VOLATILITY SPILLOVER AND CONTAGION EFFECTS BETWEEN EURODOLLAR FUTURE AND ZERO COUPONS MARKETS: EVIDENCE FROM ITALY

This paper examines the time-varying conditional correlations between the Eurodollar futures market and the zero coupons of Banca Fideuram. We apply a bivariate dynamic conditional correlation (DCC) GARCH model in order to capture potential contagion effects between the markets for the period 2005-2017. Empirical results reveal contagion during the under-investigation period regarding the twenty-one bivariate models, showing that the Eurodollar futures market has a major impact on the zero coupons of Banca Fideuram. Findings have crucial implications for policymakers who provide regulations for the above-mentioned derivative markets.


INTRODUCTION
Τhis paper investigates the potential volatility spillover and contagion effects (Dimitriou, Kenourgios & Simos 2013) of the Eurodollar futures market and the zero coupons of Banca Fideuram. We consider the zero coupons of Banca Fideuram ending from 2018 to 2033. By employing a bivariate DCC-GARCH model, we show significant volatility spillover effects (Sehgal, Ahmad & Deisting 2015;Li & Giles 2015;Aboura & Chevallier 2015;Antonakakis, Floros & Kizys 2016). Moreover, we use the definition of contagion as suggested by Forbes and Rigobon (2002). They defined contagion as a significant increase in cross-market linkages after a shock. Dynamic conditional correlations reveal contagion effects (Dimitriou & Kenourgios 2015;Sensoy & Hacihasanoglu 2015) in sub-periods between the Eurodollar futures market and all the zero coupons of Banca Fideuram.
The motivation for this paper is analyzed as follows. Firstly, there is no other empirical research investigating the conditional second moments of the distribution between the Eurodollar futures market and the zero coupons of Banca Fideuram. Secondly, the potential existence of contagion between the Eurodollar futures market and the zero coupons of Banca Fideuram provides new evidence for financial theory. Thirdly, the under-investigation period is of great importance, since it entails major economic crises i.e., the financial crisis of 2008.
The paper is organized as follows. Section 2 presents the literature review and Section 3 provides the data characteristics. Section 4 provides the methodology. Section 5 shows the empirical results. The last section provides the conclusion.

LITERATURE REVIEW
There are numerous empirical studies investigating the spillovers among different future and financial markets (Mensi et Mensi et al (2013) find evidence of spillovers between the S&P 500 and commodity price indices for energy, food, gold, and beverages over the turbulent period from 2000 to 2011. Kavussanos et al (2014) examine the existence of spillover effects between commodity and freight markets for the period 2006-2009. By using different GARCH models, they show the existence of spillovers effects. Li et al (2014) show potential spillovers and dynamic conditional correlations between spot and forward tanker freight markets. By using a multivariate GARCH model, they examine the period from 2006 to 2011. Antonakakis and Kizys (2015) find evidence of volatility spillover effects between commodity and FOREX markets: crude oil, gold, silver, platinum, CHF/USD, GBP/USD, EUR.USD. They investigate the period 1987 to 2014. Du and He (2015) found evidence of significant spillover between crude oil and stock markets using daily data of the S&P 500 stock index and West Texas Intermediate (WTI). Based on their results, they supported the existence of positive risk spillovers from stock to crude oil markets and negative spillovers from crude oil to stock markets. Ewing and Malik (2016) examine the volatility of oil and US stock market prices incorporating structural breaks using daily data from 1996 to 2013. By employing univariate and bivariate GARCH models, they find no volatility spillovers between the two markets. Bagchi (2017) investigates the dynamic relationship between crude oil price volatility and stock markets in the emerging economies like BRIC (Brazil, Russia, India and China) countries. By using a AR-APARCH model, he finds evidence of positive and negative relationships between the underinvestigation markets. Roy and Roy (2017) show the financial contagion in Indian commodity derivative markets vis-à-vis bond, FOREX, gold, and stock markets. They applied a multivariate DCC-GARCH model for the period 2006-2016. Ma et al (2019) examine the inter-connectedness between WTI oil price returns and the returns of listed firms in the US energy sector for the period 2008-2018.
They show that, although idiosyncratic information is mostly independent of oil shocks, individual energy stock returns do respond to WTI price movements. Tsiaras and Simos (2020) prove the spillover effects among S&P 500, four national equity markets and the respective FOREX markets for the period from 2010 to 2018.  investigates and proves the spillovers between JPY/USD, KRW/USD, EUR/USD and INR/USD futures markets for the period 2014-2019. In our paper, we provide empirical evidence of spillover effects between major future FOREX market and Zero Coupons derivative markets.
To the best of our knowledge, there is no previous empirical evidence providing evidence of spillover effects between the under-investigation market.  11, 2017 (3375 observations). We use the market returns generated by the equation r t = log(p t ) -log(p t-1 ) , where p t is the price of future market on day t and p t-1 is the price of future market on day t-1.

DATA CHARACTERISTICS
In tables 1, 2, 3 and 4 we see the summary statistics for the markets returns. BANCA FIDEURAM ZERO CPN. 2032 exhibits the highest mean value (0,00023071). Based on the highest maximum (0,077701), the second minimum (-0,066133) and the second highest std. deviation (0,0095707) values, BANCA FIDEURAM ZERO CPN. 2032 presents the largest fluctuations among all the markets. Additionally, all market returns are negatively skewed, except the cases of BANCA FIDEURAM ZERO CPN. 2018, BANCA FIDEURAM ZERO CPN. 2019, BANCA FIDEURAM ZERO CPN. 2020 and BANCA FIDEURAM ZERO CPN. 2021. Furthermore, we observe that all market returns show excess kurtosis. In addition, Jarque-Bera statistic results indicate the rejection of the null hypothesis of normality for all market returns. ADF (Dickey and Fuller 1979) test results reject the null hypotheses of unit root at 1% level, showing that the daily market returns appropriate for further testing.

METHODOLODY
In the first stage, we generate the daily logarithmic returns: is standardized residuals, defined as follows: , where and are i.i.d.
where is standardized errors and is conditional variance depending on and for each market lagged one period, generated by the univariate GARCH(1,1) model (Bollerslev 1986): (1) (2) EJAE 2020  17(2)  67 -88 where ω is constant, a and b are ARCH and GARCH effects .
In the second stage, we employ the Engle (2002) representation of the bivariate GARCH model inorder to estimate the bivariate conditional variance matrix ( H t is N x N matrix, with N the number of markets, i = 1,…,N) as follows: is the conditional variance matrix given by: R t is the condition correlation matrix of N x N dimension, and is defined, as follows: where the N x N symmetric positive definite matrix is given by: i s the N x N unconditional variance matrix of u t , and α and β are nonnegative scalar parameters, satisfying α + β < 1.

EMPIRICAL RESULTS
In this section, we present the empirical results generated by the multivariate DCC-GARCH model. Sub-section 5.1 shows the results of the univariate GARCH model, while in sub-section 5.2 we analyze the results of the multivariate DCC-GARCH model. In sub-section 5.3, we report an analysis of the generated Dynamic Conditional Correlations (DCCs).

RESULTS OF THE UNIVARIATE GARCH (1,1) MODEL
Tables 5, 6, 7 and 8 show the estimated values for mean equation and univariate GARCH (1,1) model. We observe statistically significant μ for all the market returns, except the case of DGCX-EUR/USD CONTINUOUS AVG.-SETT. PRICE. Additionally, empirical results report statistically significant ω for all the market returns. Moreover, ARCH (a) and GARCH (b) terms are highly significant for all the markets returns. (3) EJAE 2020  17(2)  67 -88

RESULTS OF THE BIVARIATE DCC-GARCH (1,1) MODEL, DIAGNOSTIC TESTS AND SELECTED INFORMATION CRITERIA
Tables 9, 10, 11 and 12 present the results of the bivariate DCC model estimations. We observe that the average CORij is statistically significant for the pairs of markets: DGCX-EUR/USD CONTINUOUS     In tables 13, 14, 15, and 16, we report the results of the diagnostic tests and information criteria.
x 2 (4) statistic results suggest that the null hypothesis of no spillovers is rejected at 1% significance level. Ljuing-Box test results (Hosking, 1980;Li-McLeod, 1983) provide evidence of no serial autocorrelation, suggesting the absence of misspecification errors of the estimated multivariate GARCH model. Moreover, the estimated AIC and SIC information criteria are presented.    Figures 3 and 4 plot the conditional covariances for all the pairs of market returns during the whole period. We observe a tremble trend for all the conditional covariances. Additionally, conditional covariances seem to be extremely volatile.     Figures 5 and 6 present the pair-wise Dynamic Conditional Correlations (DCCs). We observe strong co-movements for all DCCs. DCCs have positive values in sub-periods, indicating the existence of contagion, implying the specific correlations risky for any investor. Furthermore, we can notice the effects of major economic events on the DCC graphs as we see that the lines are bouncing above and beyond, i.e.

CONCLUSIONS
This paper investigates the potential volatility spillovers effects and the existence of contagion effects of the Eurodollar futures market and sixteen zero coupons of Banca Fideuram by employing a bivariate DCC-GARCH model. We set the under-investigation period from 2005 until 2017. To the best of our knowledge, this is the first empirical study investigating volatility spillovers between the Eurodollar futures market and the zero coupons of Banca Fideuram.
The main empirical results are summarized as follows. Based on the descriptive statistics, BANCA FIDEURAM ZERO CPN. 2032 returns present the largest fluctuations compared to the rest markets. Furthermore, results of the bivariate DCC-GARCH model indicate strong evidence of volatility spillover effects. DCCs analysis shows evidence of strong co-movements for all the pairs of markets. Additionally, DCCs reveal contagion for all the pairs of markets in sub-periods. The empirical results are of interest to policymakers, who provide regulations for the under-investigation derivative markets, as well as to market-makers.

ACKNOWLEDGMENTS
This article was carried out by me independently. The research is original, and has not been submitted to any other journal. I want to thank the anonymous referees for their valuable comments and suggestions which helped me to improve the paper. Any responsibility for remaining errors in the resulting work is my own. Empirijski rezultati otkrivaju zarazu tokom istražnog perioda u vezi sa dvadeset i jednim bivarijantnim modelom, pokazujući da tržište futura Eurodollar ima veliki uticaj na nulte kupone Banca Fideuram. Nalazi imaju presudne implikacije za kreatore politika koji pružaju propise za gore navedena tržišta derivata.