Lebanese University

Faculty of Economics and Business

Administration

Subject: Microstructure of Financial Market

Dr. Latifa Ghalayini

Prepared by
Haifa Mansour
Master2 – Research Accounting
2023-2024
0
The Objective of this paper is to delve into the various factors that influence market
liquidity using a simulation model of an artificial market. The authors survey the
definition of market liquidity and discuss the relationship between market liquidity
and market efficiency or stability. They consider a continuous auction market and
discuss factors affecting market liquidity.

1. Definition of market liquidity:

In the conceptual section, we encounter diverse interpretations of the market’s
definition, shaped by the specific context in which it is employed.
In retrospect, we can see that Keynes and Hickschose ”Market liquidity” as a
subject in economics, but the terminology they used consisted of phrases
such as the ”Future volatility of market prices” or the “Possibility of immediate
execution of a transaction.” When discussing whether or not a market
is liquid, Bagehot focused on factors such as the existence of adverse selection
effects due to information asymmetry, the price impact of a trade, and the portion
of trading cost which is set according to the pricing policy of the market maker.
Since the cost of losing market liquidity is large despite it simplicity nature, the
improvement and stability of market liquidity is not only important for market
participants, but also serves as a way to enhance financial market stability.
In the simulation, we assume a certain market structure and focus
on the parameters which affect an individual market participant's ordering decision,
such as expectation of the asset price, confidence in such expectation, the extent of
risk aversion, sensitivity to various kinds of market information.
Specifically, a decline of market price uncertainties accompanied by an increase in
market liquidity will, through the decline of liquidity premiums such as the bid-ask

1
spread and market impact, improve market efficiency, resulting in
efficient fund and risk allocation.
The collapse of a system, or the emergence of systemic risk, will be caused by the
market coming to a halt or by the loss of market participants' faith in the market
price discovery function.
As tautological as it may sound, a halt in the endogenous market price
discovery function depends on whether or not market participants who try
to avoid the risk of market halt form a majority.
Behavior to avoid the risk of market halt such as the reduction of market exposure
by closing positions is considered to materialize when market conditions including
price levels and the speed of changes in the price level have exceeded specified limit
values, and there will be various processes by which the system collapses.
The characteristics of boundary and collapse processes depend on market
participants' expectations of or confidence in the market itself, which are generated
from market behavior under normal conditions including a small degree of stress.
The maintenance of sufficient liquidity under normal conditions will autonomously
improve market stability by expanding the market boundaries and improving the
participants' confidence in market sustainability.

2. Static and dynamic aspects of market liquidity

In a market with no bid-ask spread or market impact, and quite a large

market resiliency, the trade execution price will be identical to the market price at
the time of trading, and there the market participants need not worry about
price uncertainties in the execution of trades.

Past studies of market liquidity have mainly focused on static aspects and have
adopted indicators which show static market depth, such as turnover, bid-ask
spread, as measures of market liquidity.
2
In order to examine how market liquidity affects the price discovery function in an
actual market, not only should the static aspects of market
liquidity be examined, but also the dynamic.
The indicators to explore market liquidity conditions:
➢ Probability of quote existence
➢ Trade frequency
➢ Price volatility
➢ Bid-ask spread
➢ Gross order book volume (buying order volume plus selling order volume)
➢ Net order book volume (buying order volume minus selling order volume)

3. Conceptual summary of the mechanism through which market

microstructure affects market liquidity.
One of the purposes of this study is to clarify conceptually the process by which
actual market microstructure affects market liquidity.
Price discovery process (2 stages):
a) Micro stage: from when market participants hold potential trade needs based
on their individual reasons to when the actually decide to place orders in the
market.
b) Macro stage: the stage in which such orders are accumulated in the market
and trades are executed.

3
3.1 Decision-making process of market participants at the micro stage
There are various channels through which traders' decision-making processes are
influenced by the market microstructure.
If there is some institution or set of rules, such as an accounting rule and a tax
system, which induce certain effects on the decision-making process of entering
the market and placing trade orders, the market microstructure will alter effective
supply and demand in the market and finally affect market liquidity.
Premiums against the risks caused by the time difference among the markets or by
the time lag between trade and settlement based on the T+ n rules in the securities
market, and the development of hedging devices such as options and futures, are
also given costs which are unavoidable for market participants and thus can be
regarded as factors which affect the magnitude of trading costs in the market.
With respect to the relationship between the difference in ordering methods
and market liquidity, there have been many theoretical studies which showed that

4
the number of limit orders affects the extent of market liquidity, or serves as an
indicator of the state of market liquidity.

3.2 price discovery process at the macro stage

Certainly! This passage discusses the price discovery process in markets, focusing
on how decisions at the micro level (individual trades) impact the macro level
(overall market prices). It highlights the role of different execution systems (like
double auctions or market maker systems) in influencing market liquidity and
information distribution.
Key points include:
➢ Micro decisions influence macro structures: Individual trades affect overall
market dynamics.
➢ Execution systems play a crucial role in information collection and
distribution, impacting market liquidity.

5
 ➢ Market transparency is vital: How information is shared affects decision-
making at both micro and macro levels.
➢ Different execution systems, like continuous double auctions or market maker
systems, differ in their information handling, impacting market efficiency.
➢ Three key types of information drive market decisions: executed trades,
existing trade needs, and potential trade needs.
Ultimately, the passage emphasizes how information collection and distribution
mechanisms within execution systems significantly impact the market's price
discovery process and overall efficiency.

4. Frame work analysis

The passage discusses using a Monte Carlo simulation to understand how traders
make decisions regarding market liquidity. The analysis involves two key parts:
a) Individual Trader Model: This part simulates how individual traders decide
on their orders based on various factors like their perception of the asset's
value, confidence in that perception, risk aversion, and sensitivity to market
feedback.
b) Order Flow Aggregation Model: It models how these individual orders
aggregate and execute within the market system, considering unique
parameters for each trader.
4.1 The simulation process includes:
➢ Conducting multiple periods (M periods) within a market with N traders.
➢ Each trader places one order per period in a random order sequence.
➢ Orders get executed based on a 'first-in' rule, similar to the Tokyo Stock
Exchange system.
➢ Traders make decisions by analyzing historical data, forecasting future market
conditions, considering their portfolio and risk preferences, and using two
types of trading strategies: short-term profit based on price trends and long-
term based on fundamental analysis.
6
 ➢ The simulation output provides detailed data on executed orders, trade
volumes, execution prices, and remaining unexecuted orders. This data helps
in understanding and analyzing trader behavior within the market
environment.

The analysis focuses on understanding how altering parameters, like traders'

composition and their expected values' dispersion, affects market behavior. While
real market participants adapt behavior based on visible variables, like market
conditions, the analysis uses fixed figures for invisible variables, like risk aversion,
to observe how changing visible factors relates to these constant invisible ones.
4.2 A trader’s decision- making model
In this model, traders operate within a continuous auction system resembling the
Tokyo Stock Exchange. There are two types of traders:
➢ value traders, who place limit orders based on expected values derived from
information beyond market data,
➢ momentum traders, who make market orders based on short-term market
trends.

7
For short-term price forecasting, traders rely on historical market data to predict
future market prices. They analyze price movements, mean price trends, and
volatility to forecast the market price for the next trading opportunity. Additionally,
traders incorporate volume information from the order book, considering gross order
volume (sum of limit orders for selling and buying) as a measure of market depth
and net order volume (imbalance between selling and buying orders) to gauge
pressure on market prices.
Traders who consider volume information modify their expectations of market price
trends and volatility accordingly. For instance, changes in gross and net order
volumes influence their forecasts, helping them anticipate potential market
movements or increased price risk.
Ultimately, this model integrates historical price data and volume information to
assist traders, whether they are value traders placing limit orders or momentum
traders following short-term market trends, in making informed decisions about their
trading strategies.

Trading behaviour of value traders

In value trading, traders combine short-term price forecasting with subjective
information about an asset's fundamental value. They use an assumed lognormal
distribution for aggregated expected values in the market, with a standard deviation
(σV). Each trader forms their expected value, initially provided but not representing
the true value. They base their orders on this expected value and the distribution of
short-term price expectations (Pt+f).
8
When placing limit orders, traders establish order prices (Pask for selling, Pbid for
buying) to maximize expected returns (E(RPask) or E(RPbid)). This involves
multiplying the return of the order by the probability of execution within the
timeframe before the next ordering opportunity.

For a selling limit order: Maximum E(RPask) = Probability(Pt+f ≥ Pask) × (Pask - P)

For a buying limit order: Maximum E(RPbid) = Probability(Pt ≤ Pbid) × (P - Pbid)

Traders acknowledge their uncertainty in forecasts as risk in placing orders. When
the expected return significantly outweighs the risk (given as the expected dispersion
of expected values), traders will place a selling or buying order with a higher
expected return.
Trading behaviour of momentum traders
Momentum traders focus on market trends and volatility, not absolute prices. They
use historical data to forecast short-term trends. Their decision to execute market
orders relies on comparing the expected earnings ratio (μ') to volatility (σ'). If μ' / σ'
is less than -B, where B is a positive constant reflecting a trader's risk aversion, they
sell in a falling market. If μ' / σ' is greater than B, they buy in a rising market. This
strategy allows them to trade based on trends and volatility, regardless of specific
price levels.
• If the ratio of expected earnings to volatility (μ' / σ') is less than -B, they
execute a selling market order when prices are falling.
• If the ratio of expected earnings to volatility (μ' / σ') is greater than B, they
execute a buying market order when prices are rising.

5. Results of analysis
The outlines simulations and analyses conducted on an artificial market model with
various trader types, exploring their impact on market behavior and liquidity
indicators.
Simulation Results Summary:

9
 ➢ Composition of Traders:
Increasing momentum traders reduces liquidity supply, raising liquidity absorption.
Probability of quote existence falls, trade frequency rises initially.
Diminishing trends observed past a certain level of momentum traders.
Volatility of execution price and mean quote increases initially, then stabilizes.
Liquidity indicators like spread and gross order volume decline but reach a lower
limit.
➢ Dispersion of Expectations and Traders’ Confidence:
Probability of quote existence and trade frequency change based on the difference
between expected and true variance.
Volatility of execution price and average spread influenced by true variance more
than expected variance.
Market liquidity linked to the underestimation/overestimation of risks by traders.
➢ Extent of Traders’ Risk Aversion:
Decreasing risk aversion for value traders increases liquidity indicators.
Price volatility rises with higher risk aversion for value traders, but decreases beyond
a certain level.
➢ Sensitivity to Order Book Information:
Higher sensitivity to order book information improves quote existence and order
book volume.
However, trade frequency, spread, and market resiliency decline with increased
sensitivity.

10
 Conclusions and Further Research:
Trading methods impact liquidity; more market order-based trading reduces
liquidity.
Traders' underestimation/overestimation of risks affects market depth and resiliency.
Lower risk aversion increases liquidity and reduces price volatility.
Higher sensitivity to order book info improves some liquidity indicators but worsens
others.
Further research is needed to explore feedback effects from market data on traders'
decisions and to incorporate more sophisticated market structures into simulations.
This study offers insights into how various trader behaviors and market conditions
impact liquidity indicators, emphasizing the importance of understanding traders'
risk perceptions and responses to market information in shaping market dynamics.

11