The Role of Gold as Haven or Diversier Invesment in Indonesia

The objective of this study is to analyzes gold ownership in Indonesia related to the treatment of gold whether gold is used as a haven or diversier investment. To determine gold as a haven or diversier investment, the research method used is the general correlation method and cross correlation. The variables used in this study are gold price variables, investment interest rates, LQ 45, exchange rates and ination, using data from 2018.1 to 2020.10. The results show that gold in Indonesia is more of a haven than a diversier investment. Based on expectations of gold prices for the next year, it shows that the relationship between gold prices - interest rates and gold prices - exchange rates is a strong complementarity, while the relationship between gold prices and LQ45 is a very strong substitution or diversication.


Introduction
Covid-19 which has developed into a pandemic in various countries has caused many investors to have high concerns because the effects of the pandemic have made the change from a health crisis to an economic crisis. The existence of the world economic crisis has caused many people to think about saving their assets so that their value does not decrease. For this reason, they are thinking about replacing assets that tend to decline in value with other assets such as gold (for example).
Gold is a precious metal that attracts the interest of various circles, not only regarding its beauty but also the value of gold itself. Gold can be used as jewelry, it can also be used as currency, input in the industrial sector, as well as an investment tool.
In the discussion about gold as an asset, the debate actually includes two terms, namely wether gold acts more as a safe haven asset or as a diversi er investment. In this study, the two terms are matched with a store of value as a safe haven and investment tool as a diversi er investment.
A store of value is an asset that is able to maintain the value it contains, not easy to depreciate. Gold and other precious metals are good stores of value because they have a long shelf life. A country's currency must be a reasonable store of value for its economy to function smoothly. In other words, an object can be used as a store of value if the object has a relatively stable value and is less likely to depreciate. In the case of gold, the price of gold is relatively stable, at least until 2010 quali ed as a store of value. However, in the following period, there were relatively high uctuations, so gold became too risky to be considered a store of value. Therefore, the idea of using gold as an investment tool began to emerge.
The results of Fernando's research (2017) state that one of the reasons for the increasing interest in gold investment is the perceived risk in the economy. Risk management is very important in today's era which is full of uncertainty in terms of economy, nance and politics. At the same time, increasing correlations between traditional diversi cation roles make diversi cation more di cult. In Sweden, gold is a great diversi cation in the equity portfolio. So, according to Fernando, gold is an investment tool, while according to Ritholtz (2016) gold is not an investment tool but gold is a store of value and currency.
Historically gold has been seen as more attractive investment than stocks and bonds or other assets classes, not only for the positive returns but also as safe haven in time of di culty (Kovinski, 2014).

Conceptual Framework
In the discussion about gold as an asset, the debate actually includes two terms, namely whether gold acts more as a safe haven asset or as an investment diversi er. According to Flavin, Morley and Panapolou (2014), safe havens are assets that have low market risk and are liquid, which are used by investors to overcome fears of losses in the capital market. Technically, the meaning of safe haven is if the correlation between an asset and other assets is negative or zero, especially in market conditions full of uncertainty (Baur and Lucey, 2010 Shakil, et.al. (2017), and de Cnijf (2019) conclude that gold acts more as a hedging tool or a safe haven, when the economy is in a state of uncertainty until it penetrates the stock market. Faugere and van Erlach (2006) compare the price of gold with stock prices, exchange rates and in ation. According to Faugere and van Erlach (2006) gold in normal conditions has various roles, but in turnmoil conditions, gold is more visible in its role as a hedging tool or acts as a safe haven. De Cnijf (2019) argues that the belief that gold acts as a safe haven does not apply when the economic and stock market turmoil is extreme so that the ambiguity of gold's role is very obvious.
Similar studies with different results were conducted by Ibrahim and Baharom (2012), and Robiyanto (2018). They have a general conclusion that gold is able to make a signi cant contribution as an investment tool when the economy leads to uncertainty to penetrate the capital market conditions. In line with this conclusion, according to Robiyanto (2018), the role of gold in the Indonesian and Malaysian economies illustrates the diversi cation of investment, both in government and corporate bonds, because stocks themselves have shown hedge properties with the magnitude of the phenomenon of comovement with related factors. Agha et al (2015) state that there are two ways to invest in gold, namely by investing in physical gold and paper gold. They tested the gold status in the Islamic perspective, qualitative research method, tested operational, contemporary application and the issue of Sharia compliance. Based on Shariah compliance, out of seven banks, only one bank offers Shariah-approved gold savings accounts. Recent trends show that Malaysians are increasingly turning to gold investment as one of the sharia-compliant mechanisms to protect wealth.
Toraman et al (2011) used data from June 1992 -March 2010 namely data on oil prices, USA Exchange rate, USA in ation rate, USA real interest rate. The results stated that the highest correlation was found between gold prices and the USA exchange rate, where the relationship between the two was negative. The second highest correlation is between the gold price and the oil price (positive correlation).

Method
In an effort to determine whether gold assets act as a safe haven or diversi er investment in Indonesia, the calculation and analysis process is designed as follows: Stage I: Identify the Relationship between Gold Price Variables and Comparative Price Variables.
This stage is an attempt to determine the pattern of movement between the gold price variable and other variables identi ed as variables related to diversi er investment or safe haven activities. Activities used to describe investment activities are: Investment Credit Interest, LQ45 Index, while the variables that describe safe haven activities are in ation and exchange rates.
Stage II: Identify price movements. Identi ed by cross-correlation using optimal lag.

Stage III: Gold Prospect.
This stage is the forecasting stage of the price variable, whose value is then compared to nd out how relevant the placement of gold is as a safe haven or diversi er investment.

Data and data sources
Data were collected from various sources ( Table 1). The observed data period was 2018.1-2020.10 (34 monthly data). The data used in this study are as follows:  (1)  The two forecast models will have the same value in the initial period of their multi-period forecast. However, both will have differences in the continuation period. This difference will also appear when determining the bandwidth of the forecast. The dynamic forecast bandwidth will follow the error distribution that occurs dynamically by following the standard deviation pattern, as well as the static forecast will follow the error distribution statically.
Cross Correlation and Correlogram. Cross correlation is a method of measuring two variables to measure association relationships that occur at several speci ed time points. The two variables measured have different roles. The rst variable acts as a driver variable while the second variable acts as an output. Mathematically, the cross correlation is not much different from the Pearson Product Moment correlation.
The degree of strength of the association between the two variables between (-1.00) to (+ 1.00) is called the correlation coe cient. The correlation coe cient is closer to the value of (1), the more it will show the identicalness of the two variables at a certain point in time. While a negative value leads to an understanding that X leads Y and if a positive value means X lags Y. Investment analysts often use this method to understand the movement of these two types of investments in order to determine diversi ers investment. The smaller Theil value (the closer to zero) the better the forecast value.

Result And Discussion
The  Note: • Normal Distribution = Skewness ± 0.80; Kurtosis ± 3.00 • JB = Jarque-Bera = Changes in variables, such as D(Gold), D(LQ45) and D(Exchange rate) show the distribution abnormality. The interpretation that can be pinned on this condition is the assumption that the three variables, the movement of which describes the speculative attitude of the holder. Between the minimum value of change, Dmin, and the maximum value of change, Dmax, there is a high deviation. When deviations occur in the change data, it can be interpreted as a speculative trait.
Cross-correlation analysis. Cross-correlation analysis is used to determine the correlation between two variables by considering the movement for a certain range. In this case, the time range used is 16 lags following the default program. From the calculation results can be described: The interest variable is considered as a representation of the movement of the real sector. Assuming that the owner of the asset will place his asset with two alternatives. The rst alternative, assets are placed in a form that is able to provide pro ts, or diversi er investment. The second alternative, assets are placed in haven positions when pro ts cannot be achieved. The interest variable describes the investment situation in the real sector. Decisions about interest rates must be based on the market. The higher interest rates can be caused by a high demand for credit. Because the real sector promises big pro ts, the demand for credit will increase which will result in an increase in credit prices (interest rate).
From the correlation results, in general, it was noted that r(d(gold), interest) = -0.229 (prob.=0.20). This means that there is a tendency for the gold-interest relationship as a safe haven or hedge. These results are supported by cross-correlation results: Lag-0 to lag-4 r(d(gold), interest) < 0 and lag-5 to lag-16 r(d(gold), interest) > 0, with correlation which tends not to be statistically signi cant, it is more directed to safe haven.
These results are supported by Flavin et.al. (2014) and Baur and Lucey (2010), that gold has a tendency as a safety asset when the economy is in a turnmoil situation.
Cross-correlation d(Gold) ↔ d(LQ45). The LQ45 variable represents asset development efforts through the capital market. From the correlation results, in general, it was noted that r(d(gold),d(LQ45)) = 0.234 (prob.=0. 19). This means that there is a tendency for gold-LQ45 to be a safe haven. From the results of cross-correlation: lag-0 to lag-2 r(d(gold), LQ45) > 0 and lag-3 to lag-5 r(d(gold), LQ45) < 0. With a correlation that tends to not statistically signi cant, the gold-LQ45 relationship is more towards safe havens.
This nding is different from Robiyanto (2018) who relates it to bonds, that gold is an investment tool because bonds themselves have dual bene ts, rstly as an investment tool, secondly, corporate bonds and government bonds, have a tendency to always follow the movement of the driving factors, such as interest rates. or the exchange rate, so bondholders do not need to place gold as a safe haven. Gold will be used as a diversi er investment.
Cross-correlation d(Gold) ↔ In ation. From the correlation results, in general, it is recorded that r(d(gold), in ation) = 0.119 (prob.=0.51) meaning, there is a tendency for the gold-in ation relationship to move positively towards in ation, which indicates that gold can be relied on as a wealth guard. From the results of cross-correlation: lag-0 to lag-2 r(d(gold), in ation) > 0; lag-3 to lag-5 r(d(gold), in ation) < 0 and the next lag tends to be positive; r(d(gold), in ation) > 0. This con rms that gold is a safety tool against a decline in the value of money with a relative increase in the general price level.
Cross-correlation d(Gold) ↔ d(Exchange Rate). From the correlation results, in general, it was noted that r(d(gold),d(Exchange)) = -0.117 (prob.=0.52). This shows that the gold-exchange relationship tends to be a safe haven. From the results of cross-correlation: lag-0 to lag-3 r(d(gold), exchange rate) < 0 and lag-4 to lag-5 r(d(gold), exchange rate) > 0, with a correlation that tends to not statistically signi cant, so it is more directed to a safe haven.
This is in line with the research results of Shakil, et.al. (2017) which shows that gold is a good safety tool for uctuations in the IDR/USD exchange rate.
Forecasting. This forecasting stage considers the optimum lag which is calculated based on the Vector Auto Regressive (VAR) analysis. The results of the lag measurement that will provide the optimum lag are as follows:  Forecasting of several price variables analyzed is carried out for the purpose of seeing in the next year, how the nature of the movement of gold price expectations on the other four variables. The correlation that occurs between the gold expectation variable and other expectation variables is as follows: Correlation (goldsf-interestsf) = -0.99; Correlation (Goldsf-in ationsf) = -0.44; Correlation (Goldsf-Exchange ratesf) = -0.69; Correlation (Goldsf-LQ45sf) = 0.96. The correlation value indicates that the relationship between the gold priceinterest and gold-exchange rate in the future is a strong complementarity, while the relationship between the gold price and the LQ45 variable is a very strong substitution or diversi cation.

Conclusions
The results of the cross-correlation calculation between gold and investment credit interest rates, the LQ 45 index, in ation and exchange rates have a low correlation in that period (2018-2020). This means that gold is more of a safe haven than an investment diversi er, meaning that gold is an asset that is able to maintain the value it contains, not easy to depreciate.
For gold price expectations for the next year, based on the correlation results, it shows that the relationship between gold prices -interest rates and gold prices -exchange rates is strong complementary, while the relationship between gold prices and the LQ45 variable is very strong substitution or diversi cation.
For now, use gold as a haven. For the next year, gold assets can be replaced with investments in the real sector or can replace gold with deposits in the form of US$ currency or short-term securities. Figure 1 Analysis owchart