Speculation on currency derivatives - the case of option and futures contract on polish market

Размещено на http://allbest.ru

Poznan University of Economics and Business

Speculation on currency derivatives - the case of option and futures contract on polish market

Morkowski J. - student, Rutkowska A. - PhD,

Poland

Abstract

The aim of the paper is to present the investment possibilities of currency derivatives. For analysis we choose option and future contracts in order to check whether derivatives with symmetrical or asymmetrical risk profiles are more effective in terms of profits. The study was carried out for 5 currency pairs EUR/PLN, USD/PLN, GBP/PLN, CHF/PLN and JPY/PLN. To predict currencies' market moves, we use two approach: technical analysis (TA) and neural network (NN). The strategies are tested using the one month for out-of-sample testing.

Keywords: foreign exchange market, derivatives, neural network

Introduction

The foreign exchange market in Poland includes spot market transactions, outright forwards, fx swaps, CIRSs, futures contract and currency options. According to the Bank for International Settlements (BIS, 2017) of foreign exchange and OTC derivatives markets survey in April 2016, the average daily net turnover in the domestic foreign exchange market amounted to USD 9.116 million and was 21% higher (at current exchange rates) than the value of transactions concluded in April 2013. However, in April 2016, as in April 2013, the average daily net turnover in the domestic currency options market amounted to USD 70 million. Furthermore, over 95% of all trades with financial institutions were transactions with foreign banks, parent banks of the reporting dealers in Poland. As many studies show derivatives contribute to improved market liquidity and increased capacity of the financial system to effectively price and bear risk.

Currency options in Poland occurred in 1996 and were offered only by 4 banks and the main currency for transactions concluded was the US dollar and the German mark. When Poland in 2004 joined the European Union, the popularity of these instruments increased by dint of to the removal of barriers to the flow of capital, services and goods. This event prompted Polish entrepreneurs to enter foreign markets and was associated with increased currency risk. A frequent form of hedging the currency exchange rate up to 2007 were currency options. Polish enterprises most often chose forward, options were the second most popular instrument. However, the currency market crisis has shaken trust in just developing options and contracts in Poland. One of the most frequently mentioned reasons for the low popularity of derivatives in Poland is the toxic currency options. In the years 2008-2010 Polish enterprises fell into financial trouble as the result of concluded agreements for currency options despite the lack of adequate security in assets. The losses were estimated at between a dozen and even 200 billion zlotys. Poland's Central Statistical Office (GUS 2010) puts the aggregate result on 2009 options operations at an estimated PLN 90 billion (Lib- eradzki, 2015).

The paper focuses on options and contracts futures as well as their investment possibilities for the underlying assets - currencies. The main difference for investors between selected derivatives is the fact that contracts are symmetrical instruments and options are asymmetrical. This means that in the case of a contract on both sides there is an obligation, while in the case of options one party has an obligation and the other has possibility. Due to the options' asymmetry, an option premium is required. Option premium is an amount the option buyer pays, and the option seller receives for granting the specified rights for the specified period under the option. Consequently, by buying put or call options we limit the possible loss to the maximum of the option premium and the possible profit is unlimited, while when we sell call or put options, the maximum profit is the amount of bonus received and the loss is unlimited. In the case of a contract, the break-even point is equal to the exercise price, whereas in the case of options it is at the exercise price +/- the amount of the bonus. Options can be divided into European and American options. The European ones are those in which the date of executing the option is known in advance. In the American option, the expiration date is specified but the option can be exercised at any time during its lifetime. Despite that, options and contracts have also many features in common, e.g. both can be with or without delivery, contracts and options may have the same underlying assets (currencies, assets, etc.).

Literature review

The literature on derivative instruments focuses on their valuation (Garman & Kohlhagen, 1983, Kanniainen et. al. 2014, Christoffersen 2015) or hedging possibilities (Geyer-Klingeberg et. al. 2018, Zhang, 2018). In our work, we test their speculative possibilities based on neural networks (NN) and technical analysis (TA).

The very dynamic development of this branch of artificial intelligence means that they also find their application in economics (Tealab, 2018, Li et. al., 2018). NN were created thanks to the inspiration of the human brain, which can make better decisions than computers that can perform very complex operations. NN are supposed to resemble the human brain.

The use of NN to predict future events is a relatively new tool that gives many possibilities, which will be wider and wider considering the continuous development of computers' computing power and, consequently, their capabilities. In the case of willingness to invest using the network, the network learning process is very important, the task of which is to find mechanisms that affect the characteristic changes in asset prices. (Banasiak, 2011).

Technical analysis is as a short-term analysis of investment profitability, made based on charts of changes in stock prices, turnover value, order volume and technical indicators.

Technical analysis is used to determine the likely future share price change based on their past evaluation, considering all factors that have or may have an impact on their supply and demand (Murphy, 1999). The following indicators EMA, SMA, RSI and strategy MACD were used in investing using technical analysis. Futher information about TA can be found in (Taylor & Allen, 1992, Avramov et. al. 2018).

Results

In the conducted study, three-year historical data (2015-2017) were analyses and one-month investment out of sample were made. The first of the analyzed currency pairs is the American dollar (USD) to Polish zloty (PLN) exchange rate. As we can see on Figure 1, there are constant declines since the end of 2016, which lasted until the end of the analyzed period. This decline is of a long-term nature.

The euro exchange rate could be considered very stable, and its deviations from the average were even more negligible. The euro exchange rate is smaller than in the case of the dollar, but it is still quite significant and over 3 years amounted to approximately PLN 0.01. Noteworthy is the very small variance of only 0.9%. The distribution resembles a normal distribution, slightly flattened, with a kurtosis of 0.0065. The currency characterized by the largest range is the British pound. PLN 6.0827 had to be paid for one pound on November 30, 2015, while the lowest rate in the analyzed period occurred on August 28, 2017 and amounted to PLN 4.5908. In the case of the Swiss franc, we are dealing again with the situation when the removal of a certain time interval would have a significant impact on the maximum value, and hence on the average and standard deviation. Here, the Black Thursday took place at the beginning of 2015. In January 15, 2015 Polish zloty lost 30% to the Swiss franc in just a few minutes. The closing price of 14.01.2015 is 3.55, while the next day, the Swiss franc closes the day with 4.32. This situation was very important for over half a million borrowers with their loans in francs. This situation resulted from the artificially defending franc rate by the Swiss government (Przygorzewski, 2015). The Japanese yen has been very popular in recent years, currently placing third in the most traded currencies on the FOREX market. The yen is the only of the analyzed currencies, in which we compare the Polish Zloty not to 1 unit of foreign currency, but to 100. This is due to the small value of 1 yen at the level of PLN 0.03-0.04.

Figure 1. Raw data of currencies pairs during data analysis period (own)

In empirical research, to receive comparable results, we assume: options are European; the strike price for investments using a neural network is equal to the estimated price, while in the case of investments using technical analysis, the strike price is the price at the money; the valuation of the option premium was made using the Black Scholes model; without delivery, the investment time horizon is the same and amounts to 1 month.

To build and train NN we use the Sta- tistica program with fixed for all surveyed currencies settings: neural network type - MLP, minimum 2 hidden layers, maximum 10 hidden layers. 20 networks are built for learning and validation, before selection the one with the highest level of fit and efficiency.

Figures 2-6 present the quotation in test period and strike prices in line with the above principles. The dotted lines are used to show strike prices in case of decrease prediction and solid ones in case of increase prediction.

The quantitative results of a hypothetical investment are presented in Table 1 and Table 2. As Figure 2 shows, after a period of short growth, the USD prices were falling. This change in the trend was foreseen by the TA, but not by the NN.

Figure 2. USD / PLN quotations during the test period and strike prices (own)

In the USD / PLN chart, you can see that through most of the investments using technical analysis and neural networks (option) they bring profit. However, at the end of the period, only investments using TA are characterized by a profit.

Figure 3. The investment in the euro has been the easiest to estimate, because all four

EUR / PLN quotations during the test period and strike prices (own) instruments give profit. Statistical indicators show large stabilization of the euro against the Polish zloty.

Figure 4. CHF/PLN quotations during the test period and strike prices.

The Swiss franc for over 25 days of investments showed losses in most positions. The change of course in the last days of January meant that three positions gave a profit.

Figure 5. GBP/PLN quotations during the test period and strike prices

The British pound chart is very interesting. For most of January, the chart will move between two lines indicating investments using options and two lines for investments using contracts. This time, the end of the month is unfavorable for investment, because the price increases, exceeding the break-even point in options using TA. exchange poland forward futures financial

Figure 6. JPY/PLN quotations during the test period and strike prices

Table 1

For all four investments in the Japanese yen, the break-even point was very close. Therefore, even the minimal change in the exchange rate, changed the value of the investment. Ultimately, we can see that the investment in the Japanese Yen brought a loss and profit in two positions. Table 1 shows the results of investments in currency options. In three out of five cases we obtained gains using both technical analysis and using neural networks. TA predicted correctly a fall in the dollar while the NN predicted wrong the growth. In the case of the Japanese yen, the NN predicted growth and TA accurately decreased. The reverse predictions of the network also concerned the British pound, but in this case, the neural network well predicted a change in the rate. In the case of investments in futures contracts, profits were obtained using of neural networks in two out of five cases, while in the case of technical analysis in four out of five.

Table 2.Futures contracts' investments results (own)

Conclusion

In empirical study futures contracts' profits are higher than in the case of currency options. It can be explained with the construction of a currency option (option premium). It is worth noting that the losses in this case are also higher. Profit / loss on currency pairs are not symmetrical because the technical analysis used the price of performance at the money and in investing with the use of the neutral network the strike price was equal to the price predicted by the neural network.

References

1. Avramov D., Kaplanski G., & Levy H. (2018). Talking Numbers: Technical versus fundamental investment recommendations. Journal of Banking & Finance, 92, 100-114.

2. Banasiak A. (2011), Zastosowanie sieci neuronowych w analizie tech- nicznej, w: A. Adamska, A. Fierla, Inwestowanie. Instrumenty klasyczne i alter- natywne, Oficyna Wydawnicza Szkoly Glownej Handlowej w Warszawie, Warszawa, 231-266.

3. BIS (2017), Turnover in FX and OTC derivatives markets in Poland. Re- treived from: https://www.nbp.pl/homen.aspx?f=/en/system-

4. finansowy/obroty.html

5. Christoffersen P., Feunou B., Jeon Y. (2015), Option valuation with observable volatility and jump dynamics, Journal of Banking & Finance, vol. 61, Supplement 2, 101 -120,

6. Garman M., Kohlhagen S. (1983), Foreign currency option values, Journal of International Money and Finance, 2(3), 1983, 231-237.

7. Geyer-Klingeberg J., HangM., RathgeberA.W. (2018) What drives financial hedging? A meta-regression analysis of corporate hedging determinants, International Review of Financial Analysis, ISSN 1057-5219.

8. Kanniainen J., Lin B., Yang H. (2014), Estimating and using GARCH models with VIX data for option valuation, Journal of Banking & Finance, vol 43, 200-211.

9. kurencja.com, (2017, June 1) Jen japonski (JPY) - wszystko o walucie, Re- treived from: http://blog.kurencja.com/jen-japonski-jpy-o-walucie/

10. Li Y., Jiang W., Yang L., Wu T. (2018), On neural networks and learning systems for business computing, Neurocomputing, Vol. 275, 1150-1159.

11. Liberadzki K. (2015). Toxic Currency Options in Poland as a Consequence of the 2008 Financial Crisis (April 20, 2015). Available at SSRN: https://ssrn.com/abstract=2669391

12. Murphy, J.,J. (1999). Technical Analysis of the Financial Markets: A Comprehensive Guide to Trading Methods and Applications, Penguin, ISBN110165919X.

13. Przygorzewski M. (2015, January 16), Czarny czwartek kredytobiorcow frankowych, Retrieved from https://www.bankier.pl/wiadomosc/Czarny- czwartek-kredytobiorcow-frankowych-3271785. html

14. Taylor, M.P., & Allen H. (1992), The use of technical analysis in the foreign exchange market. Journal of international Money and Finance, 11(3), 304-314.

15. Tealab A. (2018), Time series forecasting using artificial neural networks methodologies: A systematic review, Future Computing and Informatics Journal, ISSN 2314-7288.

16. Zhang W.G., Yu X., Liu Y. (2018), Trade and currency options hedging model, Journal of Computational and Applied Mathematics, Vol 343, 328-340.

Размещено на Allbest.ru