Macroeconomic News and Price Discovery in Indonesian Government Bond Market

The study of macroeconomic news impact on government bond gets little attention, especially in emerging markets. Andritzky et al. (2007) and Nowak et al. (2011) study this impact for some emerging countries, but little attention given to Asian countries. The question about whether macroeconomic news have impacts on government bond is important, considering the large amount of government bonds outstanding in Indonesia and the importance of regulation to maintain the stabilization of bond price. This research use daily returns of Indonesian government bond benchmark series over five-year period to investigate the impact of domestic and global macroeconomic news announcements. We study the relationship using event study approach. Following common literature we use surprise component of macroeconomic news announcements, which will be defined as the difference between market expectation and the actual release of the macroeconomic news. We use economic forecast survey conducted by Bloomberg as the proxy of market expectations needed to calculate domestic (Indonesia) and global (US) macroeconomic news announcements surprises. We find that, for bond returns, surprises of global macroeconomic news announcements is more important than domestic ones, especially for recent years, while both surprises of global and domestic macroeconomic news announcements affect bond returns volatility. Keyword: Bond pricing, Bond returns, Macroeconomic news announcements, emerging market

with investors who invest in government bonds. They will value financial and non-financial information about the government of the countries so they can adjust their expectation toward the bonds.
Many studies examine the relationship between domestic or global macroeconomic factors and government bond yield or spread, such as Rozada and Yeyati (2008) or Erba et al. (2013). The studies find that the relationship is significant. These studies usually use a long period data as macroeconomic data usually available for monthly at the shortest and yearly at the longest. They cannot capture the immediate reaction of the market when the new information arrived. When new information arrived, investors immediately adjust their expectations and at that time the new value or price of the investments is being formed. This paper tries to capture the impact of the macroeconomic news announcements on Indonesian government bond price and volatility. Previous studies, which are few, mainly focus in advanced markets. Studies in emerging markets by Andritzky et al. (2007) and Nowak et al. (2011) did not include Indonesia in their sample. Indonesian government bond size has increased tremendously after the economy crisis in 1998 and the secondary market is developing fast, show the importance of studying Indonesian government bond price and volatility.
The paper is organized in six parts -introduction, review of literature, data and data resources, methodology, results and discussion, and conclusions and recommendations.

Literature Review
Literature about impact of macroeconomic news announcements mainly focus on stocks or currencies returns instead of bonds or government bonds. And many of those researches focus on advanced markets and very few on emerging markets. Early researches capture the impact of macroeconomic news announcements on bond returns by comparing returns on days when there were macroeconomic news announcements and days when there wasn't, such as Fleming and Remolona (1999) and Andritzky et al. (2007). Recent studies use 'surprise' effect of macroeconomic news announcements and use forecast survey from Money Market Services (MMS) International or Bloomberg as proxy of market expectation, such as Balduzzi et al. (2001), Goeij and Marquering (2006), Brenner et al. (2009), Nowak et al. (2011), and Paiardini (2014. Balduzzi et al. (2001), using intraday data of US Treasury Bill and Treasury Note, find that several announcements affects bond price, but in different manner according to maturity of the bond. Rigobon and Sack (2006) use new technique to handle noise in news found higher significance of the news on US Treasury bond returns. Brenner et al. (2009) study the impact of macroeconomic news on US Financial markets find that the news surprises have significant impact on all assets type including government bond, and especially sensitive for 'good' news surprise. Outside US, Paiardini (2014) study the effects of macroeconomic news on Italian government bond. The study use domestic, neighboring countries, regional and global macroeconomic news surprises. The results show 25 of 68 news types has significant impact on Italian government bond returns. Only 1 domestic news is significant and mostly US news has significant impact. Andritzky et al. (2007) and Nowak et al. (2011) study macroeconomic news effects on emerging bond markets. Andritzky et al. (2007), using GARCH (1,1) and dummy variables to indicate days with news, find that domestic macroeconomic news has no significant effect on emerging bond markets spreads at the mean level, but the US interest rate news has. Nowak et al. (2011) find that global and regional macroeconomic news are the same important as the domestic news. US macroeconomic news that has large surprise found to have large effect compare to same surprise size of domestic news.
Several studies also try to capture the news effect on bond returns volatility. Balduzzi et al. (2001) find that news has significant and persistence increases on bond returns volatility. Goeij and Marquering (2006) find that macroeconomic news surprises have significant impact on daily returns of US Treasury bonds. This impact on volatility is difference between several news types as some news has longer impact and other have shorter impact. Brenner et al. (2009) find that news have significant impact and that conditional bond return volatility increase before the news released and decrease afterward. In emerging markets, both Andritzky et al. (2007) and Nowak et al. (2011) find that news have significant impact on market volatility and Nowak et al.
(2011) also find that the volatility is more drawn-out compared to advanced markets. Most previous studies indicates the significance of news on government bond returns. Studies outside US always use US news as global news indicator, and find that the news significantly affects government bond markets outside US. Most news types also has significance on government bond return volatility, with emerging market volatility tend to be more drawn-out compared to mature markets.

Data and Data Resources:
This study using daily average price in secondary market to get daily returns on benchmark series of Indonesian government bond. Average price is used because there is no data information about the hour and minute when the transaction done. Transactions mainly done through over the counter and then reported to the Bank Indonesia as the central registry for all Indonesian government bond. The daily return is calculated as follow: Where Rt is daily return at time t of a benchmark series of Indonesian government bond. Pt is price of a benchmark series of Indonesian government bond at time t. Pt-1 is price of a benchmark series of Indonesian government bond at time t-1.
The new information in macroeconomic news is proxied by 'surprise' following common literature by Balduzzi et al. (2001), Nowak et al. (2011) and Paiardini (2014). This surprise is the difference between actual macroeconomic news and the expectation of the market. For the expectation of the market I use Bloomberg forecast survey to market analysts which is also used in many previous studies such as Nowak et al. (2011) and Paiardini (2014). This survey results is released prior to the scheduled date of the news. The surprise is calculated as follow: is standard deviation of forecast error of macroeconomic news k. We use standardized surprise following Balduzzi et al. (2001)  transactions data for each series. Although the data is available for longer period, but the transaction activities will be different when a series appointed as benchmark or not, so the data after 1 year is not comparable.
Data of news dates, actual macroeconomic news, and market analyst survey was obtained from Bloomberg. There

Methodology
This paper attempts to examine the impact of surprises from macroeconomic news on Indonesian government bond price and volatility. Using the time series technique, tests of stationarity and multicollinearity needed. After the data become stationary and free of multicollinearity, equation ( Where Rt is daily return at time t of a benchmark series of Indonesian government bond, Rt -1 is lagged t-1 return of a benchmark series of Indonesian government bond Sk,t is surprise macroeconomic news k at time t. To know the impact of the macroeconomic news on returns volatility of Indonesian government bond, we apply GARCH (1,1) following Andritzky et al. (2007), but we use surprise variables instead of dummy variable in the equation. GARCH (1,1) was developed by Bollerslev (1986) to allow the conditional variance to depend on its previous own lags. We use 'Bollerslev-Wooldridge Heterokedasticity Constant Covariance' if the residuals is not following normal distribution, to avoid bias on standard errors estimations.
Where ht is daily returns variance of a benchmark series of Indonesian government bond, ℎ −1 is one lagged daily returns variance of a benchmark series of Indonesian government bond and −1 2 is one lagged of residual squared.

Results and Discussion
The result for estimation of equation (4.1) can be found below for Indonesian government bonds period 2014.

Conclusion and Recommendations
This paper study impact of the macroeconomic news on Indonesian government bond price and volatility.
Using daily returns of benchmark series of Indonesian government bond period 2010-2014 as dependent variable and macroeconomic news surprises as independent variables, we find that domestic news impact is very little, except for the surprise of BI rate news. The global news have no or little impact between 2010-2012 and starting to have impact in 2013, and become dominant in 2014. While for the impact on volatility, the results show that both domestic and global news have impact on bond returns volatility, but with inconsistent manner over time. This means the market response is changing overtime in processing the new information. These results are useful for regulators, investors, and academicians. The suggestion for the regulators are to be aware of macroeconomic news that affect the government bond market, so that immediate actions can be taken if the market about to become unstable. It is also suggested to create and maintain the regulations for foreign investors' participants in the market. The regulators also need to encourage investors to do transactions in the exchange than through OTC as this may lead to better system and data of secondary market transactions. Future researches can be done by using intraday data with information about precise time the transactions happen, which is currently not available, to see more comprehensively about the instantaneous market reactions to macroeconomic news surprises.