Technological progress may be changing what we learn and how we trade

Laura Veldkamp, Maryam Farboodi 02 January 2019

a

A

Technological change is making it possible to process more and more information. In financial markets, this change has been accompanied by changes in trading strategies. Fundamental, value investing is on the wane, while strategies that make use of order flow data on others’ trades have flourished. Relative to their predecessors, today’s investors favour trading strategies that depend sensitively on how many buyers and sellers are eager to trade at that moment (Hendershott et al. 2011). What is the logical relationship between progress in financial information technology, trading strategies, and market efficiency? 

Why might information technology plausibly be related to trading strategies? If investors make optimal portfolio choices and learn about the same asset fundamentals over time, their investment choices should react to good news and bad news in a consistent, stable way. However, if an investor switches what they learn about, their investment choice should react to changes in the new variable. This change in trading pattern is a change in trading strategy. Put differently, if an investor wants to trade in a different way and still have their portfolio be optimal, they should acquire different information to support that new trading strategy. A technological change that affects information acquisition may therefore change trading strategies.

Yet, the relationship between trading strategy and information technology is not obvious. More data processing does not imply that the data must have a different composition. Perhaps when technology changes, everyone learns more about the same variables and strategies become more successful but not different in nature. 

In our paper (Veldkamp and Farboodi 2017), we take a standard noisy rational expectations framework and add two ingredients”

  • First, investors obtain a growing amount of processed data over time. 
  • Second, they can choose how much of that data they would like to be about the fundamental value of a firm, and how much will be about the non-fundamental demands of other investors. 

These two new ingredients are essential to explore changes in information technology and changing choices of trading strategies. We find that growth in the amount of data investors can process is a logical and predictable cause of an aggregate shift from fundamentals-based to order flow-based trading strategies. 

To understand how the volume of data affects data choices and thus trading strategies, consider how each form of data is used. The use of fundamental data is straightforward – if the data predict that the firm is more valuable than the price indicates, buy; otherwise, sell. This strategy generates some value, no matter what others do. 

The use of demand (order flow) data is more subtle. When an investor sees lots of uninformed buy orders coming in and learns that demand is high, she should sell because it is likely that the price is high, relative to fundamental value. This strategy looks like ‘trading against dumb money’. More charitably, standing ready to trade against agents who need to sell for non-fundamental (liquidity) reasons could also be called ‘market making’. Like our order-flow traders, market makers use order flow data to try and distinguish information-driven and non-information-driven trades, and to trade accordingly. An alternative way to describe this strategy is that the investors use their knowledge of price noise to remove that noise from the price signal, and to better extract the information that others know, from prices. The idea that investors can use order flow data to extract others’ information, or ‘follow smart money’, captures some of the flavour of front-running or trend-chasing. While following smart money and trading against dumb money sound very different, in this simple model, they are formally equivalent. What is crucial is that these strategies are only valuable when some investors are informed. There is no point in searching for dumb money, or trying to follow smart money, if everyone is dumb. As data technology improves and investors become better informed, following the informed orders and avoiding trades against informed order flows becomes paramount. 

This is why, when information technology is poor and quality data is scarce, it makes sense for investors to focus on learning about asset fundamentals. As financial information technology improves, it becomes more useful to identify the uninformed component of order flows to avoid being burned by the increasingly more informed fundamental traders.  Thus, a shift from fundamentals-based trading strategies to order-flow based trading strategies is a logical outcome of technological change. 

However, there is a limit to this shift. If every investor learned exclusively about non-fundamental demand, there would be no fundamental information in prices, and no fundamentally informed traders to avoid trading against. Order flow data would return to having no value. So, fundamental traders are essential to sustain order-flow strategies. The model predicts that, in the long run, the demand for fundamental and order flow data both grow proportionately with the growing amount of processed data. When data are sufficiently abundant, technological progress ceases to affect the composition of fundamental and order-flow traders. But in the transition, trading strategies should change. While this does not prove that technological changes are triggering observed shifts in strategies, it offers a way of thinking through the incentives to adopt each type of trading strategy, and how these incentives might depend on technology. Furthermore, despite the abundance of data in the long run, prices do not become fully informative and data stay valuable. The key insight is that data are a double-edged sword. They make prices more responsive to fundamentals not only today, but also in the future. The possibility of unpredictable large price collapses in the future creates risk today, and prevents the equilibrium from unravelling.

A model is a laboratory that allows us to see whether outcomes are harmful or beneficial. We consider two criteria to assess social benefit, or ‘efficiency’. 

  • The first efficiency measure is price informativeness, which is the sensitivity of prices to changes in future firm value. One might be concerned that as order flow trading grew and crowded out fundamental trading, prices might reflect less information. This does not happen. As technology improves, price informativeness rises. In this sense, markets become more efficient. 
  • The second measure is liquidity, or the price impact of an uninformed trade. One would expect that a strategy that looks like market making would enhance market liquidity. Yet, as order-flow trading becomes prevalent, liquidity stagnates. 

Finally, the model reveals empirical moments that can identify the strength of this mechanism. Specifically, the covariances between portfolio holdings and future earnings or current price noise can be used to infer data choices. The model predicts that order flow data demand will be higher in times and for assets where the signal-to-noise ratio in prices is higher. 

The study of how information is used in asset markets is still in its early stages. But sound logic supports the notion that technology is changing not only how much information we see, but also what kinds of information we choose to make use of. These changes in data processing choices rationalise new trading strategies, similar to those popular among financial managers today. More work remains to be done to determine the importance of this mechanism in today’s financial economy.

References

Hendershott, T, C Jones, and A Menkveld (2011), “Does Algorithmic Trading Improve Liquidity?", Journal of Finance 66: 1-34.

Farboodi, M and L Veldkamp (2017), “Long Run Growth of Financial Technology," NBER Working Paper 23457.

a

A

Topics:  Financial markets

Tags:  order flow data, technological change, trading strategies, financial markets

Professor at Columbia Business School

Assistant Professor of Economics, Bendheim Center for Finance, Princeton University

Events

  • 17 - 18 August 2019 / Peking University, Beijing / Chinese University of Hong Kong – Tsinghua University Joint Research Center for Chinese Economy, the Institute for Emerging Market Studies at Hong Kong University of Science and Technology, the Guanghua School of Management at Peking University, the Stanford Center on Global Poverty and Development at Stanford University, the School of Economics and Management at Tsinghua University, BREAD, NBER and CEPR
  • 19 - 20 August 2019 / Vienna, Palais Coburg / WU Research Institute for Capital Markets (ISK)
  • 29 - 30 August 2019 / Galatina, Italy /
  • 4 - 5 September 2019 / Roma Eventi, Congress Center, Pontificia Università Gregoriana Piazza della Pilotta, 4, Rome, Italy / European Center of Sustainable Development , CIT University
  • 9 - 14 September 2019 / Guildford, Surrey, UK / The University of Surrey

CEPR Policy Research