HEDGING BY USING WEATHER DERIVATIVES IN WINTER SKI TOURISM HEDGING BY USING WEATHER DERIVATIVES IN WINTER SKI TOURISM

Tourism, as one of the main driving forces of economic development, is exposed to many risks. Besides frequent fluctuations in foreign currency exchange, prices of fuel and transportation, the tourism industry has become more sensitive to weather conditions lately. One of the new instruments which can be efficiently used for weather risk hedging is weather derivatives (forwards, futures, options and swaps on chosen weather variables temperature, rain, snow, wind etc.).


Introduction
The effects of climate changes and weather conditions on company's business results are increasing, and affecting a growing number of economic sectors. According to MünichRe Group (2000), over 80% of business activities on a global level are depending on the weather. Most affected sectors are energetics, agriculture, construction, transportation, and tourism (Bank, Wiesner, 2011;Zapranis, Alexandridis, 2013). The companies concerned are the ones that are highly sensitive to weather conditions and seasonality of their operations (Müller, Grandi, 2000a;2000b). Tourism, as an economic sector, is very sensitive to all kinds of changes, of both economic and noneconomic character (Becken, Hay, 2007;Agrawala, 2008). It is exactly the second type of change, the non-economic one, which is gaining importance lately (for example, EP 2018 (65) 1 (125-142) Bojan S. Đorđević safety risk, and climate change and weather risk). As emphasized by Simpson et al. (2008), besides increased exposure and sensitivity to various kinds of risks, modern tourism companies features higher fixed costs and lower profitability. Such structure makes them "slow" in accepting changes, vulnerable under the influence of different sets of risks and affecting the level of tourism service quality (Scott, McBoyle, 2007). Šperanda (2008) suggested that companies operating in tourism sector should eagerly acquire new knowledge and useful tools for risk management, such as, and currently available in our market, currency futures -forwards and swops, which can efficiently manage currency and interest risk.
Weather risk is a product of increasing climate changes, which effects we all feel. Extremely high or low temperatures, droughts and lack of rain, overflow of falls and floods, storms and hurricanes, definitely affect the business of the tourism sector companies. Business protection can be partly found at insurance companies, which cover only specific types of risk when paid compensation for damage in many cases compensates the partial loss. That is why, in order to protect or hedge the loss and reduction of product/service production caused by unfavourable weather conditions, weather derivatives are constructed, as a new tool for risk management (Russ, 2004). As Cao et al. (2004) and Brockett et al. (2005) pointed out, it is about weather risk which presents uncertainty in money flows and revenues as a consequence of non-catastrophic weather conditions (fluctuations in temperature, the quantity of precipitation -rain and snow, thespeed of wind etc.). Edrich (2003) and Ali (2000) made a clear distinction between weather risk and other sources of risk. What is vital here, is that weather risk affects the quantity of demand and sales of certain products/services and not the price of those goods and services. It is all about so-called volumetric risk. The weather condition risk is highly geographically localized and it is not possible to control it. Considering the exposure to weather risk and possibility for hedging, Leggio (2007) concluded that companies usually decide to use weather derivatives in order to lessen the volatility of revenues and money flows, to cover and compensate expenses (companies -hedgers), but also to stimulate and enhance sales (making profits and earningsspeculators) (Đorđević, Đorđević, 2014). It is definite that the main motive for trading these derivatives is hedging, but Zapranis and Alexandridis (2013) also emphasized the inevitable motive for profit and existence of speculators, as an important segment of weather derivatives market. Like other financial instruments, weather derivatives are traded in theorganized marketstock exchanges (standard contracts), but also out of the market (over the counter -OTC), in arrangement with traders -banks, insurance houses, hedge funds etc. (Marković, 2013;Đorđević,Đorđević, 2014). If the trade is in an organized market, then we are talking about futures, while forwards and swaps are usually traded in OTC. Weather options are present at both markets, and they usually contain chosen weather index in their bases. The biggest and most important weather derivative market today is American CME, while in Europe it used to be London market LIFFE, known today as ICE Europe.
Forward represents a simple contract which, at given time, arranges purchase and sale of certain assets at some future time point, where commitments are realized in the future, i.e. on a previously set date. One side (forward buyer) hedges from an eventual rise of assets price, while the other side (forward seller) protects from the fall of assets price (Hull, 2009;2011). Forwards are not standardized contracts, which mean that they are traded on OTC markets, with the mutual agreement of both sides on all elements of the contract (contract price, maturity date). Basic motive for forwards trading is protection (hedging) from markets' risk and final delivery of assets on the set date. These contracts are not liquid, meaning that they cannot be re-sold, thus they are not speculative (Đorđević, Đorđević, 2013; 2014). Compared to other derivatives, forwards have both advantages Bojan S. Đorđević and disadvantages. They are very flexible because they are adjusted toboth sides and have small transaction cost. Disadvantage refers to the absence of clearing house in the trade, so the risk of contract realization is growing. Also, if there is no margin (which is the case with futures), this type of contract exposes traders to credit risk (Golden et al., 2007). In case of weather derivative, forward would contain weather index in its base on certain weather variable (temperature, rain, snow, wind, frost etc.). On weather derivative market forwards usually appear in the form of swap contracts (Marković et al., 2012). Contractors would be obliged to set the elements for construction of weather forward precisely: weather index with variable, index unit value -tick, reference meteorology station -weather station, strike level, upper and lower limit -L lower and L upper , maximum pay off -maxPayOff, validity period, due date and margin, in order to lessen credit risk (for example 5-10% of contract value).
Today, weather derivatives represent a device in weather risk management, and their usage in the world is increasing, mostly in options trading. Since weather derivatives present general uncertainty in our region, apart from few works on the subject of their application in agriculture, this paper aims to show the possibility of hedging by applying chosen weather derivatives (snowfall forward) in winter ski tourism, on an example of ski lift operator in Kopaonik.
The paper is structured as follows: in section 2, we define the problem and the aims of our research. In section 3 we give an overview of most significant research results on the subject of weather risk management and derivatives application in winter ski tourism, ski resorts and ski lift operator business. In section 4 we present research methodology, data collection and results. In this section, werepresent analyses of hedge effectiveness when forwards are applied to snowfall index (weather snowfall forward) in ski lift operator business in Kopaonik mountain. Eventually, section 5 gives conclusions and recommendations for future research.

Defining research problem and aims
Ski season in Kopaonik ski resort lasts from December until the end of April (winter ski season). In that period ski resort, that is ski lift operator, makes amajor and most significant share of its incomes from ski tickets sales and ski lift operations. During a winter season, when the profits are the highest, the quantity of snowfalls and height of natural snow on ski tracks is of great importance, excluding artificial snow systems, which efficient work depends on the temperature (it takes from at least -2,5 o C up to -5 o C for snowmaking system to work efficiently). Due to climate changes, which in our case can be summed up to extreme weather (un)conditions (insufficient or overflowing snowfalls), the exposure of ski resorts and their income to weather risk is growing. In order to ensure their business and protect (hedge) from potential losses, ski operator can use weather derivative with weather index on snowfalls in their base. World markets (for example CME) offer monthly and seasonal weather derivatives (forwards, futures, options, and swaps), which appear in both organized (standardized contracts) and OTC markets. In weather-sensitive and seasonal industries, such as tourism, monthly and seasonal optional contracts on various weather variables (temperature, falls -rain, snow, frost etc.) are widely used. In winter tourism, dominant weather variable is snowfall quantity in ski season, although we shouldn't neglect temperature and wind, on whose index base ski resorts can also trade.
In this paper, we presumed that the quantity of snowfall highly affects incomes from ski tickets during the winter ski season at Kopaonik mountain. Ski lift operator (in our case JP Skijališta Srbije) wants to ensure (hedge) its business and ski ticket sales profits from weather risk of insufficient snowfall during the season (financial hedging), without taking into consideration existing artificial snowmaking system, so that operational hedging is not considered. Let us assume that weather derivatives on a quantity of snowfall are available at OTC market. Ski lift operator chooses weather forward, or seasonal snowfall index forward, to be more precise. Snowfall quantity data is gained from Kopaonik weather station, which is settled in Sunčana dolina (Sunny Valley) so that geographical basis risk is avoided, while the inclusion of basis risk will depend on a correlation of weather index with snowfall and ski ticket sales revenue. Basis risk will be included as the most important element of snowfall forward hedging effectiveness, as a coefficient of simple linear regression of income value and snowfall weather index (Castelino, 1992;Rohrer, 2004).
When evaluating weather forward, we will use the simplest and at the same time efficient method for price setting -method of historical simulation, which is based on historical data of Kopaonik snowfall for the past 10 years. As Hnilica (2007)ascertained, a method of historical simulation represents a simple model for weather derivative evaluation and a base for further statistical analyses, with the assistance of appropriate software (Crystal Ball, for example).
Therefore, we have two goals: firstly, we want to show that there is a statistically significant connection between snowfall and ski lift operators' income, and secondly, we'd like to analyze forward implementation effect in hedging weather risk and reduction of weather risk in ski lift operators' business at Kopaonik.

Literature review
The largest number of weather derivatives implementation research is found in the agriculture sector, while in tourism that is not a case. Researching winter ski tourism, most researchers focused on basic questions of climate changes and snowfall influence on tourism companies business and tourists' behaviour. Thus, researching the connection between lack of snow, skier's activities and ski ticket sales in winter Based on formerly presented studies, it is evident that all researchers emphasize the significant influence of weather variables (snowfall) on business results of world's ski resorts, as well as positive results of derivative's hedging. Since these kinds of studies haven't been conducted in our region so far, it is our task to ascertain to what extent ski centre's business depends on snowfall quantity, to present weather derivative as a new risk management tool and to measure the effectiveness and efficiency of chosen derivative hedging.

Research methodology and results
To investigate the possibility of efficient weather risk derivative hedging, we must go through several steps: 1. Weather index determination (snowfall index -SFI), 2.
Investigation of ski resort revenue dependence on tickets sales based on snowfall quantity (regression and correlation analyses, autocorrelation test and test of heteroscedasticity), 3. Construction of chosen weather derivative (weather snowfall forward), and 4. Evaluation of snowfall forward hedging effectiveness.

Weather index determination
In our research we will use cumulative snowfall index -CumSFI, corresponding to the sum of overall monthly falls during winter skiing season  16), which in our case represent season cumulative indices CumSFI for Kopaonik ski resort. Named period of 10 winter seasons has been chosen because this period had available data about income generated from ski tickets sales in given ski resort, which is Kopaonik ski operator. Cumulative index CumSFI presents overall snowfalls data in five months period (December -April). Snowfalls, as well as ski resort monthly income from ski tickets in November, are excluded because ski resort opens from 1 st December. Snowfalls and incomes from ski tickets are shown in Figure 1. Coefficient β is the most important parameter for us because it shows the change of dependable y (ski resort income from ski tickets sales) when the independentx (quantity of snowfall) changes per 1 unit, that is 1 cm. The value of this coefficient will be of importance when constructing weather derivative, i.e. determination of its tick size.
Evaluation of set regression model will be reliable only if standard assumption model is fulfilled. Otherwise, the evaluation of regression coefficient would be biased. That is why it is necessary to examine several assumption models. The first assumption that needs to be fulfilled is testing residual homoscedasticity and heteroscedasticity. We will use Gained coefficient evaluation forms final regression model of the influence of snowfalls on ski lift operator's ticket sale sincome. The results gained by regression and correlation analyses are shown in Figure 2. and in Table 2.  The most significant data in Table 2. are standard deviation and variance of snowfall and incomes during given months, which present a measure of risk, i.e. ski resort weather risk income volatility. We can see that highest values of standard deviation are achieved for January, February, and March when ski resort generates most of its income. This is the most important information for ski resorts when it comes to decision making about contracting certain monthly derivative in order to hedge, regardless of its type (forwards, futures or options).

Construction of weather snowfall forward
Let us assume that ski lift operator wants to hedge from the lack of snowfall. For this purpose, weather forward on snowfall index is arranged, with strike level SL = 214.6 cm of snow. Ski lift operator holds a short position in this contract. The payoff is expected in acase that CumSFI < SL, i.e. (SL -CumSFI) x T.
The other contractor, let's say an insurance house,for example, holds a long position and any snowfall quantity bigger than SL (214.6 cm) is in order since the house will be expecting a payoff in the value of (CumSFI -SL) x T.
Weather forward value is gained through the following formula (short position): Where is The weather index unit value T is gained based on regression equation coefficient and snowfall average in given time. Mathematical calculation shows T = 11.000 € / cm of snowfalls. Since the trading takes place on OTC market, we assume that both buyer and seller of weather forward agreed on the T money value.
Basic elements in the construction of weather forward are weather index (CumSFI), the weather index unit value in money (tick -T), strike level (SL), maturity periodt, interest rate -r, and reference climate point or measuring station. Since forwards are traded on bilateral, OTC market, it is up to contracting parties (buyer and seller) to agree upon all named elements. In this case, there is a possibility of arranging a monthly, several months or seasonal forward on snowfall. Depending on the maturity period (monthly or season), we get strike level SL, which presents snowfall average for a certain period (Mean). The elements of weather forward are shown in Table 3. Hedging effectiveness is gained on the base of formula (Ederington, 1979;Tang, Jang, 2012, Gupta, Kaur, 2015: Results of weather forward hedging effectiveness for the past 10 winter seasons at Kopaonik are shown in Table 4. Obtained results are shown in Table 5.  If we look at the variable costs values -V c , we can see higher values next to CumSFI lower values. The explanation lies in greater use of snowmaking system, which the ski lift operator is using if there is the lack of natural snow. According to Kopaonik ski resort, an optimum level of snow on the ski track is 30 cm for daily use, with a maximum temperature of -2,5 o C needed for efficient system operation. For example, during 2007/08 season, Kopaonik ski resort made 300.000 m 3 of artificial snow (average of 33cm of snow on the tracks) with costs of 20,00 RSD/m 3 (6,6 million RSD). If we have in mind that CumSFI = 222cm in that season (more than the historical average of 214.6 cm), we can conclude that ski resort was forced to use snowmaking systems during certain periods, which affected the growth of V c .

Conclusion
We have presented seasonal weather forward as a new and efficient tool for weather risk financial hedging. Our research was focused on analyses of a possible implementation of snowfall forward in ski lift operator business at Kopaonik mountain ski resort, in order to reduce the lack of snowfall risk and ski ticket income volatility decrease. Using basic techniques of risk management we can achieve more than expected ski ticket income hedging effectiveness by over 40%. Considering the fact that companies have a different understanding of risk and aversion to risk, this paper presents in what way and with what kind of hedging strategies (short and long hedge) tourism companies can control such risk. In our example, we used the dominant variable for winter ski tourism -snowfall, but we must also emphasize the possibility of using a temperature-based index derivative, which can be of importance for ski centre operative hedging, i.e. for snowmaking system.
This research is based on weather snowfall forward as basic and most simple future contract. Besides its advantages, we must point out some operational flaws of forward, which are reflecting in higher level of credit risk, insufficient liquidity (not transferable) and difficulties to find the other partner in OTC trading. Therefore, tourism companies have liquid contracts, such as futures and options, available, both standardized (exchange-traded derivatives) and in OTC market. The stock exchange is offering contracts which are monthly and seasonal. In this paper, besides seasonal, we have shown monthly digression of snowfall index and incomes, which make a basic data for the potential choice of weather derivative and construction of the same.
Given results are the foundation for further exploration of possibilities and effectiveness of weather derivatives application in winter ski tourism in Serbia. Since there haven't Bojan S. Đorđević been any researches of this kind, author's intention was to explore the options of implementation and weather option effects in Kopaonik ski resort in a future period, as well as in other ski resorts -Zlatibor and Stara Planina. This will enable an overview of geographic diversification -geographical basis risk and construction of weather basis derivative, based on chosen contract -forward, future or option. It is up to companies in winter tourism sector to keep track of new discoveries, inspect all possibilities and be prepared for timely reaction in terms of greater exposure to different kinds of risk. They already have different strategies of financial and operative hedging at their disposal.