1. Real-time price discovery in global stock, bond and foreign exchange markets (by T. G. Andersen, T. Bollerslev, F. X. Diebold, C. Vega) Olesia Kiiashko Kseniya Bortnikova Khrystyna Pastukh Ksenia Pogodina

2. OUTLINE 1. Introduction 2. High-frequency return data 3. Asset returns and macroeconomic surprises: basic results 4. A generalized specification 5. Summary and directions for future research

3. Traditional “efficient markets” thinking (asset prices should completely and immediately reflect movements in fundamentals) Asset prices and fundamentals may be largely and routinely disconnected 1. Introduction How is news about macroeconomic fundamentals incorporated in pricing stocks, bonds and foreign exchange? Macroeconomic fundamentals - are statistics that indicate the current status of the economy of a state (Interest Rates Announcement, GDP, CPI).

4. Price discovery question: Price discovery question: 1). Previous researches: the connection only for a single asset class and country (e.g., Balduzzi et al., 2001, study the U.S. bond market), multiple asset classes but only a single country (e.g., Boyd et al., 2005, who study U.S. stock and bond markets). multiple countries but only a single asset class (e.g., Andersen et al., 2003b, study several major U.S. Dollar exchange rates). 2). In this paper authors examine multiple countries and asset classes.

5. Scientists methodology: Simultaneously combining: 1) high-quality and ultra-high frequency asset price data across markets and countries  to study price movements in (near) continuous time; 2) a very broad set of synchronized survey data on market participants' expectations,  to infer “surprises” or “innovations” when news is announced; 3) advances in statistical modeling of volatility, which facilitate efficient inference.

6. 2. High-frequency return data Object of research - a unique high-frequency* futures dataset Investigation the response of U.S., German and British stock, bond and foreign exchange markets to real-time U.S. macroeconomic news * High-frequency financial data are observations on financial variables taken daily or at a finer time scale (computer technology and data recording and storage have made these data sets accessible to researchers).

7. Reasons for using futures market data 1) Futures prices are easily available on a tick-by-tick* basis 2) Most significant U.S. macroeconomic news announcements are released at 8:30 Eastern Standard Time (EST) when the futures markets are open, but the equity markets closed 3) Transaction costs are lower in the futures markets, and the contracts are very actively traded Focus on price adjustments measured over very short time intervals Futures markets tend to lead cash markets in terms of price discovery *Tick - minimum change in price a security can have, either up or down.

8. High-frequency returns around macroeconomic news announcements News announcement regressions are based on the period ranging from ten minutes before to one-and-a-half hours after an announcement Macroeconomic news does move the markets The summary statistics confirm the usual rank ordering in terms of volatility, with stock markets being the most volatile, followed by foreign exchange rates, and then fixed income. All results reported below are based on five-minute local currency continuously compounded returns

9. report Unconditional sample correlations report The table provides a sense of the comovements among the asset markets during news announcement times U.S. macroeconomic news affects the Dollar and the equity market in the same direction

10. The positive bond and equity market cross- correlations explanations (during U.S. macroeconomic news announcement) Stock market prices around the world would tend to be positively correlated if the discount rate is the dominant effect or if the domestic and foreign real output and/or monetary shocks are positively correlated Although the high positive contemporaneous correlation across countries may be explained by the common bond market response to U.S. macroeconomic news, a number of other influences, including market microstructure, contagion, and cross-market hedging effects, could also account for the high-frequency correlations.

11. Correlations in expansions vs. contractions Whether the interactions among asset returns and their responses to macroeconomic fundamentals vary systematically across the business cycle? the unconditional correlations separately for the expansion period the contraction period the stock-bond correlations are positive and small the stock-bond correlations are negative and large explanation the cash flow effect dominates during contractions while the discount effect dominates in expansions (due to central bank policy) data for 1 period cannot claim as a general pattern

12. 3. Asset returns and macroeconomic surprises: basic results С haracterization of the dynamic effects of U.S. macroeconomic news announcements for each of our nine asset markets 1) Describing the measurement of macroeconomic news 2) Estimating the responses of asset returns to news Data used: MMS (Money Market Services) real-time data Period from January 1, 1992 through December 31, 2002 The units of measurement obviously differ across the macroeconomic indicators

13. Dynamic news effects Response equations estimation using the two five-minute returns directly preceding and the eighteen five-minute returns following each announcement Methods used: 1) OLS consistent (but the disturbance terms for the five-minute return regressions are heteroskedastic) 2) WLS consistent+efficient 3 markets: Domestic and foreign bond markets, Foreign exchange markets Domestic and foreign equity markets.

14. Estimation results. Bond Markets and Foreign exchange markets Bond Markets Immediate reaction positive real shocks and inflationary shocks produce lower bond prices ( higher yields ) Strongest effect U.S., because U.S. macroeconomic fundamentals also significantly impact the foreign bond markets Foreign Exchange Markets Immediate reaction positive domestic real shocks lead to an appreciation of the Dollar (during the recent contraction regime) U.S. inflation rate news do not systematically affect the foreign exchange rates

15. Estimation results. Equity markets. Full sample almost no significant responses Splitted sample (expansions vs contractions) positive real economic shocks are met with a negative response in expansions and a positive response in contractions. negative impact of inflation during expansions, because of the presence of stronger anti- inflationary monetary policies the responses occur almost instantaneously and tend to “drown” in the overall day-to-day price movements it is crucial to focus on the high-frequency returns just around the announcement time.

16. A closer look at stock–bond correlations state dependence in the equity markets news reaction the time-invariant reaction of the bond markets state dependent stock–bond market correlations positively correlated during expansions negatively correlated during contractions the existence of time-varying cross- market dependence daily realized correlations monthly realized correlations correlation patterns are driven by the macroeconomic news releases? correlation patterns simply indicate the general relationship among the markets? on the nonfarm payroll announcement

17. Daily realized correlations on the nonfarm payroll announcement January 1, 1992 - December 31, 2002 The shaded area corresponds to the U.S. contraction sample from March 1, 2001 to December 31, 2002

18. Monthly realized correlations on the nonfarm payroll announcement January 1, 1992 - December 31, 2002 The shaded area corresponds to the U.S. contraction sample from March 1, 2001 to December 31, 2002

19. 4. A generalized specification Why to implement the generalized specification? - so far immediate news reaction is observed ( synchronous high-frequency data around the time of the announcement)  mitigates other influences & potentially omitted variables biases  simultaneous equations model The model is identical to already presented model, except that the contemporaneous asset returns are included

20. 4. A generalized specification Problem: without any additional restrictions or modeling assumptions, contemporaneous coefficients, that capture cross-asset linkages and/or spillover effects are not identified, not explained by news announcements Solution: applying QMLE techniques to the multivariate GARCH model Estimate the model for 3 markets at a time

21. Estimation Results 1) Trivariate domestic system (U.S. T-Bond, S&P500, and $/ Euro returns) 9301 five-minute announcement at period 1(expansion) All coefficients have natural interpretation, however, in contrast to previous results(positive correlation),we have two opposing effects: bond prices affect stock prices positively, but stock prices affect bond prices negatively. 6463 five-minute announcement at period 2(contraction) Qualitatively as before. But! For S&P500: has opposite sign positive stock returns, ceteris paribus, always raise interest rates, while positive innovations to interest rates have a strongly regime- dependent impact on stock returns

22. Estimation Results 2) Interdependence among the national stock markets estimated coefficients are positive (expansion sample)  important cross-country linkages over-and-above those explained by the U.S. macroeconomic news releases. estimating the same relations over the more recent contraction period suggests even stronger contemporaneous cross-country linkages.

23. 5. Summary and directions for future research What was done? Authors characterized the real-time interactions among U.S., German and British stock, bond and foreign exchange markets in the periods surrounding U.S. macroeconomic news announcements. What was found? Announcement surprises produce conditional mean jumps high-frequency stock, bond and exchange rate dynamics are linked to fundamentals.

24. Bad macroeconomic news has the traditionally-expected negative equity market impact during contractions, but a positive impact during expansions Distinct correlation patterns are not limited to the period around announcements; rather, they apply generally for trading day returns in expansions and contractions Generalized estimation approach documents highly significant contemporaneous cross-market and cross- country linkages, even after controlling for macroeconomic announcement effects Important direct spillover effects among foreign and U.S. equity market, that let observe the interaction of actively- traded financial assets around announcement times The discount rate effect dominates in each of the three stock markets during U.S. economic expansions, coupled with the U.S. interest rate playing the role of the “world interest rate” and hence dictating the move of the foreign bond markets.

25. Directions for future research Using high-frequency data to quantify the three separate channels of private information, contagion, and public information that link the markets Recent studies Brandt and Kavajecz (2004) Evans and Lyons (2003, 2005) Pasquariello and Vega (2004) Increased the role of order flow in the price formation process Combining the information in order flow and other liquidity measures in concert with the new statistical procedures and rich high-frequency return and news announcement data

26. Thank you for your attention!