This research paper studies the shocks in the gold market and uses the econometrics of time series from Jan 2007 to Dec 2011, thereby, determining the fluctuation in gold prices in Pakistan using the ADF unit root test. The results give a clear view regarding the shocks in the Gold Market. The unit root test showed that the prices of gold (10gm) are not fixed and are to face wide changes. However, the first difference unit root had showed that the prices are not drifting and remained stagnant.
Key words: Gold, Unit Root Test
Introduction
Gold had been widely considered to be best for saving resources in long run with an expectation of strong purchasing power and profit at selling. As Gold is found worldwide, it is a source of exchange for money and is considered to be cash equivalent. People throughout Pakistan invested a lot in this metal and even earned profits. And because Pakistan is going through a market crunch these days, due to political distress, decrease in stock returns and negative economic indicators, people are keener to move towards gold market.
The concerns regarding the shocks in gold market increases due to this reason and for revealing the truths Augmented Dickey-Fuller Unit Root Test has been applied.
Literature review
In the paper by P K Mishra, J R Das and S K Mishra, is an attempt to analyze the causality relationship between gold price and stock market return based on BSE-100 index in India. For the purpose of study secondary data was taken from the data source of the Reserve bank of India. Average monthly data was collected for the period January 1991 to December 2009. To analyze Augmented Dickey-fuller unit root test and granger causality test have been used. In this time series data there is stationarity and also causality is running between two variables. One can be used to predict the other. (Mishra, Das, & Mishra, 2010)
In the article, by Young Kim, Jouahn Nam and Kevin J. Wynne, the relationship between gold prices and their affect on abnormal return to the company is explored and demonstrated. The SUR model is used to find out the positive correlation between firms implementing gold derivative positions and their lesser variable returns. This study examines the market response for both users and non users of god derivatives, when a shock in the prices is introduced in market place. Applying a Seemingly Unrelated Regression (SUR) Model, we find evidence that firms with gold derivative positions gain lower abnormal returns when gold prices rise. Correspondingly, when there is a negative shock on gold prices, non-gold derivative using firms experience much greater negative abnormal returns.
The study focuses on gold mining firms as a well-documented price fluctuation in this industry has been shown over time, considerably in 1997 and 1999.
Applying an SUR model, the authors provide supporting evidence that gold derivative using firms, which have lower long-term sensitivities to gold price changes, gain lower levels of abnormal returns due to their hedging position when gold prices went up. When gold prices fell, the gold derivative using firms experienced little to no change in firm value while the non-gold derivative using firms experienced negative abnormal returns. The result provides us with the evidence that indeed risk management strategies are priced in the market place and they do affect firm value as returns of the firms are highly dependent on the prices, and which are a result of the strategy adopted by that particular firm. (Kim, Nam, & Wynne, 2009)
The article by Wan-Hsiu-Cheng, Jung-Bin Su and Yi-PinTzou study about spot price effect but also the endogenize power and models which are used in this article are the combined BHK (Brenner, Harjes, and Kroner) and power GARCH (PGARCH).PGARCH and PGARCH-HV model is the best model to measure and forecast the value at risk. This article investigates the value at risk in gold market by considering both oil volatilities and flexible model construction. Article is representing the relationship between oil and gold that oil and gold prices raise and fall in synchronicity with each other. As the prices of oil rise and fall then gold prices react immediately and high volatility of oil plays an important role with regards to gold return. (Cheng, Su, & Tzou, 2009)
Empirical result shows the existence of co-integrations between variations in the gold price, oil price, exchange rate of dollar with various currencies, and the stock markets in various countries (Wang, Wang, & Huang, 2010). In their article Relationships among Oil Price, Gold Price, Exchange Rate and International Stock Markets use the data from Database of Taiwan Economy Journal (TEJ), Taiwan Bureau of Energy, Ministry of Economic Affairs, and cnYES. After applying the unit root test non stationarity in time series data exist, so to further test the data Co-Integration, and Vector Error Correction Model has been applied which results in significant, and after that Granger Causality test has been applied to check the cause and effect of variables among each other.
Hypothesis:
H1: The gold prices in Pakistan have the unit root.
Methodology
To explore non- stationarity in the gold market and gold prices of 10 grams in Pakistan, the data from 31, Jan 2007 to 31, Dec 2011 has been taken. The values of Gold prices have been taken from online site (forex.com).
Different econometrical tool can be used to investigate non- stationarity in the gold price, which includes PAC, ADF Unit Root Test, but in the following research ADF Unit Root Test has been used to check non- stationarity in the gold market of Pakistan. Augmented Dickey-Fuller Unit Root test also identifies that either shocks are temporary or permanent.
Results
Table1: Augmented Dickey-Fuller Unit Root Test
At level
At 1st Difference
t-Statistic
Prob.*
t-Statistic
Prob.*
Augmented Dickey-Fuller test statistic
0.875911
0.9945
Augmented Dickey-Fuller test statistic
-8.42377
0.0000
Test critical values:
1% level
-3.5504
Test critical values:
1% level
-3.5504
5% level
-2.91355
5% level
-2.91355
10% level
-2.59452
10% level
-2.59452
At Level
Coefficient
Std. Error
t-Statistic
Prob.
GOLDRATE (-1)
0.019953
0.02278
0.875911
0.385
D (GOLDRATE (-1))
-0.50332
0.142431
-3.53379
0.0009
D (GOLDRATE (-2))
-0.425644
0.140763
-3.02383
0.0038
C
605.5272
640.0128
0.946117
0.3484
At 1st Difference
Coefficient
Std. Error
t-Statistic
Prob.
D (GOLDRATE (-1))
-1.861542
0.220987
-8.423765
0
D (GOLDRATE (-1), 2)
0.394841
0.136006
2.903124
0.0053
C
1114.443
267.823
4.161119
0.0001
The focus of this study is to investigate non-stationarity levels of gold prices, and for this purpose ADF unit root test is applied at level and at first difference. Findings in table 1 reveal that at level ADF statistic value is 0.875911 which is greater than Mackinnon critical values at 1%, 5%, and 10% so we do not reject the null hypothesis which means that gold prices at level has a unit root.
t-1 +β1ðš«Yt-1 +ETt Equation 1
0.01995 -0.5033
[0.875] [-3.533]
α and β1 in the equation 1 are the parameters of Yt-1 and ΔYt-1 and are estimated 0.01995 and -0.5033. Here α< 1 i.e. α = 0.01995 at t = 0.875 which means that data has non-stationarity and shocks are temporary at level.
t-1 +β1ðš«Yt-1 +ETt Equation 2
-1.86 0.394
[-8.4237] [2.9031]
ADF test statistic at 1st difference is -8.4237 and is smaller than the Mackinnon critical values at 1%, 5%, and 10% in equation 2 which means that data has now become stationary and the null hypothesis is rejected and gold prices has no unit root. Furthermore here α < 1 which means that at 1st difference shocks are temporary.
Conclusion
In conclusion we can say that shocks are temporary in this time series data of gold rates, which means that fluctuations in the gold prices occurs temporary. Here we used monthly data in the given time series from January 2007 to December 2011 and Augmented Dickey-Fuller Unit Root Test is applied to depict the non-stationarity in data and after applying the test we get to know that for a certain time period shocks does occur in the gold prices but this trend is only temporary.
