The non-linear market impact of large trades: evidence from buy-side order flow

We perform an empirical study of a set of large institutional orders executed in the US equity market. Our results validate the hidden order arbitrage theory proposed by Farmer et al. [How efficiency shapes market impact, 2013] of the market impact of large institutional orders. We find that large trades are drawn from a distribution with tail exponent of roughly 3/2 and that market impact approximately increases as the square root of trade duration. We examine price reversion after the completion of a trade, finding that permanent impact is also a square root function of trade duration and that its ratio to the total impact observed at the last fill is roughly 2/3. Additionally, we confirm empirically that the post-trade price reverts to a level consistent with a fair pricing condition of Farmer et al. (2013 Farmer, D., Gerig, A., Lillo, F. and Waelbroeck, H., How efficiency shapes market impact, 2013. Available online at: http://arxiv.org/abs/1102.5457 [Google Scholar]). We study the relaxation dynamics of market impact and find that impact decay is a multi-regime process, approximated by a power law in the first few minutes after order completion and subsequently by exponential decay.     

The immediate price impact of trades on the Australian Stock Exchange

We study the immediate price impact of a single trade executed in the Australian Stock Exchange (ASX). By ordering the top 300 stocks on the ASX in order of their free float market capitalization, a clear pattern emerges, with higher cap stocks experiencing lower price impact than lower cap stocks for the same traded volume. We investigate this relationship in detail, and show that the price impact and liquidity scale as a power of the market capitalization. This relationship is used to obtain a single market impact curve which shows average price shift as a function of volume traded. We obtain similar results for every year from 2001 to 2004.     

The impact of aggressive orders in an order-driven market: a simulation approach

This article investigates resiliency in an order-driven market. On basis of a vector autoregressive model capturing various dimensions of liquidity and their interactions, I simulate the effect of a large liquidity shock, measured by a very aggressive market order. I show that, despite the absence of market makers, the market is resilient. All dimensions of liquidity (spread, depth at the best prices and order book imbalances) revert to their steady-state values within 15 orders after the shock. For prices, a long run effect is found. Furthermore, different dimensions of liquidity interact. Immediately after a liquidity shock, the spread becomes wider than in the steady state, implying that one dimension of liquidity deteriorates, while at the same time, depth at the best prices increases, meaning an improvement of another liquidity dimension. In subsequent periods, the spread reverts back to the steady-state level but also depth decreases. Also, I find evidence for asymmetries in the impact of shocks on the ask and bid side. Shocks on the ask side have a stronger impact than shocks on the bid side. Finally, resiliency is higher for less-frequently traded stocks and stocks with a larger relative tick size.     

Performance-weighted ensembles of random forests for predicting price impact

For any large player in financial markets, the impact of their trading activity represents a substantial proportion of transaction costs. This paper proposes a novel machine learning algorithm for predicting the price impact of order book events. Specifically, we introduce a prediction system based on ensembles of random forests (RFs). The system is trained and tested on depth-of-book data from the BATS and Chi-X exchanges and performance is benchmarked using ensembles of other popular regression algorithms including: linear regression, neural networks and support vector regression. The results show that recency-weighted ensembles of RFs produce over 15% greater prediction accuracy on out-of-sample data, for 5 out of 6 timeframes studied, compared with all benchmarks. Feature importance ranking is used to explore the significance of various market features on the price impact, finding them to be highly variable through time. Finally, a novel procedure for extracting the directional effects of features is proposed and used to explore the features most dominant in the price formation process.     

New evidence on price impact of analyst forecast revisions

Previous research shows a positive relationship between consensus forecast revisions and stock returns in developed markets. We obtain new evidence from four major Asia-Pacific markets that suggest that abnormal returns are related to latest forecast revisions. The price impact of negative revisions is consistently stronger than that of the positive revisions. We also found considerable differences in price impact between developed and emerging markets for positive revisions, while no such difference is detected for negative revisions. The latest forecast revisions and category of analysts (those working in international broking firms) appear to be two key determinants of abnormal stock returns.     

The price impact of order book events: market orders,limit orders and cancellations

While the long-ranged correlation of market orders and their impact on prices has been relatively well studied in the literature, the corresponding studies of limit orders and cancellations are scarce. We provide here an empirical study of the cross-correlation between all these different events, and their respective impact on future price changes. We define and extract from the data the ‘bare’ impact these events would have if they were to happen in isolation. For large tick stocks, we show that a model where the bare impact of all events is permanent and non-fluctuating is in good agreement with the data. For small tick stocks, however, bare impacts must contain a history-dependent part, reflecting the internal fluctuations of the order book. We show that this effect can be accurately described by an autoregressive model of the past order flow. This framework allows us to decompose the impact of an event into three parts: an instantaneous jump component, the modification of the future rates of the different events, and the modification of the jump sizes of future events. We compare in detail the present formalism with the temporary impact model that was proposed earlier to describe the impact of market orders when other types of events are not observed. Finally, we extend the model to describe the dynamics of the bid–ask spread.     

Price impact of block trades: the curious case of downstairs trading in the EU emissions futures market

Using high-frequency data from the European Climate Exchange (ECX), we examine the determinants of price impact of €21 billion worth of block trades during 2008–2011 in the European carbon market. We find that wider bid-ask spreads and volatility are characterised by a smaller price impact. Larger levels of price impact are more likely to occur during the middle of the trading day, specifically the four-hour period between 11 a.m. and 3 p.m., than during the first or final hours. Purchase block trades induce a relatively smaller price impact on price run-up, while sell block trades exhibit a larger price impact on price run-up. We conclude that block trades on the ECX induce less price impact than in equity or conventional futures markets, and that a significant proportion of the effects contradict findings on block trades in those markets; thus, we provide the first evidence of the curious bent to block trading in the European Union emissions trading scheme.     

Private information flow and price discovery in the U.S. treasury market

Using intraday data, we identify the intensity of private information flow in the U.S. Treasury market. Our results show that the intensity of private information flow is highly correlated with public information shocks and higher for longer maturity bonds. More importantly, we find that bond price changes associated with high intensity of private information flow tend to be persistent, whereas those associated with low intensity of private information flow are more likely reversed. While public information and private information are the main determinants of bond price variations on days with news announcements, private information and liquidity shocks are important determinants of bond price variations on days with no significant events. Finally, we show that the depth of limit order book is inversely related to the intensity of private information flow. Nevertheless, informed dealers do not seem to use hidden orders to disguise their trading intentions.     

An information-based model of target stock price runup in the market for corporate control

The stock price runup of target firms in the market for corporate control has been anecdotally attributed to inside trading. Moreover, the empirical merger and acquisitions literature documents a time-varying level and duration of the stock price runup of target firms. Using a market microstructure approach, we model stock price runup as a stochastic process that shifts between a random walk without drift and a predictable process dependent on a parsimonious set of state variables. Consistent with the market microstructure literature, predictability in prices can be exploited only by the informed trader. The model is capable of explaining the complex stylized facts observed in stock price runup. It is also consistent with the merger wave literature, as we find that capital liquidity, economic growth, and market valuations drive the complex dynamics of stock price runup.     

Universal features of price formation in financial markets: perspectives from deep learning

Using a large-scale Deep Learning approach applied to a high-frequency database containing billions of market quotes and transactions for US equities, we uncover nonparametric evidence for the existence of a universal and stationary relation between order flow history and the direction of price moves. The universal price formation model exhibits a remarkably stable out-of-sample accuracy across a wide range of stocks and time periods. Interestingly, these results also hold for stocks which are not part of the training sample, showing that the relations captured by the model are universal and not asset-specific.

The universal model—trained on data from all stocks—outperforms asset-specific models trained on time series of any given stock. This weighs in favor of pooling together financial data from various stocks, rather than designing asset- or sector-specific models, as is currently commonly done. Standard data normalizations based on volatility, price level or average spread, or partitioning the training data into sectors or categories such as large/small tick stocks, do not improve training results. On the other hand, inclusion of price and order flow history over many past observations improves forecast accuracy, indicating that there is path-dependence in price dynamics.     

The impact of share price on seasonality and size anomalies in Australian equity returns

Given the high correlation between a firm's stock price and market capitalisation, it is possible that the well-documented size anomaly is masking a share-price effect. Using a seemingly unrelated regression model to accommodate contemporaneous correlation between portfolios, we estimate the separate effects of firm size and share price on returns to Australian equity portfolios. The analysis is also extended to estimate seasonal components of size and price effects. Our major findings are: (i) firm size and share price have significant and independent effects on portfolio returns averaged over all months, (ii) the familiar negative relation between size and returns is confirmed across all months, and (iii) the relation between share price and returns is negative in July and positive in all other months (with the exception of January where no price effect occurs). These findings, which are consistent across sub-periods and robust to method variations, highlight the need for future research to provide an economic foundation for the relation between average returns, size and price.     

The impact of business group affiliation on stock price informativeness: Evidence from an emerging market

This paper examines the relationship between business group affiliation and stock price informativeness in an emerging market setting. We use stock price synchronicity as a measure, and study the impact of group affiliation ‐specifically the extent of affiliation, ownership structure and existence of group bank‐ on firm specific information content. Results reveal that the amount of firm‐specific information capitalized into stock prices tends to be lower (higher) when the firm is group‐affiliated (unaffiliated), indirectly (directly) owned, and affiliated group has (does not have) a group bank. Additionally, the extent of group affiliation maintains a non‐linear relationship with synchronicity, suggesting that the perception of higher versus lower levels of group ownership differs.     

The price discovery role of day traders in futures market: Evidence from different types of day traders

Using proprietary account-level transaction data in the futures market where day traders are self-declared ex ante, this study investigates whether day traders enhance price discovery at the market level. From a natural classification of day traders, we find that heterogeneous day traders have differential effects on price discovery. Self-declared day traders, who benefit from low margin requirement, do not improve price discovery measured by information share. In contrast, non-declared traders, who are not self-declared as day traders, improve price discovery. Their positive impacts on price discovery are particularly significant during periods of high volatility and arrival of new information. Overall, a margin stimulating policy may encourage more day trading, but may also attract overconfident investors, especially inexperienced ones, and who do not enhance price discovery.     

Dynamic analysis of stock return volatility in an integrated international capital market

This paper examines the dynamic behavior of the stock return volatility for Canada, Japan, Germany, and the United Kingdom. The evidence indicates that international stock return volatility is mainly influenced by the U.S. stock return volatility and the exchange rate volatility, supporting the international capital market integration hypothesis. There seems to be some correlation between stock return volatility and macroeconomic volatility, but the effect is relatively weaker. In addition to the economic fundamentals, the noise component is found to be time varying, confirming the AR(MA)CH specifications in the stock return models.     

On the microstructure of price determination and information aggregation with sequential and asymmetric information arrival in an experimental asset market

Summary. Experiments were conducted on an asset with the structure of an option. The information of any individual is limited, as if only the direction of movement of the option value known for a single period without information of the value from when movement was initiated. However, if all information of all insiders were pooled, the value of the option would be known with certainty. The results are the following: (1) Information becomes aggregated in the prices as if fully informative rational expectations operated; and (2) The mechanism through which information gets into the market is captured by a path dependent process that we term The Fundamental Coordination Principle of Information Transfer in Competitive Markets. The early contracts tend to be initiated by insiders who tender limit orders. The emergence of bubbles and mirages in the markets are coincident with failures and circumstances that prevent the operation of the Fundamental Principle.The financial support of the national science foundation and the Caltech Laboratory for Experimental Economics and Political Science are gratefully acknowledged. The authors have benefited from helpful comments of David Grether, Kerry Back, Ivana Komunjer and Pete Kyle.