Motivation

● Exploit marketing inefficiency: utilize the delayed reactions to earnings information to identify
undervalued or overvalued stocks
● Leverage psychological biases: take advantage of the common investor behaviors, such as the
underreaction or overreaction to new financial data
● Active risk management: enables active traders to manage risks dynamically, responding swiftly to
market changes and adjusting positions accordingly

Importance of Intraday Trading Strategies

● Enhance returns: offers opportunities to achieve substantial returns, crucial for traders looking to
maximize gains from market movements
● Enhance marketing efficiency: contributes to correcting mispricing in the stock market more
quickly, helping the market reflect true asset values more accurately
● Diversification of trading strategies: diversify trading approaches and reduce risk

1 | Earnings Momentum Trading Strategy

 Marketing Microstructure

Algorithmic Trading's Role: it has transformed how traders respond to earnings announcements in the
stock market. In the A-share market, where individual retail investors have traditionally been a dominant
force, the rise of algorithmic trading has introduced a new dynamic.

Earnings Momentum Strategies: Strategies that capitalize on earnings momentum aim to leverage the
predictive power of earnings surprises. In A-share markets, it is both a signal that triggers algorithmic
trading responses and a metric around which momentum strategies are built. As A-share markets
mature and algorithmic trading becomes more prevalent, the interplay between these strategies and
market microstructure will likely become more pronounced

2 | Earnings Momentum Trading Strategy

Dataset Acquisition and Preprocessing

● Data Source: Tushare

The data used for analysis originates from Tushare, a comprehensive financial data provider
offering a vast array of datasets covering various financial instruments, including stocks, bonds,
funds, and macroeconomic indicators in China. Specifically, we obtained the CSI 300 index data
from Tushare's API, which provides historical and real-time financial data for Chinese securities.

● Data: CSI 300

The CSI 300 is a capitalization-weighted stock market index designed to replicate the
performance of the top 300 stocks traded on the Shanghai Stock Exchange and the Shenzhen
Stock Exchange. It has two sub-indexes: the CSI 100 Index and the CSI 200 Index. Over the
years, it has been deemed the Chinese counterpart of the S&P 500 index and a better gauge of
the Chinese stock market than the more traditional SSE Composite Index.
The time range for the data is from January 1st, 2010 to December 31st, 2021. We used data
from 2010 to 2012 to collect signals, started our backtest from 2012, and ended in December
2021.
3 | Earnings Momentum Trading Strategy
Data Analysis

● CSI 300 Index was 3,584.27 as of

2024-04-26, Historically, CSI 300
Index reached a record high of
5,877.20 and a record low of 818.03,
the median value is 3,337.58.
Typical value range is from 3,138.74
to 4,521.48. The Year-Over-Year
growth is -9.45%.

● During 2010-2021, the mean value

of CSI 300 Index is 3410.47, with a
standard deviation equals to 603.37.

4 | Earnings Momentum Trading Strategy

 Technical Analysis

SUE (Standardized Unexpected Earnings)

● Calculation:
- EPSt: The actual earnings per share reported by the company for period t.
- E(EPSt): The expected earnings per share for the company for period t. According to the information, it is
calculated using the earnings per share from the same quarter of the previous year.
- Unexpected EPS: This is the difference between the actual EPS and the expected EPS for period t
- σ(EPSt−E(EPSt)): The standard deviation of the unexpected EPS over the previous quarters, which
provides a measure of variability or volatility in the earnings surprise.
● Capture significant deviations in earnings that are likely to influence stock prices, hence providing a trading
signal

5 | Earnings Momentum Trading Strategy

 Technical Analysis

Trading Signals Based On SUE

● Stocks are ranked based on their SUE scores
● Buy signals are generated for stocks with high SUE values, and sell signals are generated for stocks
with low SUE values
● Leverages the hypothesis that stocks with unexpectedly high earnings compared to their last year’s
performance (high SUE) will see price increases

Dynamic Universe of Stocks:

The trading universe is dynamic, updated quarterly based on the available data and the
specific conditions like earnings announcements and quarter changes, which again is based on the
structured schedule of financial quarters.

6 | Earnings Momentum Trading Strategy

 Trading Strategy Development

The development of the trading strategy described leverages intraday market data and combines
elements of quantitative analysis, primarily using the SUE (Standardized Unexpected Earnings) metric
and volatility measures like Beta, to execute trades and manage hedge.

1. Data Utilization:
Intraday Market Data: our strategy uses high-frequency intraday data, which includes stock prices
(open, high, low, close) and volume. This data type is crucial for capturing market movements and
allows the strategy to react quickly to market changes
Earnings Data: this includes earnings per share (EPS) data, the dates of earnings announcements
and end date, which are used to calculate SUE. SUE reflects how actual earnings compare to
expectations, providing a signal for potential price movements

7 | Earnings Momentum Trading Strategy

 Trading Strategy Development

2. Calculation of SUE:
Expected and Unexpected EPS: SUE calculation involves determining expected EPS (often based
on past performance) and comparing it to actual reported EPS to find unexpected earnings.
Standardization of SUE: the unexpected earnings are then standardized by the standard deviation
of unexpected earnings over a relevant historical window, making it a more reliable indicator by
normalizing the magnitude of earnings surprises.

3. Beta Calculation for Risk Management:

Volatility and Systematic Risk: Beta is calculated to assess the systematic risk of a stock relative
to the broader market. It measures how much a stock’s price is expected to react relative to market
movements.
Use in Strategy: this measure helps adjust the trade sizes based on the stock's risk profile,
integrating a layer of risk management into the strategy.

8 | Earnings Momentum Trading Strategy

 Trading Strategy Development

4. Signal Generation and Trade Execution:

Buy/Sell Signals: The strategy uses SUE to generate buy signals for stocks with high positive SUE
and sell signals for those with negative or low SUE, hypothesizing that stocks with positive surprises
will gain in value.
Dynamic Adjustment: The strategy is designed to dynamically adjust to new data, recalculating
SUE and Beta as new earnings information and price data become available.

9 | Earnings Momentum Trading Strategy

 Backtesting & Performance Analysis
Backtrader Environment Setup:
Cerebro Engine: Initializes the main engine to set up, run, and evaluate the trading strategy.
1) Starting cash: 1,000,000 CNY
2) Commission: 0.01%
3) Slippage: 0.05%

Data Preparation for Backtesting:

CSI 300 Data: Uses historical CSI 300 stock data from the past 10 years
Data Cleaning:
1) Filter symbols based on data availability
2) Align data with trading calendar
3) Clean and normalize data
Data Loading: Loads historical data from pandas DataFrames into Backtrader using the
‘PandasDirectData’ feed.

10 | Earnings Momentum Trading Strategy

Backtesting & Performance Analysis

Running the Backtest:

Execution: Simulates trades using the ‘My Strategy’ logic over the historical data
Logging: Records initial and final portfolio values to gauge the strategy’s performance

Portfolio Performance Analysis:

Extra Key Metrics: Utilize PyFolio and Drawdown analyzer to get portfolio performance metric
● Profit and Loss (PnL): Calculates the net profit or loss by comparing the ending vs.
starting portfolio values
● Maximum Drawdown: Measures the largest peak-to-trough drop during the backtesting
period as a key risk indicator
Comprehensive Report: Create a comprehensive performance report using Quantstats
library, including detailed statistics
Shareable HTML format

11 | Earnings Momentum Trading Strategy

Performance Metrics

Sharpe Ratio 0.71 Cumulative Return 26.39%

Maximum Drawdown -6.67% Sortino Ratio 1.07

Win-Loss Ratio 1.29 Calmar Ratio 0.31

CAGR 2.04% Expected Shortfall (cVaR) -0.45%

Volatility (annualized) 4.44%

12 | Earnings Momentum Trading Strategy

Performance Metrics

13 | Earnings Momentum Trading Strategy

Performance Metrics

14 | Earnings Momentum Trading Strategy

Performance Metrics

15 | Earnings Momentum Trading Strategy

 Optimization

Enhanced Trading Rule for Earnings Announcements:

To ensure that trading decisions are based on the most accurate and timely financial data. It checks
whether the earnings reports are not only announced but also correspond to the fiscal periods relevant to
the current trading strategy.

1. Identify Key Parameters and Rules

● Announcement and End Dates: Ensure the earnings reports used for trading decisions are for the
intended fiscal quarters to avoid trading based on outdated or irrelevant information.
● Exclusion Criteria: Define specific rules to exclude certain periods, such as not trading based on
fourth-quarter earnings reports, which might contain year-end adjustments that could skew analysis.

16 | Earnings Momentum Trading Strategy

 Optimization

2. Validate Timing and Relevance of Earnings Reports

● Daily Checks for Earnings Reports: Implement checks to identify if an earnings report has been
published on the current or previous day, ensuring timely trading actions.
● Fiscal Quarter Confirmation: Use the end_date from the earnings report to confirm that the
earnings are indeed for the relevant fiscal quarter, which ensures the strategy is acting on accurate
and relevant data.

3. Adjust Trading Strategy Based on Data Accuracy

● Handling of Multiple Announcements: Adapt the strategy to handle instances where multiple
earnings announcements might occur on the same day, prioritizing the most recent based on the
end_date.
● Avoidance of Ambiguous Data: If the ann_date and end_date do not align properly, the trade is
not executed, preventing decisions based on potentially misleading information.

17 | Earnings Momentum Trading Strategy

 Results
Performance Against Benchmarks

● Average Monthly Return: Our strategy realized an average monthly return of 0.4% from 2012 to
2019, which aligns with the benchmark range of 0.2% to 0.63% noted in Chen et al. (2009).

18 | Earnings Momentum Trading Strategy

 Results

Influence of SUE on Portfolio Returns

● High vs. Low SUE Stocks: Monthly returns differ within our portfolio depending on whether stocks
have high or low Standardized Unexpected Earnings (SUE).
● Holding Period: Returns improve progressively with longer holding periods, which affirms that an
extended duration correlates with higher returns.

Market Impact on Profitability

● Earnings Momentum: Profitability decreases when the holding period extends from 3 to 12 months.
● Seasonal Variation: The strategy experiences a drop in profitability around mid-year, with negative
returns typically observed in July and August.

19 | Earnings Momentum Trading Strategy

Comparison with Existing Strategies

Strategy CSI300

Sharpe 0.71 -0.42

Ratio

Maximum -0.07 -0.11

Drawdown

Win-Loss 1.29 1.08

Ratio

20 | Earnings Momentum Trading Strategy

 Conclusion and Limitations

Findings
● Earnings Momentum Strategy Performance: The
strategy achieved an average annualized return of
26.39% with a Sharpe ratio of 0.71, indicating modest risk-
adjusted returns.
● Reactivity to Earnings Reports: The strategy capitalizes
on the delay in market reactions to earnings
announcements, aiming to profit from subsequent
adjustments in stock prices.
● Market Sensitivity: It was observed that the strategy's
performance is sensitive to market conditions and the
timing of earnings announcements.

21 | Earnings Momentum Trading Strategy

 Conclusion and Limitations

Limitation:
● Market Sensitivity: The strategy’s performance is heavily dependent on timely and accurate
earnings reports. Delays or inaccuracies in earnings disclosure could skew the effectiveness.
● Crowded Trades: As earnings momentum is a well-known strategy, increases participation can
diminish returns, as evidenced by reduced profitability over the evaluation period.
● Data Dependency: The strategy relies on precise data from financial reports. Inaccuracies or
inconsistencies in reported earnings can lead to false signals.
● Model Overfitting: There is a risk of overfitting to historical data, which may not necessarily predict
future performance accurately.

22 | Earnings Momentum Trading Strategy

 Implementation

Future Research Suggestions:

● Alternative Data Sources: Exploring less traditional data sources might provide new insights and
less crowded trading angles
● Machine Learning Techniques: Applying advanced machine learning models to predict earnings
surprises could enhance predictive accuracy and trading outcomes
● Broader Market Conditions: Incorporating macroeconomic indicators to adjust the trading strategy
according to broader market conditions could reduce drawdowns and improve returns

Real-World Trading Scenarios:

● Risk Management Enhancements: Implementing stricter risk management rules, such as stop-loss
orders or dynamic position sizing based on volatility, could help mitigate losses during drawdown
periods
● Integration with Other Strategies: To diversify risks, this strategy could be combined with other
trading strategies that are based on different principles, such as price momentum or macroeconomic
indicators
23 | Earnings Momentum Trading Strategy
References

● Price, Earnings, and Revenue Momentum Strategies (Chen et al. 2009)

● Chan, Louis, Narasimhan Jegadeesh and Josef Lakonishok (1996), Momentum

strategies, Journal of Finance 51, 1681-1713.

● Jason, Fernando. “Earnings Per Share (EPS): What It Means and How to Calculate It.”
Investopedia. Investopedia, August 23, 2022.

● Chen, Hong-Yi, Sheng-Syan Chen, Chin-Wen Hsin, and Cheng-Few Lee. “Price,
Earnings, and Revenue Momentum Strategies.” Asian Academy of Management Journal
of Accounting and Finance , August 2015.

24 | Earnings Momentum Trading Strategy