توقف معاملات روزانه در بازار معاملات آتی نیکی

The Osaka Securities Exchange (OSE) halts Nikkei 225 index-futures trading when the next transaction is to take place at a price more than ¥30 (prior to February 1994) or ¥60 (from February 1994) away from the previous trading price. This paper examines the efficacy of the intraday price limit rule in terms of price discovery, liquidity and volatility. We also include transaction data from the Singapore International Monetary Exchange (SIMEX) where Nikkei futures are traded simultaneously. The intraday price limit rule generally appears to be ineffective in reducing volatility and avoiding price jumps, at least partly because OSE traders have access to the alternative market at SIMEX.

The trading systems of Japanese securities exchanges differ strongly from the main other markets around the world. Most importantly, Japanese markets do not employ specialists, such as the ones on the New York Stock Exchange, who continuously provide liquidity to limit the magnitude of price changes between consecutive transactions. Instead, Japanese systems supply through a limit order book, managed by so-called saitori members. Saitori members solely function as middlemen in transactions between regular members of the exchange. They cannot trade for their own account, nor can they accept orders from the public. A special feature of Japan's exchanges is that liquidity is organized through restrictions on maximum price changes from trade to trade. Thus, immediacy is sometimes sacrificed by the trading system in order to slow down trading when there is an order imbalance and to generate liquidity by actively soliciting counterorders with other exchange members. At least two studies have looked at this intraday price limit rule on the Tokyo Stock Exchange (TSE). Lehmann and Modest (1994) study transactions, best-bid and best-offer quotes for all individual stocks on the TSE over a period of 26 months. They find that intraday price limits are hit only very rarely, occurring less than once every 4 days. In general, they conclude that the TSE's trading system functions well. Hamao and Hasbrouck (1995) performed an extensive study of stock prices using transaction data from the TSE. They conclude that immediacy at the TSE appears to be “quite high” and the intraday price limit rule causes some trades to be executed more slowly, but allows many market orders to transact at better prices. Many articles study the impact of other types of trading halts on price discovery in an effort to provide answers in the debate in policy circles on the use of trading halts in financial markets, especially since the market breaks of 1987 and 1989. Proponents argue that price limits reduce overreactions, while others suggest that price limits prevent trading altogether and therefore harm the price discovery process. Ma et al. (1989), for example, study the influence of daily price limits on the price formation process after the market has resumed trading. They conclude that price limits serve a useful function in giving the market “time to breathe”. Lee et al. (1994) study circuit breakers on the New York Stock Exchange and find that trading halts are associated with increased volume and volatility, which persist for 1 day and 3 days after reinstatement, respectively. The authors therefore conclude that trading halts are not successful at fulfilling their goal of reducing “excess volatility”. These results support their claim that the trading halt disrupts “learning by trading”. Kryzanowski and Nemiroff (1998) investigate the price discovery process around trading halts on the Montreal Exchange. They find that volatility and trading activity increase significantly around trading halts and return to lower levels in less than 2 days after the resumption of trading. However, the halts on the Montreal Exchange appear to be less of an impediment to uncertainty resolution than halts on the NYSE (see Lee et al., 1994). Kim and Ghon Rhee (1997) study daily price limits on the Tokyo Stock Exchange and conclude that price limits are ineffective. Berkman and Steenbeek (1998) study the influence of daily price limits on the price formation process of the Nikkei 225 stock index futures contract. The Nikkei 225 futures contract is traded on both the Osaka Securities Exchange (OSE), a market with strict price limits, and the more lenient Singapore International Monetary Exchange (SIMEX). The authors concentrate on the relative price, trading volume and volatility on these two markets and relate these variables to the distance of the actual price from the price limit. No evidence is found of a “gravitation effect”, which states that the price of a security is drawn towards the price limit. The results do provide support for the “satellite-market” model described in Subrahmanyam (1994): if the transaction price gets closer to the price limits, price variability and market liquidity decrease on the dominant market (the OSE) relative to the satellite market (SIMEX). In this paper we study the efficacy of the intraday price limit rule on the market for Nikkei 225 stock-index futures in Osaka (hereafter: Nikkei futures). Until February 1994, the Osaka Securities Exchange (OSE) would halt trading whenever the price moved up or down by more than ¥30 within 6 minutes. As a result, the OSE halted trading frequently and for longer periods, because the 6-minute intervals overlapped whenever prices moved substantially up or down. After the rule was changed, prices were allowed to move by ¥60 every 3 minutes, causing the rule to be triggered only in very volatile periods. An important feature of this paper that makes it different from existing studies on intraday price limits is the presence of an alternative market, the Singapore International Monetary Exchange (SIMEX), where Nikkei futures are traded simultaneously and in similar design. Trading on SIMEX continues whenever the intraday price limit rule triggers a trading halt at the OSE. With the exception of Berkman and Steenbeek (1998) who study daily price limits in the Nikkei futures market, all existing studies focus on assets for which there is no main alternative trading venue to the one for which trading halts are studied. The SIMEX transaction data during OSE trading halts allow us to investigate whether OSE traders actually make use of the alternative trading venue when their own market is closed. If this is the case, then the next logical question is whether the trading halts at the OSE are effective at all in reducing price volatility. The availability of SIMEX transaction data provides us with a continuous time series to study price volatility before, during and after trading halts, whereas existing studies do not have any price information during trading halts. We exploit this opportunity to estimate a GARCH model for the intraday Nikkei futures returns and use the results to study volatility around trading halts. The results indicate that first, until February 1994, the intraday price limit rule led to an excessive number of trading halts on the OSE. In about 50% of the cases a trading halt was followed by another trading halt within 15 minutes. This may well have caused many OSE traders to move at least part of their trading from the OSE to SIMEX as between the old (strict) intraday price limit rule and the new (more lenient) intraday price limit rule, SIMEX' market share increased from less than 5% just before the rules were tightened in September 1991 to over 55% just before the second rule change in February 1994, in terms of the number of contracts traded. Second, trading activity increases around trading halts, but soon returns to normal levels after trading resumes. For the latter part of our sample (after the gain in market share for SIMEX) longer trading halts result in a significant increase of between 11% and 15% in SIMEX trading activity compared to just prior to the halt, presumably due to OSE traders trading on SIMEX. Third, the futures price in Osaka on average makes a jump when trading resumes after the halt. This indicates that trading halts frequently lead to price discontinuities, instead of smoothing the price discovery process. Fourth, conditional volatility remains high for at least the first 30 minutes after the trading halt. This either indicates that trading halts lead to an increase in volatility, or that they are ineffective at decreasing volatility. Hence, in contrast with earlier findings on the TSE's trading system, the presented evidence leads to the conclusion that the intraday price limit rule in its current form does not function the way it was intended on the Nikkei futures market. Abolishing the rule would have a positive influence on price discovery process as well as on the volume share on the OSE. The remainder of this paper is organized as follows. Section 2 describes the trading mechanisms of the OSE and SIMEX. Section 3 elaborates on the dataset and descriptive statistics. Section 4 discusses the methodology and presents the empirical results. Finally, Section 5 concludes.

In this study we look at trading activity, price discovery and volatility in Nikkei futures at SIMEX and the OSE around Osaka trading halts. The OSE changed the intraday price limit rule triggering trading halts in February 1994. Prior to this date, the intraday price limit rule was such that it was frequently triggered and potentially lasted very long. The new intraday price limit rule, still applied today, is much more lenient and leads to a substantially lower amount of trading halts.The old rule led to an excessive number of trading halts, without yielding the desired reduction in volatility. In fact, in approximately 50% of the cases was a trading halt followed by another trading halt within 15 minutes. Under the new rule in 1994r1995 we find SIMEX activity to increase during OSE trading halts, especially when trading halts last longer than 2 minutes. Hence, it appears Žand it is certainly the case today. that at that time many OSE members were a member of SIMEX as well. As such, although the new trading rule is much more flexible, many OSE members now trade on SIMEX when the trading halts occur under the new rule. This renders the new intraday price limit rule more or less ineffective in reducing volatility, and the OSE has lost a substantial amount of its volume share to SIMEX. We even find that the conditional volatility of SIMEX Nikkei futures prices is significantly higher in the first minute after a trading halt than it is in the final minute prior to the trading halt. In terms of price discovery the Osaka price jumps when trading resumes. Hence, trading halts interrupt the continuous path of prices, rather than enhancing it. Of course, on the other hand, it is true that trading often resumes at a price closer to the previous trading price, which improves the continuity of prices. In general, however, the presence of SIMEX substantially reduces any potential effects the OSE might have hoped for. SIMEX does seem to underreact during non-trading, prices rising Ždecreasing. only half as much as the final change when trading resumes within 1 minute after that. In general the OSE and SIMEX prices are in line again within 2 minutes after trading resumes. Over the past years, more and more financial markets have changed their Aold-fashionedB trading systems into modern electronic trading systems. However, the weakness of electronic trading systems remains its handling of very volatile markets. The OSE’s AsolutionB to the problem, the intraday price limit rule, does not seem to be effective, partly because of the presence of an alternative market in Singapore. Abolishing the rule would have a positive influence on price discovery process as well as on the volume share on the OSE.