OFR Paper Models Impact of Market Structure on Prices Under Stress

Concern has focused recently on the apparent fragility of market liquidity, which refers to the ability to buy and sell securities with a minimal price impact and is essential for markets to operate efficiently. Although liquidity may seem adequate under normal conditions, it seems to disappear abruptly in times of stress.

An OFR working paper released today suggests that sharp price declines may be exacerbated by funding constraints of large dealers and other market-makers and by the slow reaction of liquidity providers such as insurance companies. These factors were apparent during recent incidents such as the Flash Crash on May 6, 2010, and the extreme volatility in the U.S. Treasury market on October 15, 2014.

Understanding the relationship between sharp price declines and liquidity remains a challenge, although researchers are making progress. Liquidity events are rare and data from non-crisis periods are not necessarily meaningful for predicting behaviors during crisis periods. The task is further complicated by the increasing dominance of automated algorithmic trading systems.

Rather than extrapolate from non-crisis to crisis periods, the OFR researchers built a model that incorporates dynamics among market participants into the process of assessing liquidity.

The model uses an agent-based framework that can be calibrated to mirror the behavior of market participants, including algorithmic traders. Agent-based models specify rules that dictate how agents will act in different situations. In this paper, the agent-based model is centered on the actions of three types of market participants: liquidity demanders (hedge funds, for example), market-makers (dealers), and liquidity suppliers (insurance companies, pension funds, and asset managers). The decision cycles of market participants vary between the liquidity demanders and suppliers, and the market-makers’ capacity and willingness to hold inventory, which may be impaired during times of crisis.

Two components determine the price impact of the demand for liquidity: the ability or willingness of the market-makers to take on inventory, and the time frame for the liquidity suppliers to take the other side of the trade. In simulations of the model, the absence of either a market-maker or a liquidity supplier causes an increase in market impact, with the effect becoming increasingly severe as the liquidity demand increases.

The model is targeted toward the dynamics of individual market participants and can be tailored to specific conditions. This type of model can contribute to our understanding of market liquidity by capturing characteristics of liquidity specific to crisis periods. It can also help regulators and policymakers test the effects of potential future policy changes. The next step will be to use market data to calibrate the model to fit observed behaviors of market participants.

The OFR is committed to understanding the causes and effects of fragile market liquidity. Another recent OFR working paper explored patterns that connect daily liquidity conditions across a broad range of financial markets to financial conditions in the aggregate, both in normal times and under stress. This research represents one potential framework for monitoring liquidity to extract signals of impending disruptions.

The OFR is also developing a monitor using a suite of statistical tools that measure the breadth, depth, or resiliency of markets using inputs such as price and volatility (see the OFR’s 2014 Annual Report, pages 30-35).

Greg Feldberg is Acting Deputy Director for Research and Analysis at the Office of Financial Research