old is one of the treasurable and oldest metals among the inventions of the humankind. From ancient times, gold is accepted as the medium of exchange and as highly valuable wealth used in India. Times may change still gold stands as a traditional investment for everyone beyond the economic status. There are more than 16,000 tons of gold are available in India, and almost every family possess gold for its sentimental attachment. Gold has three purposes in general, namely -like ornaments, as household investments, and for industrial consumption.

In the 21st century, investors are more sensible enough to avoid high risk. In recent days, as the share prices are volatile, the investors are not anxious to invest their funds in the stock; they prefer investing in gold for its safety and liquidity. Volatility is the deviation of the return around its mean values either in the positive or negative direction. The Estimating of volatility becomes an essential task in the management of the portfolio. In the case of gold, its price always depicts an increasing trend to benefit the investors even during the recession period. However, the investment in gold needs a huge fund due to its excessive demand and less supply. Due to this, small investors find it hard to invest in bullion markets.

To enable the small-time investors to park their fund in the gold market, the Gold Exchange Traded Fund (Gold ETF) scheme introduced in the USA in the year 1993. In India, the ETF method is in vogue since 2007 onwards. The scheme Gold ETFs is just like other stock which is traded generally in the stock markets. It is a kind of mutual fund that is listed and operated in the stock market. They can be bought and traded through the online Demat account. As the system is more accessible and requires a less small amount of fund to invest, the schemes under GTFs are more lucrative for small investors, unlike conventional gold. This study aims to observe the casual relationship and volatility among the natural Gold and Gold ETF.

A very few studies have conducted in the field of gold Exchange traded fund, which is referred to in this segment of review of the literature. (Pandey, 2010) felt that buying Gold ETF is safer, convenient, and adds some tax-oriented benefits when compared with physical gold. (Mishra, Das & Mishra, 2010) attempted the casual relationship between Indian natural gold price and stock market indices return found that there is a long-run equilibrium relationship exists, and the one-way causal relationship was noticed between Gold price and Stock market return. (Athma, Prashanta, & k, 2011) Stated that gold investment is less volatile while comparing with equity share investment. Gold ETF is one of the best methods for portfolio diversification. (Kumar, Kumar, & Roy, 2012) analyzed the Gold ETFs performance by conducting risk and returns, and proved that Gold ETF had given good return compared with equity fund. (Nemavathi & Nedunchezhian, 2013) Attempted to estimate the volatility of gold and Gold ETF by using EGARCH model. They concluded the volatility of the fund is based on their yield performance. (Narend, 2014) Analyzed performance of Gold ETFs by Jensen Alpha Return. They estimated that ETFs are giving more returns and performing better than physical gold. (Tripath, Parashar & Singh, 2014) studied the causal relationship between the gold price and other macroeconomic factors, the result shows that there are Granger Causes that exists in the exchange rate and oil price in India.

(Mishra P. K., 2014) stated that gold is the best investment option for an investor in the alternative of other investments and there was a bidirectional connection between the gold price and stock market return in India. (Gencer & Musoglu, 2014) empirically analyzed gold price in Turkey by using the GARCH model and result expressed that market return has an impact on gold price volatility. (Anand, 2017) Found that Gold ETFs are influenced by gold price in India. In a long-run relationship, gold ETFs are giving more return, so it attracts the investors to invest in Gold ETF. (Jain & Mary, 2018) Found that gold ETF is a new concept in the investment portfolio for small investors. Gold ETFs is an easy and convenient way of trading. Investors need not worry about the security on holding storage and other physical damage like theft and due to natural scarcity of demand in gold.

However, the literature review reveals a certain aspect of the real truth of the research on the existence of Gold and Gold ETF. This study aims to estimate the short run and long run relationship between gold and Gold ETFs and its volatility.

? To investigate the volatility effect in Real Gold and

Gold ETFs;

? To analyze the short run and long run association between Gold and Gold ETFs in India; ? To observe the impact of Real Gold price return in Gold ETFs return in India.

IV.

This study was aimed at estimating the volatility and relationship between Gold and Gold ETF in India. For this purpose, the top five gold ETFs which are listed in the National Stock Exchange was chosen based on the returns in the field of Gold ETFs in India. The Gold ETF selected for study are presented in Table 1. This study is empirical in nature, based on secondary data. The daily historical data from 1 November 2015 to 31 October 2018 (739 Daily Observations) were collected from National Stock Exchange (for GOLD ETF) and World Gold Council (for Physical Gold) for estimating volatility and relationship of gold and gold old ETF. The return on the gold price was calculated as the logged difference between time period t and t -1 .

Where R t is the daily return of gold price at time t. P t denotes the price of gold per gram at time periodt,and P t-1 indicates the price of gold per gram in the selected ETFs at time period t-1. For analyzing the data, various econometrics tools used such as the Augmented Dickey-Fullertest, Granger Causality, Johenson Cointegration test, and GARCH model and LM-ARCH test were applied by using "Eviews 10" statistical software package.

V.

Descriptive Statistics: The descriptive statistics of Gold and Gold Exchange Traded Funds are presented in Table 2. Table 2 displays the summary of statistics of the gold and gold ETF data. As per the table, the mean values of gold ETFs are varied from each gold funds. Among these funds, IDBIGOLD had a more standard deviation that means that gold fund more volatile in the market and it has high risk; KOTAKGOLD was less volatile compared with other gold funds. Gold is very high because of various macroeconomic factors. The skewness for the gold funds and gold was almost negative (asymmetrical value), and Kurtosis concerned for Gold funds it was found to be nearly 3 (approximately). The Jarque Bara test for regularity was significance at the level of 5%, indicating the data used for this study were not normally distributed.

The correlation coefficient is used to measure the extent of the strength and direction of the connotation between the indices and stock returns in different countries. The analysis of Karl Pearson's Correlation is presented in Table 3. Table 4 shows the result of Multiple Regression Analysis, which is conducted on non-stationary data. As per the table, the R-squared value is nearly 96% that means, the independent variable is having a high impact on the dependent variable. Almost all the independent variables are significant at 5% level other than QGOLDHALF, among those variables HDFC gold ETF has an extremely high coefficient value. Augmented Dickey-Fuller Unit Root Test: Before examining the relationship between Gold and Gold ETF, it is crucial to check the univariate properties of the data sequence are non-stationary, or they comprise a unit root. For that, the ADF unit root test is employed, which was developed by Dickey-Fuller. A time series is said to be stationary; that means, the alteration of the series does not symmetrically fluctuate over time. Nonstationary data will lead to incorrect values. (Dickey & Fuller, 1979). Unit root was calculated as per the following equation:

Where ???? ???1 is the first difference operation, ?? 1 , ?? 2 ?? are coefficient to be estimated. ?? = 0, ?? ?? is white noise error term, if the estimated slope of coefficient not in this regression ?? (hypothesis) is zero or not. if it is zero, then ?? ?? is nonstationary. The best lag length was taken with the Akaike Information Criterion (AIC) and maximum, lagwas put to 36. (Gujarati, 2009)The ADF null hypothesis is established as unit root in the time series, whereas, the alternative is -there is no unit root and it is stationary, which is observed and presented in table 5.

Volume XIX Issue III Version I Year 2019 ( ) B From the table, all the probability value of 'At level' is not significant that means the data series is not stationary; it has a unit root. In the first difference, all the index probability value is less than 5%, that means it rejects the null hypothesis to accept the alternative, so the data is stationary.

The Granger causality test is directed to inspect the direction of causality amongst Real gold and Gold ETFs. This test is applied only to stationary time series data. Granger causality was calculated as per following formula (Gujarati, 2009)

Where Î?" is the difference operator, ?? ????? and ?? ????? are represent as the lagged value of ?? ?? and ?? ?? . ?? 1 and ?? 2 are error terms assumed white noise. The lag length was picked by using Akaike Information Criteria (AIC) as the most favourable number. The Granger Causality Test results are presented in Table 6 - The table 6, expresses that Granger Causality Test results in which the value of probability is less than 5% that means it rejects the null hypothesis, showing that the Axis Gold ETF and HDFC Gold ETF are bidirectionally caused with gold, other variables are causing unidirectionally. So, all the data are having cause and relationship with Real gold. Particularly the price of real gold has more effect on the rate of Gold ETFs.

Results of Co-integration Test: (Johansen, 1990) Cointegration test is the most commonly used method in investigative the long-run equilibrium association of the different time series or integration in the financial market. The data becomes stationary after the first difference in the ADF test. Following table 7, shows the cointegration relation between Real Gold and Gold ETFs. The result of the cointegration test, Trace and Max-Eigen values are checked at 5% significant level.

Here the null hypothesis is that 'these series are not integrated with the Real Gold Price'. As per the table 7, all the Gold ETFs are less than 5% level of significance. That means to reject the null hypothesis and accept the alternative one. All the Gold ETFs are having a long-run equilibrium relationship with Real Gold Price. Estimation of Volatilit: GARCH model is useful in analyzing the financial time series such as market indices. A unique feature of these models is that the error variance may be correlated over time because of the phenomenon of volatility clustering. The AutoRegressive Conditional Hetroskedacity model was developed by (Engle, 1982) or Generalised AutoRegressive Conditional Hetroskedacity Effect (Bollerslev, 1986). GARCH model was initially proposed by Bollerslev the simplest model GARCH (1,1) can be inscribed as (Gujarati, 2009)

Where ?2t is the variance for the time t. ?i and ?j are coefficients. ?t-i is the lagged residual from the mean equation and, ?2t-j is the lagged variance from the period t-j. ?? is the coefficient measuring the impact of real gold price on Gold ETFs. Estimated GARCH coefficient and prob value for returns are presented in Table 8. From table 8, ARCH and GARCH coefficient were all significant at 5% level, other than KOTAKGOLD Which means there is an autoregressive effect in all the Gold ETFs. So, the future is influenced by the past movement of gold fund returns and GARCH implies there is strong volatility clustering effect was found in the data. The impact of Real gold in Gold ETFs was significance ag 5% level. However, the coefficient values are very less that means even though it has an impact,butit is not a strong impact over the ETFs. ARCH-LM test: ARCH LM tests were conducted to find out whether any autocorrelation was found in the residuals of the GARCH equation, which is necessary to verify any arch effect that has remained in the data or not. The null hypothesis of this test is that the residuals from the Generalised AutoRegressive Conditional Hetroskedacity equation do not have the ARCH type of heteroskedasticity.

Residuals are free from autoregressive heteroskedasticity, the estimated coeffect (Obs*R-Squared) of the ARCH-LM test and its P values are obtainable in Table 9.

This research article was planned to examine the relationship and volatility between real gold and gold ETFs in India. For the period of three years from 1 November 2015 to 31 October 2018, with selected Gold Exchanged traded funds by using various econometric analysis. The study was found that there is a strong positive relationship amongst the real gold and gold ETFs. By measuring the correlation test implies there exist short-run relationship and by using Johenson cointegration test, it shows that there is a long-run equilibrium relationship also. The data for the study was non-stationary while calculating ADF at first level it was significant at 5% level that means the data termed to be stationary. By using the Granger causality analysis, it shows that there are a cause and relationship between gold and Gold ETFs either in one way or both the ways. The volatility effect and clustering effect was found in the Gold ETFs; all the gold funds are performing almost the same as the past period performance. These results help to investors, market research's, companies and other financial institutions to make the best decision towards the Gold ETFs. It will also help to increase the Gold ETFs trade in the future.