پناهی, حسین, خداوردیزاده, صابر. (1394). تأثیر غیرخطی تورم و توسعهی گردشگری بر رشد اقتصادی ایران: رهیافت مارکوف-سویچینگ. مجله برنامه ریزی و توسعه گردشگری, 4(14), 8-25.

حسین پناهی; صابر خداوردیزاده. "تأثیر غیرخطی تورم و توسعهی گردشگری بر رشد اقتصادی ایران: رهیافت مارکوف-سویچینگ". مجله برنامه ریزی و توسعه گردشگری, 4, 14, 1394, 8-25.

پناهی, حسین, خداوردیزاده, صابر. (1394). 'تأثیر غیرخطی تورم و توسعهی گردشگری بر رشد اقتصادی ایران: رهیافت مارکوف-سویچینگ', مجله برنامه ریزی و توسعه گردشگری, 4(14), pp. 8-25.

پناهی, حسین, خداوردیزاده, صابر. تأثیر غیرخطی تورم و توسعهی گردشگری بر رشد اقتصادی ایران: رهیافت مارکوف-سویچینگ. مجله برنامه ریزی و توسعه گردشگری, 1394; 4(14): 8-25.

تأثیر غیرخطی تورم و توسعهی گردشگری بر رشد اقتصادی ایران: رهیافت مارکوف-سویچینگ

^{2}دانشجوی دکترای اقتصاد بینالملل، دانشگاه تبریز

چکیده

دستیابی به نرخ رشد اقتصادی بالا و باثبات ازجمله مسائل مهم هر کشور است. از طرف دیگر تورم و آثار زیانبار آن (بهویژه بر رشد اقتصادی) نیز یکی از مشکلات اساسی کشورها به حساب میآید. امروزه توسعهی گردشگری در تمامی عرصهها، چه در سطح ملی و منطقهای و چه در سطح بینالمللی موردتوجه برنامهریزان دولتی و شرکتهای خصوصی قرار گرفته است. آگاهی جوامع از اینکه گردشگری منبع درآمدی ارزی بسیار مناسب و قابل ملاحظهای در اختیار اقتصاد یک کشور قرار میدهد، باعث شده است که گردشگری مفهوم بسیار گستردهای در ابعاد مختلف اقتصادی، اجتماعی و فرهنگی پیدا کرده و به عنوان یک صنعت تلقی شود. به این منظور، مقاله حاضر با استفاده از رویکرد غیرخطی مارکوف- سویچینگ به بررسی تأثیر درآمدهای گردشگری و تورم بر رشد اقتصادی ایران به صورت فصلی و طی دورهی زمانی 1391-1374 پرداخته است. نتایج حاکی از وجود رابطه مثبت و معنیدار درآمدهای حاصل از گردشگری بر رشد اقتصادی در هر سه رژیم صفر، یک و دو بوده است. بهطوریکه در رژیم صفر بیشترین تأثیر و در رژیم دو کمترین تأثیر را بر رشد اقتصادی گذاشته است. همچنین سرمایهگذاری و تورم به ترتیب تأثیر مثبت و منفی بر رشد اقتصادی گذاشته است.

Tourism Development and Economic Growth
Markov Switching Model

نویسندگان [English]

Hossein Panahi^{1}؛ Saber Khodaverdizadeh^{2}

^{1}Associated Professor, University of Tabriz

^{2}PhD Student of Economics, University of Tabriz

چکیده [English]

Extended Abstract Nowadays tourism industry is the world’s third profitable industry after the oil and the automotive industries. Moreover it is anticipated to become the first in the world by 2020. Here in from like other developing countries, despite a high potential for tourism, it has not benefited from tourism-related opportunities. However, tourism industry, as the main source of currency for the country, plays a major role in economic growth and solving problems including youth employment and also monetary and financial problems. Given the successful experience of numerous countries its tremendous impact on the economic development of countries we used nonlinear Markov- Switching models to investigate the influence of seasonal tourism on economic growth of Iran (1995-2010). Findings of the present study suggested a positive significant relationship between tourism revenues and economic growth (in Iran). As well, investment and inflation have respectively positive and negative impact on economic growth. Introduction World economy is very heterogeneous and significant. Nevertheless, economic development is the primary goal of the economic policy because their development of natural resources can be accomplished with the process of the economic growth and development. Tourism as the main source of industry for the country is one of the sectors that can help countries in addressing quickly this menace especially the ones that have great potential to attract international tourists. According to the forecast by United Nations World Tourism Organization (UNWTO) although currently tourism industry is the world’s third profitable industry after the oil and automotive industries, it is going to be ranked first by 2020. Creating new regional job opportunities, development of transportation, increased growth of the financial sector in the economy involving foreign currency, domestic sale of goods and services, introducing the country to the other people, exporting culture to other countries and its impact on international relations and foreign investments are examples of benefits supporting the lucrative tourism section. Iran has been named among the world top 10 tourism destinations with several thousand years of civilization and unique tourist culture, natural, historical and religious attractions. However, capacity utilization is quite low. On the other hand this country is working hard to reform its petroleum-based economy over the past few decades. So, knowing as the Tourism-Led Growth, tourism development and the benefits is the logical solution for solving the country’s economic problems. In the recent years analyzing the importance of tourism and its effects on other macroeconomic variables are considered among the challenging issues. So that several studies have acknowledged the prominent role of this industry on the balance of payments, tax revenues and unemployment (Betisle & Hoy, 1980, West, 1993, Durbarry, 2002, Cheong, Seng, Khan, 1990).According to Oh (2005) it is essential to verify the validity of the tourism led growth hypothesis in host countries. Hence, in the present study we tried to investigate the contributions of tourism to economic growth seasonally in the period of 1995-2012 for Iran using a nonlinear method. The remainder of this paper is structured as follows. Section 2 is a relevant literature review. Section 3 explains the methodology of this study and Section 4 presents model estimation and the results of the analysis. Finally, Section 5 summarizes the paper and concludes it. Methodology Markov- Switching model was introduced for the first time by Quandt (1972) and Quandt- Goldfold (1973) then was developed by Hamilton (1989) used in business cycle analysis. Unlike other nonlinear models such as STAR and ANN that enable gradual transition between two or more regimes. Markov- Switching model implies sudden switching. This model is also different from a model of structural change and time of change- point occurs in any number. While in structural change model just exogenous transition at a specific time are available. So Markov- Switching model performs well for describing the data that exhibit different behaviour patterns in different periods of time. One major advantage of this model is its flexibility namely permanent and temporary changes are available and these changes can occur frequently in a short period of time. Yet this model determines endogenously the exact time of the changes and structural failures. Capabilities of Markov- Switching model in predicting the behaviour of economic variables have led to its widespread use in the economy (Fallahi, Hashemi, 2010). In Markov- Switching model it is assumed that the regime whish happens in t is not observable and is dependent on an invisible process ). In a model with 2 regimes , we can easily assume that gets the value of 1 and 2. An AR(1) with 2 regimes can be written like this:

Or

To complete the model, we must specify the characteristics of . In Markov Switching Model, is considered a Markov process of first degree. This assumption says ) is just dependent on . The following section introduces the transition probabilities from one state to another to complete the Model. (4)

Where represents the probability of the Markov chain moves from state i at time t-1 to state j at time t. must be non-negative and the following conditions are met: , (5) Model introduced above can be generalized to the case involving m regimes and P gaps. In other words, it is an AR(p) process and has values of 2, 1, ..., m. in this case, depending on which of these equations is dependent to status variable, some general states happens: Table 1-1: States of Markov Switching Model

By combining the first and second modes with the second and third modes we can obtain more detailed models where there is a possibility of the various components of the equation of the regimes. Table (3-2) summarizes the different states of the Markov-switching model.

Table 1-2: Summary of Different Modes of MS-AR Model

In this paper Markov- Switching approach will be used to study international tourism real receipts growth on Iran’s economic growth. Therefor the main model of this research has been modified version of Wan-Chen Po, Bwo-Nung Huang (2008) which is: (6) Where dependent variable is (GY) and independent ones are tourism revenue (TR) gross formation of capital (GFC) and inflation rate. Statistics have been collected from the world development indicators seasonally and for the period of 1995-2012. Model Estimation and Results Analysis Evaluation of reliability of the variables is the first step for estimating the time series. Reliability was assessed using augmented dickey-fuller (ADF) and the results are presented in table (1-3) and all of the variables with stationary one-time difference is folded in the model. Table 1-3: The Results of Unit root ADF Test: Intercept and Procedure Apart from this model, acceptability of a criterion is that the features of data should be nonlinear. An LR test was used for the model to ensure nonlinearity of the data. The statistics were calculated using maximum likelihood estimation of two competition models: a model with one regime (a linear model) and the other with two regimes (nonlinear model). It had a chi-square distribution. If the value of the statistic is greater than the critical value, it can be concluded that the linear model was inadequate for this confidence level. So the using the nonlinear model will be suitable. The LR test is shown in table 1 Table 1: The Results of the LR Test

As indicated in table 1 the observed value of the test statistic is significantly (5%) higher than critical value. Therefore it can be assumed that it is better to use Markov- Switching nonlinear methods instead of linear one for estimating the model parameters. After determining optimized intervals the number of regimes are introduced using Akaike criterion. Mont Carlo simulation results indicated that comparing to the likelihood function method, Akaike criterion is more appropriate indicator for determining the number of regimes. Table 2 shows the Akaike information statistics summary for the number of the regimes II and III. Table 2: Using Akaike Criterion Method for Determining the Number of Regimes

Table 2 identify 3 regimes as the optimal number. As mentioned in the methodology chapter, Markov- Switching method has served features that each one is a component of a regime- dependent equation. Then to select the best one, we may use the maximum likelihood of them. The model with maximum likelihood reflects the optimal model. Characteristics of each regime are as indicated in Table 3. The first and second column shows respectively the number of observations and the presence probability for each observation. For example the probability that randomly selected observation be in zero regime is about 30.65%. The third column of the table indicates the mean of successive observations in the regime studied. In other words, by traverse stock price index from one-regime towards zero-regime, about 1.58 duration will remain in the regime. Table 3: Characteristics of Each Regimes

The probability of each of the years over each of the regimes is the presented in figure 1. As seen in this figure, sum of these probabilities is one. That is the year considered can be in the regimes I, II or III.

The transition probability from one regime to another is presented in Table 4 that shows stability and instability among regimes. Table 4: The Transition Probability from One Regime to Another

The transition probability over each of the regimes is found to be significantly high (96%). Then this regime is more stable than the other two ones. As noted in the chapter on introducing the model, the error terms of Marko- Switching must be normal and should be tried to overcome autocorrelation and heteroskedasticity in it. The results of these characteristics are listed below. Table 5: Tests for Error Term's Normality

Regarding findings of the autocorrelation analysis (significance of 5%) an absence of autocorrelation cannot be rejected. Thus it indicates no autocorrelation in the error terms. A normality test suggests normality for distribution in estimated error terms. The heteroskedasticity test results revealed that the variance of the error terms was consistent. Estimation of the Model:

Results The results of the model estimation reported in the table below, indicates that the intercept in the regimes zero, one and two were respectively -6.24, 7.59 and 21.02. Interruption in economic growth across all three regimes will have significant positive impact on current economic development with greatest impact on the regime of zero. So that one percent increase in TR will negatively affect economic growth in the regimes of zero and two, while its influence in the regime one is significantly positive. But one percentage transition in INF of each of these three regimes would lead to negative impact on economic growth. GFC interruption is the regimes zero and two and in regime one have respectively negative and positive effects on economic development. Conclusion and Policy recommendations Achieving high and stable economic development requires a great understanding of its effective factors. Among various factors affecting economic growth of a country, tourism industry has successfully influenced some of the countries. In this contest the main objective of the present study is to assess the effect of quarterly revenues of tourism industry in the economic development during 1995 -2012. Then nonlinear Markov- Switching method was used to assess the impact of tourism and other explanatory variables on economic growth in terms of research model studies. Findings suggested positive effect of this industry on economic developments. The results of estimating the model shows that an increase in tourism development will lead to increase in per capita GDP, too. The spending of international tourists is economically export income. So the influence of this industry is widespread. Because help to transfer foreign exchange to country by the multiplier coefficients and thereby simulate the economy. So tourism industry is a major source of income for many countries and promoting this industry will stimulate the economy via foreign currencies and increase the country's income. Overall the results suggest that common sources of growth including investment in physical and human capital and household consumption expenditure develops the economy. In other words, 1% tourism growth will lead to 4% economic development. Keywords: Tourism, Economic Growth, Markov-Switching References

Belisle, F. and Hoy, D. (1980). The perceived impact of tourism by residents, Annals of Tourism Research, 8: 83–97. Butcher, G., Fairweather, J.R. and Simmons, D.G. (2003). The economic impact of tourism on Christchurch city and Akaroa Township, Tourism Research and Education Center, 37:135-143. Chen Po, W. and Nung Hung, B. (2008). Tourism development and economic growth–a nonlinear approach, Physica A: Statistical Mechanics and its Applications, 387: 5535-5542. Chor, F.T. and EU C.T. (2015). Does tourism effectively stimulate Malaysia's economic growth, Tourism Management, 17:158-163. Durbarry, R. (2002). The economic contribution of tourism in Mauritius, Annals of Tourism Research, 29: 862–865. Espinoza, R.A., Prasad, A. and Leon, H.L. (2010). Estimating the Inflation Growth the Nexus Smooth Transition Model IMF, Working Paper. Fallahi,F. and Hashemi, A. (2010). The Casuality Between GDP and Energy Consumptions in Iran using Markov Switching Models, Energy Economies Studies, 7(26):131-152 (In Persian) Gokal, V. and Hanif, S. (2004). Relationship between Inflation and Economic Growth, Working Paper. Hamilton. J.D. (1989). A new approach to the economic analysis of nonstationary time series and the business cycle, Econometrica, 57: 357–384. Kazerooni, A., Khodaverdizadeh, S. and Shomali, A. (2014). Examination of tourism development Effect on the economic welfare of the eight D; static panel approach, The First Conference of Outlook on Torbat Heidariyeh Region in 1404. (In Persian) Khan, H., Seng, C. and Cheong, W. (1990). Tourism multipliers effects on Singapore, Annals of Tourism Research, 17: 408–418. Krolzig, H.M. (1997). Markov-Switching Vector Autoregressions Modelling, Statistical Inference and Applications to Business Cycle Analysis, Springer, Berlin. Kuan-Yu, C. (2011) .Combining linear and nonlinear model in forecasting tourism demand, Expert Systems with Applications, 38: 10368–10376. Lee, C.C. and Chien, M.S. (2008). Structural breaks, tourism development, and growth: Evidence from Taiwan, Mathematics and Computers in Simulation, 77: 358–368. Lee, C.C. and Chine, M. S. (2008) Tourism development in economic growth, evidence from Taiwan, journal of policy Modeling, 25: 180–192. Leitato, N.C. (2011). Tourism, trade and growth: a panel data, Approach Actual Problems of Economics, 343-349. Lin, C-F.J. and Terasvirta, T. (1994). Testing the constancy of regression parameters against continuous structural change, Journal of Econometrics, 62: 211-228. Marin, D. (1992). Is the export-led hypothesis valid for industrialized countries? The Review of Economics and Statistics, 4: 678–688. Oh, C.-O. (2005). The contribution of tourism development to economic growth in the Korean economy, Tourism Management, 26: 39–44. Omay, T. and Kan, E.O. (2010). Re-examining the threshold effects in the inflation–growth nexus with cross sectionally dependent non-linear panel: Evidence from six industrialized economies, Economic Modelling, 27: 996-1005. Quandt. R.E. (1972). A new approach to estimating switching regressions, Journal of the American Statistical Association, 67: 306–310. West, G.R. (1993). Economic significance of tourism in Queensland, Annals of Tourism Research, 20: 490–504. Yavari, K., Reza Golizade, M., Aghaei, M. and Mostafavi, S.M.H. (2010). The impact of tourism spending on OIC economic growth, The Journal of Economic Studies, 45(2): 221-244. (In Persian)