Signing trades and an evaluation of the Lee–Ready algorithm

We use techniques developed to analyze the Supply Curve in liquidity models in order to analyze the accuracy of the Lee and Ready algorithm, both for highly liquid and relatively liquid stocks. Through the use of order book data combined with tick data, we are actually (somewhat tediously) able to tell whether or not a given trade is buyer or seller initiated. For those trades where such knowledge is certain, the accuracy of the Lee and Ready algorithm is not as accurate as has been assumed previously. We can essentially prove that the Lee and Ready algorithm is always at least 55% accurate, and is around 61% accurate for highly liquid stocks (i.e., the top 50 of the S&P 100).     

Trade signing in fast markets

This study assesses the accuracy of trade signing algorithms in fast trading environments using NASDAQ and NYSE trade and quote data. Using data that contain true trade signs, we show that the Lee and Ready algorithm outperforms the tick rule and classifies trades at least as well as in earlier studies from slower trading environments, even in subsamples where the market is particularly fast. We conclude that trade signing remains viable in fast markets, and that the use of quote data continues to increase trade classification accuracy.     

A test of the accuracy of the Lee/Ready trade classification algorithm

We analyze the accuracy of the Lee/Ready (1991) trade classification algorithm and the tick test. Our definition of true trade classification is based on whether the Makler (the equivalent of the specialist on the Frankfurt Stock Exchange) bought or sold shares. The Lee/Ready method classifies 72.8% of the transactions correctly. The simpler tick test performs almost equally well. We document that misclassification of trades may systematically bias the results of empirical microstructure research. Finally, we show that estimation of the bid–ask spread from transaction data results in a reasonably accurate estimate of the relative liquidity of our sample stocks. This is an important finding because quote data for the German stock market is not available on a regular basis.     

Anonymity,Stealth Trading,and the Information Content of Broker Identity

This paper examines whether the identity of a broker involved in transactions contains information. Using a sample of transactions from the Australian Stock Exchange—where broker identity is transparent—we provide evidence that consecutive buyer‐/seller‐initiated transactions by the same broker have a relatively high permanent price impact. This implies that broker identity conveys information to market participants, and that markets in which broker identity is disclosed are likely to be more efficient. We also find that medium‐sized trades by the same broker convey greater information than large and small trades, which is consistent with stealth trading by informed investors.     

Tick test accuracy in foreign exchange ECN markets

Recent computer quoting activity has increased the allure of the tick test because the quote rule and its variants require matching asynchronous trade and quote records. We find tick test accuracy of 1.2 million forex trades is about 67% which falls to 63% for zerotick trades (half the sample). Accuracy declines as quoted spreads decrease and as time to the previous trade increases. We observe extreme asymmetry for midquote changes, where buyer accuracy is 96% (27%) for up (down) changes, respectively. The quote rule is about 77% accurate. The group tick test is superior to the bulk volume classification method.     

Transaction duration and asymmetric price impact of trades—Evidence from Australia

This paper investigates the asymmetric price impact of buyer and seller initiated trades and the informational role of the trade duration. Using trade data from the Australian Stock Exchange (ASX), our results indicate that buyer initiated trades increase the ask price more than the bid price, and seller initiated trades decrease the bid price more than the ask price. The transaction duration is modeled in a Box-Cox ACD framework. The unexpected portion of the duration is found to play a more significant role in causing price impact in both purchases and sales than the expected duration. A trade shortly after the previous trade results in higher price impact than one after a long period. We found evidence that increased trading activity is associated with larger price impact, therefore implying a higher degree of information-based trading.     

Inferring investor behavior: Evidence from TORQ data

We use a unique dataset (TORQ) to calibrate several techniques commonly used to infer investor behavior from transactions data. Specifically, we evaluate the Lee–Ready (1991) Journal of Finance 46, 733–746) algorithm for distinguishing trade direction, and we examine the use of trade size as a proxy for trader identity. We find that, due to complexities in the NYSE auction process, up to 40% of reported trades cannot be unambiguously classified as either buyer- or seller-initiated. However, for those trades that can be classified, the Lee–Ready algorithm is 93% accurate. In addition, we construct a firm-specific trade size proxy that is highly effective in separating the trading activities of individual and institutional investors. These findings should be useful to researchers interested in inferring trader sophistication and buy/sell direction from transactions data.     

On the occurrence and consequences of inaccurate trade classification

The validity of many economic studies hinges on the ability to properly classify trades as buyer or seller-initiated. This study uses the TORQ data to investigate the performance of the Lee and Ready (1991, Journal of Finance 46, 733–746.) trade classification algorithm. I find that the algorithm correctly classifies 85% of the transactions in my sample, but systematically misclassifies transactions at the midpoint of the bid–ask spread, small transactions, and transactions in large or frequently traded stocks. I then provide evidence of the biases induced by inaccurate trade classification.     

Trade classification algorithms for electronic communications network trades

Ellis et al. [Ellis, K., Michaely, R., O’Hara, M., 2000. The accuracy of trade classification rules: Evidence from Nasdaq. Journal of Financial and Quantitative Analysis 35 (4), 529–551] find that trade classification rules have limited success in classifying trades which execute inside the quotes. We reconfirm this result and propose an alternative algorithm to improve the classification accuracy for trades inside the quotes. This alternative algorithm improves the overall success rate for classifying trades, especially for trades that occur inside the quotes. Additionally, we show that the Lee and Ready [Lee, C., Ready, M., 1991. Inferring trade direction from intraday data. Journal of Finance 46, 733–747] and Ellis et al. (2000) trade classification algorithms provide biased estimates of the actual effective spreads and price impacts, while our algorithm provides statistically unbiased estimates of actual effective spreads and price impacts.     

Block Trade Price Asymmetry and Changes in Depth: Evidence from the Australian Stock Exchange

This paper examines the price response to large block transactions on the Australian Stock Exchange during the 1999 sample period. We find asymmetry in the price reaction between buyer‐ and seller‐initiated trades with respect to size and resiliency following the trade. We extend previous research by examining order book changes surrounding block trades and relating price effects to changes in book depth. Purchases are associated with persistent order book imbalance, while the sales imbalance is insignificant. Cross‐sectional analysis demonstrates that price resiliency following a trade is related to the speed at which limit orders arrive to replenish book depth.     

Integrating auction and search markets: The slow Dutch auction

The issue of choosing to sell property by auction or by traditional negotiated search markets is addressed in this article. A general selling institution called the slow Dutch auction is introduced. This general selling mechanism reduces to either a conventional auction, a posted offer, or some time dependent mix of these selling institutions depending on the pricing rule chosen by the seller. We model search by having potential buyers whose private valuation for the property is unknown to the seller arrive randomly over time. With this general framework the seller's problem is to choose a selling mechanism that maximizes expected wealth. Surprisingly, we find that the optimal selling institution is always a posted offer market. The seller chooses an optimal posted price and waits until a buyer arrives who is willing to pay this price. Auctions are never optimal.     

Microstructural dynamics in a foreign exchange electronic broking system

This paper explores the relationships between quotations, spreads and transactions in the Foreign Exchange market. Such interactions have been the subject of much work in markets such as the NYSE, but until now have gone unexamined in the FX market owing to a lack of data. Using a 7 hour, transactions-based data set we examine the determinants of both quote revisions and spreads. The results indicate that trades are a major factor in spread determination and quote revisions. Furthermore, there is evidence that the widely documented negative auto-correlation in quote returns is at least partially caused by the ‘thinness’ of this particular segment of the FX market.     

Short sales,long sales,and the Lee–Ready trade classification algorithm revisited

Asquith et al. (2010) conclude that short sales are often misclassified by the Lee–Ready algorithm. The algorithm identifies most short sales as buyer-initiated, whereas the authors posit that short sales should be overwhelmingly seller-initiated. Using order data to identify true trade initiator, we document that short sales are, in fact, predominantly buyer-initiated and that the Lee–Ready algorithm correctly classifies most of them. Misclassification rates for short and long sales are near zero at the daily level. At the trade level, misclassification rates are 31% using contemporaneous quotes and trades and decline to 21% when quotes are lagged one second.     

The impact of trading halts on liquidity and price volatility: Evidence from the Australian Stock Exchange

This study examines market behaviour around trading halts associated with information releases on the Australian Stock Exchange, which operates an open electronic limit order book. Using the Lee, Ready and Seguin (1994) pseudo-halt methodology, we find trading halts increase both volume and price volatility. Trading halts also increase bid-ask spreads and reduce market depth at the best-quotes in the immediate post-halt period. The results of this study imply that trading halts impair rather than improve market quality in markets that operate open electronic limit order books.     

Locals, foreigners, and multi-market trading of equities: Intraday evidence from Thailand

We study stock market orders and trades in a developing country, Thailand, where foreign ownership limits partially segment local and foreign investors into two distinct markets. Some foreigners forgo voting rights and distributions to trade on the “local board”, while some locals forgo such benefits and pay a price premium to trade on the “foreign board”. Regardless of nationality, these cross-market traders typically submit orders when liquidity is high, fill orders at relatively beneficial prices, exploit patterns in stock prices across markets, display profitable holding-period returns, and enhance price discovery. This suggests that skilled, informed trading that affects market quality does not depend on trader nationality.     

Do cosmetic reporting variations affect market behavior? A laboratory study of the accounting emphasis on unavoidable costs

Eight laboratory market sessions are conducted in which sellers have both opportunity costs (default redemption values from not trading) and unavoidable costs (fixed outlays that must be paid irrespective of trading). The treatment variation is whether sellers’ intermediate accounting reports compare sales revenues to unavoidable costs (four sessions) or default redemption values (four sessions). Findings indicate systematically higher seller asks and buyer bids in the condition where accounting emphasizes unavoidable costs. However, the price and efficiency implications of these higher asks and bids are offset by simultaneous shifts in the relative frequencies of seller-initiated and buyer-initiated trades.     

The anatomy of buyer–seller dynamics in decentralized markets

In this paper I investigate the nexus between buyer–seller dynamics, financial frictions and market efficiency in decentralized markets. To do so, I introduce financial frictions in a dynamic market with heterogeneous traders. Heterogeneously constrained buyers sequentially enter the market to acquire units of a generic good from heterogeneously endowed sellers. I characterize two closely related classes of equilibria, respectively called homogeneous equilibrium with no entry (HEWNE) and homogeneous equilibrium with entry (HEWE). Both equilibria prescribe a market where only the efficiently endowed type of seller exists in the limit. However, the two equilibria diverge in the specification of agents’ behavior subsequent to trade. In HEWNE, sellers and buyers exit the market upon successful trading. In HEWE, like in supply chains, in every period certain types of buyers replace exiting sellers, thus becoming potential sellers for subsequent waves of buyers. First, I identify the critical role of frictions in steering the complex evolution of market heterogeneity for both classes of equilibria. Secondly, I operationalize the combined study of HEWNE and HEWE to obtain sharp predictions on market efficiency for a range of empirically-relevant situations in which buyer–seller dynamics are decoupled, for example when entry of new sellers is delayed or stopped. Third, I test the theoretical findings against a simulated artificial market.     

Ripples through markets: Inter-market impacts generated by large trades

This paper uses a sample of large trades executed on the London Stock Exchange's SEAQ-I market for European cross-traded firms to investigate their impact on home market prices when parallel markets suffer from information frictions. I find that (a) large London trades produce price impacts in home markets even though no timely information is published, (b) market makers appear to pre- and post-position their inventories by splitting orders across markets, and (c) the price discovery process across markets changes significantly around large trades with the foreign market making a significantly bigger contribution to price discovery at this time, even though information opaqueness exists.     

NYSE order flow,spreads, and information

This paper uses unique NYSE audit trail data to evaluate spreads and information content for different order types. Actual spreads are positive for liquidity-demanding orders and negative for liquidity-supplying orders after controlling for order direction. However, because a large fraction of liquidity-demanding orders get price improvement, the actual spread for liquidity-demanding orders is up to 50 percent less than the Lee and Ready (1991) algorithm would suggest. Regression results show that the order composition of trades affects traditional measures of spreads and information. They also show that NYSE non-displayed liquidity reduces trading costs facing market orders, and that liquidity-demanding floor broker orders are the most informative order type.