اطلاعیه های اقتصاد کلان، نوسانات، و روابط متقابل: بررسی نرخ بهره انگلستان و بازارهای سهام

This study investigates the intraday and daily pricing behavior of UK interest rate and equity index futures contracts. The paper initially examines the response of Short Sterling, Long Gilt, and FTSE100 to the release of scheduled macroeconomic announcements before employing dynamic time series techniques in order to reveal the nature of causal transmission patterns between these variables. In brief, short-term interest rates were found to be highly sensitive to indicators of prevailing economic conditions. However, the release of data important in the formation of inflationary expectations had a relatively subdued impact on long-term rates. Announcement effects appear somewhat ambiguous for the stock market. The analysis also reveals the bid–ask bounce and swift mean reversion in volatility to be important behavioral features of the return-generating process. Whilst the three variables appear to be bound by two cointegrating relationships, the tests for lead/lag relationships produce mixed results.

What information might be expected to move UK interest rate and equity markets, and to what extent do these markets respond in a similar manner to the arrival of news? Do these markets exhibit established empirical properties like volatility clustering and the bid–ask bounce around the time of announcements? Furthermore, what is the nature of the interrelationships within and across the UK interest rate and stock markets? In attempting to address these questions, this study conducts an empirical investigation into the intraday and daily pricing behavior of UK interest rate and equity index futures contracts. A large volume of empirical work has now identified a significant bond market reaction to macroeconomic news announcements. Whilst much of this analysis has been based on the US treasury market, studies of announcement effects in UK interest rate markets have also established similar reactions. For instance, significant movements in daily Short Sterling and Long Gilt prices have been attributed to the release of money supply information (see Dale, 1993, Goodhart & Smith, 1985 and Haldane & Read, 1999) and retail price index (RPI) data (see Goodhart & Smith, 1985 and Joyce & Read, 1999). Meanwhile, Brooke, Danton, and Moessner (1999) find that short-term and long-term interest rates are sensitive to a range of domestic and US announcements, including average earnings and RPI in the UK, and consumer price index (CPI), nonfarm payrolls, retail sales, industrial production, and gross domestic product (GDP) numbers in the United States. Studies using very high-frequency data also document a similar effect on the Short Sterling and Long Gilt around the time of news releases (see ap Gwilym et al., 1998 and Becker et al., 1995). In contrast to interest rate markets, the impact of macroeconomic news releases on equity markets appears somewhat mixed. For instance, Cutler, Poterba, and Summers (1989) find that, in most cases, information cited by the press as causing market movements in the United States was, in fact, quite unimportant. This reinforces the earlier studies of Hardouvelis (1987), Pearce and Roley (1985), and Schwert (1981), who all conclude that there is little evidence to suggest that the US stock market responds to macroeconomic news other than monetary information. On the other hand, the UK-based studies of Goodhart and Smith (1985) and Joyce and Read (1999) establish RPI releases to be significant for the UK equity market, and ap Gwilym et al. (1998) find that the FTSE100 is sensitive to RPI, producer price index (PPI), and public sector borrowing requirement (PSBR) data. Of US news items, Becker, Finnerty, and Kopecky (1996) report that PPI, merchandise trade, nonfarm payrolls, and CPI announcements are important for the FTSE100. This overview of existing research leads us onto the main aims of this paper. First, the only study that conducts a simultaneous examination of Short Sterling, Long Gilt, and FTSE100 responses to macroeconomic news releases was that of Goodhart and Smith (1985), which was based on the use of daily data and just four domestic announcements. In contrast, this paper examines tick data in 5-min intervals and a more complete array of potentially market-moving news releases. The application of high-frequency data in announcement studies is critical given that ap Gwilym et al. (1998) find that the adjustment of prices to news releases in the UK interest rate and equity index futures markets is completed within 3 min. Second, only the study of Buckle, ap Gwilym, Thomas, and Woodhams (1998) has attempted to model both UK interest rate and stock market return and volatility dynamics around announcement periods in a framework capable of capturing the stylized facts of financial asset behavior. However, their study excludes the Long Gilt from analysis and also fails to distinguish the relative importance of economic announcements. This study addresses both limitations. Finally, to our knowledge, no study has attempted to identify whether the three assets that are the focus of this paper are related not just in their reaction to news, but more generally in a longer-term comovement sense. Our analysis employs dynamic time series techniques in order to establish the nature of short-term and/or long-term causal transmission patterns between the Short Sterling, Long Gilt, and FTSE100. From a broader perspective, it is the contention of this paper, unlike many of its predecessors, to seek a generalized explanation of interest rate and stock market behavior. The current exercise should be of interest as, amongst other things, it investigates how UK markets respond to the arrival of information reflecting underlying economic conditions, and, in the process, this may shed light on whether traders responses are in accordance with widely accepted views about how the economy operates. The findings may also be of more general interest since the UK is considered to be a relatively large, open, and liquid market, situated in the financial ‘epicenter’ of Europe. In brief, the results suggest that traders distinguish between the information content of different news items. As expected, short-term interest rates are found to be highly sensitive to indicators of prevailing economic conditions. However, the release of data important in the formation of inflationary expectations had a relatively subdued impact on long-term rates. Announcement effects appear somewhat ambiguous for the stock market, in line with some of the earlier studies of the US equity market. The analysis also reveals that the bid–ask bounce and subdued volatility persistence are important behavioral features of the return-generating process. Whilst the three variables appear to be bound by two cointegrating relationships, the exact nature of the lead/lag relationships is unclear. The remainder of the paper is structured as follows. Section 2 presents a brief discussion on information effects and possible causal transmission patterns in interest rate and equity markets. Section 3 describes the features of the data, while the empirical design is presented in Section 4. We report the results and discuss their implications in Section 5. Section 6 concludes the paper.

This study has addressed a range of issues in investigating the intraday and daily pricing behavior of UK interest rate and equity index futures contracts. The major findings are as follows. First, the reaction of the markets to macroeconomic data suggests investors and portfolio managers distinguishes between the information content of different news items, in line with previous central bank studies in the United States (see Fleming & Remolona, 1997) and Australia (see Campbell & Lewis, 1998). We find that for all three markets, the announcement of changes in domestic monetary policy is the most important of the news items considered. Of the other announcements, short-term interest rates are found to be highly sensitive to indicators of prevailing economic conditions, such as changes in the US interest rates, retail prices, retail sales, and producer prices. We also document that the Long Gilt responds significantly to retail sales announcements, but much less to factors influencing inflation expectations. As for the FTSE100, traders appear to react swiftly to the figures in industrial production and PPI announcements, in addition to changes in domestic interest rate settings. Furthermore, the GARCH estimation of intraday returns and volatility demonstrate that both the Short Sterling and the Long Gilt are characterized by consecutive 5-min returns of opposite signs, which is suggestive of a bid–ask bounce, yet no such pattern emerged for the FTSE100. As is common in high-frequency return-generating processes, error term variance appears conditional on the information contained in the variance and volatility in the previous 5-min interval. In addition, volatility shocks display no real signs of persistence in any of the markets, which perhaps offers further support to the study of ap Gwilym et al. (1998), who finds that the adjustment of prices to news releases in both the UK interest rate and stock index futures markets is completed in less than 3 min. In a final examination of dynamic behavioral patterns in the Short Sterling, Long Gilt, and FTSE100 futures contracts, the results show that the three assets are bound by at least one, but most likely two, cointegrating relationships. The (within-sample) error correction analysis initially identifies the Short Sterling as the leading variable within the trivariate system. However, the variance decompositions suggest that beyond the sample period, the FTSE100 is actually the most exogenous variable, followed by the Short Sterling. Both procedures nevertheless confirm that the Short Sterling is exogenous and an important contributor to Long Gilt price behavior, consistent with the expectations theory of the term structure of interest rates.