 2005; Alvarez, 2007; Allen

Bundit Chaivichayachat

http://www.iaeme.com/IJMET/index.asp 145 editor@iaeme.com

and Yop, 2009; Onder et al., 2009; Monebi and Rahim, 2010; Song and Wei, 2010;

HanafionHarun and Jamaluddin, 2011; Ibranim, 2011; Skuflic and Stokovic, 2011; Betonio,

2013; Altindag, 2013; Bentum-Ennin, 2014; Deluna and Jeon, 2014; Laframboise et al., 2014;

Moorthy, 2014; Bundit, 2017; and Bundit, 2018). The increasing in demand for tourism in

Thailand induces a significant final demand in tourism related sectors. Finally, the linkages

between the tourism sector and the tourism related sectors initiate an economic expansion for

Thai economy. In order to prepare for the growing in tourism sector, the number of visitors

should be forecasted. This paper aims to forecast the number of foreign tourists by used the

economic condition for tourism index (ECT) and the transfer function. The results can be used

to setup the policy to support the expansion of the tourism sector in Thailand

2. METHODOLOGY

The number of foreign tourists, the research method are the composite index and transfer

function. First, the composite index named economic condition for tourism index (ECT) will

be calculated. This index summarized all economic condition that effect the demand for

tourism in Thailand by foreign tourists. The increasing ECT will be followed by the

increasing in tourism demand and tourism revenues. In contrast, the declining in ECT will be

followed by the slowdown in tourism demand and tourism revenues. Then, after the ECT

index was calculated, the transfer function is employed to estimate a linear equation to

forecast the number of foreign tourists. For the first step, the set of economic variables the

determined the demand for tourism will be defined following the demand theory, maximized

behavior and the recent empirical works. The economic variables that determined demand for

tourism by foreign tourists are tourist‟s income (million Baht: YM), tourism price (consumer

price index in Thailand: PT), exchange rate (Bath: US dollar: NE), market share of retail trade

sector (ratio of GDP in retail trade to total GDP: RT), tourism promoting policy (million baht:

TB), number of hotels and guest houses (number: HG), number of hospital (hospital approved

by Ministry of Health: HS), crime rate (time: CR), and economic, tourism price in AEC

(average consumer price index in AEC countries: PO) and political condition (dummy

variable given 1 for instability occurred and 0 for another: PS). These economic variables are

normalized to cancel the difference in measurement unit. Following UNDP (2013), the

economic variables which generate the positive effect on demand for Thai tourism can be

normalized as:

iii

P - Min (P )

Y Max (P ) - Min (P )



(1)

where

is the positive economic variable on tourism including YM, RT, TB, NE, HG,

HS, and PO,

is normalized of positive economic variables. The normalized variable value

is lie between 0 and 1.

For the negative factors, they can be normalized as

iii

Max (N ) - N

Y Max (N ) - Min (N )



(2)

Where

is the negative economic variable on tourism including PT and PS.

The normalized variable values, both

and

, are lie between 0 and 1.

Once the normalized variables are calculated, the method to setup a composite index is

implemented for calculate the ECT index. There are many methods for setup the composite

index. For example, WTTC (2013) employed simple average. Gooroochurn and Sugiyarto

(2005) and Fernandez and Rivero (2009) used factor analysis to calculate weight of each

Forecasting Foreign Tourists in Thailand by Economic Condition for Tourism Index

http://www.iaeme.com/IJMET/index.asp 146 editor@iaeme.com

economic variable. Freedman (1994, 1996), Lack (2003), Goodhart and Hofmann (2001),

Chong (2014) and Chang, Hsu and McAleer (2014) estimated weight of each normalized

factor by calculated variance decomposition from vector autoregressive (VAR). This paper

will used for 3 methods to calculate the ECT index. The first method is the simple average

method. The index can be setup as

i 1

I k





(3)

where

is the economic condition for tourism index by simple average, i = 1, 2, …, k

and k is the number of normalized economic factor.

Second, ECT index will be calculated by estimated the variance decomposition in vector

autoregressive (VAR). It is the same methods as the composite index for financial condition

index (FCI) and monetary condition index (MCI). The variance decomposition equation

presented as

2 2 2 2 2 2 2

FTN 1 2 k

Y Y Y

1 2 k

(L) (L) ... (L)          

(4)

where

FTN



is variance of the forecast error for the number of foreign tourists,



is the

variance of the jth economic factors and

j(L)

is lag operator in the impulse response function

from vector autoregressive.

For the third way, this paper proposes new method for the calculation for the weight of

economic factors based on artificial neural network (ANN) for forecast the number of foreign

tourists. The factor important measures the importance of ith factor on the forecasting of

foreign tourist. The summation of factor important overall economic factors equal one. (see

Bundit,2015 and Bundit,2017).

Once the ECT index is calculated, the next step is the estimation of the transfer function.

(Kulendran and Witt, 2003 and Santos and Macedo, 2008). The basic idea of the transfer

function is autoregressive integrated moving average (ARIMA) which including 2 parts, AR

term and MA term. In the transfer function, the ECT is invited explicitly to play a role for

forecast tourist number. Therefore, there are three parts; moving average (MA),

autoregressive (AR) and the ECT index. The transfer function can be presented as

t t-b t

(B)FTN (B)ECT    

(5)

where FTN is output series or number of foreign tourists, ECT is input series or the

economic condition for tourism index, B is lag operator where

kt t-k

B FTN FTN



is r

whitenoise residual process,

1 2 r

(B) 1 - B - B -...- B    

(B)

(B)  

and



whitenoise term.

3. RESULTS AND DICUSSION

Thailand, National Statistic Office of the National Economic and Social Development

Board, were arranged. Then, the normalized economic factors which used for organized the

economic condition for tourism index shown in Table 1. Then, the weight for each factor

which estimated by simple average, variance decomposition and factor important are

presented in Table 2. Based on Table 2, the economic condition for tourism index following 3

methods are shown in Figure 1 and Table 3. The results indicate that the ECT index by 3

methods move at the same pattern. In comparing with normalized number of foreign tourists,

Bundit Chaivichayachat

http://www.iaeme.com/IJMET/index.asp 147 editor@iaeme.com

the ECT index move very close to the normalized number of foreign tourists especially in

2013 and 2014. The correlation between the ECT by simple average, variance decomposition

and factor important are 0.931, 0.891 and 0.931 respectively. Moreover, before employ the

ECT as a determinant for forecasting number of foreign tourist in transfer function, the

Granger causality test was performed. The results by Granger causality represent the causality

between normalized number of foreign tourists and the ECT index. The Granger causality test

shown in Table 4. Along the null hypotheses, only one null hypothesis can be rejected. The F

statistic indicates the null hypothesis, ECT_VDC does not Granger causality, was rejected

with statistical significant. Only the economic condition for tourism index by variance causes

normalized number of foreign tourists with statistically significance. Therefore, in the next

step, the ECT index by variance decomposition will enter to the transfer function for

forecasting number of foreign tourists. Moreover, the weight of variance decomposition

shown that tourism cost, tourism cost in AEC and crime rate are the most important

macroeconomic variables to determine demand for Thai‟s tourism.

Next, the ECT index weighted by variance decomposition will enter to the transfer function a

written in equation (5). At first, we proposed lag length equal to 4 for quarterly data for all term in the

transfer function (autoregressive, moving average and economic condition index) and developed the

estimation results by General-to-Specific method. Finally, the best estimated transfer function

presented in Table 5. The ECT index lag 1 and lag 3 determine number of foreign tourists with

statistically significance. Then, the ECT index weighted by variance decomposition can be used as

leading indicator for number of foreign tourists. (Kulendran and Witt, 2003; Santos and Macedo,

2008; Fernandez and Rivero, 2009 and Chang, Hsu and McAleer, 2014).

Table 1: Normalized Economic Factors

Table 2: Weight of Economic Factor

YM NE PT RT PO TB HG HS CR PS

2008 Q1 0.0083 0.0136 0.0904 0.0788 0.0178 0.0002 0.0000 0.0557 0.0000 0.0360

Q2 0.0087 0.0130 0.1190 0.1050 0.0200 0.0000 0.0011 0.0522 0.0007 0.0360

Q3 0.0085 0.0117 0.1162 0.1174 0.0210 0.0002 0.0023 0.0493 0.0011 0.0360

Q4 0.0075 0.0101 0.0570 0.1198 0.0160 0.0009 0.0035 0.0470 0.0003 0.0360

2009 Q1 0.0000 0.0078 0.0297 0.1035 0.0075 0.0045 0.0048 0.0431 0.0012 0.0360

Q2 0.0000 0.0157 0.0000 0.1039 0.0029 0.0052 0.0061 0.0428 0.0009 0.0360

Q3 0.0017 0.0261 0.0071 0.0942 0.0000 0.0054 0.0075 0.0440 0.0011 0.0360

Q4 0.0050 0.0274 0.0546 0.0823 0.0027 0.0052 0.0089 0.0467 0.0019 0.0360

2010 Q1 0.0152 0.0365 0.0753 0.0762 0.0063 0.0001 0.0104 0.0572 0.0008 0.0360

Q2 0.0197 0.0263 0.0694 0.0976 0.0073 0.0008 0.0119 0.0602 0.0010 0.0000

Q3 0.0237 0.0222 0.0700 0.1096 0.0078 0.0027 0.0135 0.0621 0.0028 0.0000

Q4 0.0273 0.0219 0.0658 0.1078 0.0085 0.0059 0.0151 0.0629 0.0034 0.0360

2011 Q1 0.0297 0.0202 0.0672 0.0995 0.0094 0.0149 0.0168 0.0660 0.0037 0.0360

Q2 0.0326 0.0000 0.0797 0.1341 0.0118 0.0190 0.0185 0.0633 0.0018 0.0000

Q3 0.0354 0.0274 0.0801 0.1370 0.0109 0.0227 0.0203 0.0581 0.0035 0.0000

Q4 0.0381 0.0271 0.0782 0.1202 0.0096 0.0259 0.0221 0.0504 0.0076 0.0000

2012 Q1 0.0400 0.0328 0.0715 0.0000 0.0073 0.0275 0.0240 0.0303 0.0043 0.0000

Q2 0.0426 0.0510 0.0614 0.0131 0.0042 0.0303 0.0259 0.0218 0.0054 0.0360

Q3 0.0452 0.0570 0.0662 0.0149 0.0042 0.0331 0.0279 0.0148 0.0050 0.0360

Q4 0.0480 0.0423 0.0696 0.0106 0.0056 0.0360 0.0299 0.0094 0.0068 0.0000

2013 Q1 0.0510 0.0386 0.0681 0.0087 0.0060 0.0394 0.0320 0.0002 0.0073 0.0000

Q2 0.0538 0.0317 0.0590 0.0238 0.0063 0.0421 0.0341 0.0000 0.0074 0.0000

Q3 0.0565 0.0316 0.0516 0.0314 0.0071 0.0447 0.0363 0.0034 0.0080 0.0360

Q4 0.0591 0.0385 0.0517 0.0223 0.0064 0.0471 0.0386 0.0105 0.0061 0.0360

2014 Q1 0.0620 0.0474 0.0554 0.0090 0.0062 0.0538 0.0408 0.0368 0.0045 0.0000

Q2 0.0645 0.0316 0.0609 0.0249 0.0059 0.0540 0.0432 0.0450 0.0053 0.0000

Q3 0.0668 0.0385 0.0554 0.0343 0.0049 0.0522 0.0456 0.0505 0.0048 0.0000

Q4 0.0690 0.0474 0.0451 0.0324 0.0045 0.0485 0.0480 0.0535 0.0040 0.0000

Forecasting Foreign Tourists in Thailand by Economic Condition for Tourism Index

http://www.iaeme.com/IJMET/index.asp 148 editor@iaeme.com

Figure 1: Economic Condition for Tourism Index and Number of Foreign Tourist (normalized)

Table 3: Economic Condition for Tourism Index

Simple Average Variance Decomposition Factor Important

Tourist's Income YM 0.1000 0.0862 0.1070

Exchange Rate NE 0.1000 0.0718 0.0970

Tourism Price PT 0.1000 0.2718 0.1570

Size of Retail Trade Sector RT 0.1000 0.1077 0.1770

Tourism Price in AEC PO 0.1000 0.2030 0.0570

Tourism Promoting Budget TB 0.1000 0.0057 0.0870

Hotel and Guest House HG 0.1000 0.0063 0.0870

Hospital HS 0.1000 0.0444 0.1070

Crime Rate CR 0.1000 0.1627 0.0470

Economic and Political Instablity PS 0.1000 0.0404 0.0770

1.0000 1.0000 1.0000

Total

Forecasting foreign tourists in Thailand by economic condition for tourism index

This paper set up a new composite index called economic condition for tourism (ECT) index. This index summarizes the information of the macroeconomic variables which effect the demand for tourism in Thailand.

Có thể bạn quan tâm