Paper: arXiv 2310.00753
Authors: Vaibhav Sherkar, Rituparna Sen
Abstract
A property of data which is common across a wide range of instruments, markets and time periods is known as stylized empirical fact in the financial statistics literature. This paper first presents a wide range of stylized facts studied in literature which include some univariate distributional properties, multivariate properties and time series related properties of the financial time series data. In the next part of the paper, price data from several stocks listed on 10 stock exchanges spread across different continents has been analysed and data analysis has been presented.
Complexity vs Empirical Score
- Math Complexity: 4.0/10
- Empirical Rigor: 6.5/10
- Quadrant: Street Traders — empirical and applied, lower math complexity
Why this score: The paper employs standard statistical concepts like skewness, correlation, and power-law tail estimation without advanced theoretical derivations, but conducts extensive, multi-market empirical analysis with real data and specific statistical tests.
Research Flowchart
flowchart TD
A["Research Goal: Investigate Stylized Empirical Facts"] --> B{"Methodology"}
B --> C["Collect Stock Price Data<br>10 Global Exchanges"]
B --> D["Define Multivariate & Time Series Properties"]
C --> E["Computational Analysis<br>Statistical Tests & Modeling"]
D --> E
E --> F["Key Findings & Outcomes<br>1. Confirmed Univariate Stylized Facts<br>2. Identified Multivariate Patterns<br>3. Validated Time Series Properties"]
Evidence Card
What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card
Proposes
The paper presents a wide range of stylized empirical facts and analyses price data from several stocks listed on 10 stock exchanges across different continents.
Source sentence
This paper first presents a wide range of stylized facts studied in literature which include some univariate distributional properties, multivariate properties and time series related properties of the financial time series data. In the next part of the paper, price data from several stocks listed on 10 stock exchanges spread across different continents has been analysed and data analysis has been presented.
Assumes 2
- Log returns are used instead of prices to achieve stationarity. statistical
Source sentence
To overcome this issue, we have considered log returns for statistical analysis instead of stock prices. Log returns are defined as follows. rt = log pt pt−1 (1) where rt denotes the log returns at time t and pt denotes the prices at time t.
- Mean returns for many stocks are zero, allowing squared returns to be used as a volatility measure. statistical
Source sentence
Since, mean returns for many of the stocks are zero, we consider squared returns as a measure of volatility.
Data
- Daily closing price and volume of constituents of leading indices of 10 markets market-prices public
Source sentence
Section 5 concludes the paper presenting the conclusions. ∗1Indian Statistical Institute, Kolkata ,Email: [email protected] †2 Indian Statistical Institute, Bengaluru ,Email: [email protected] 1 arXiv:2310.00753v1 [q-fin.ST] 1 Oct 2023 2 Description of Data Daily closing price and Volume data of number of constituents of leading index of the 10 chosen markets for 8 to 10 years in 2010-2019 has been
Costs
Not applicable
Finds 5
- Aggregational Gaussianity is observed: the distribution of returns becomes more normal as the time scale increases from daily to quarterly. statistical
Source sentence
According to the stylized empirical fact, the distribution should look more and more like normal distribution as we move from daily to quarterly returns. By visual inspection of figure 2,which represents the QQ plots of daily, weekly, monthly and quarterly returns of ’China Tourism Group Duty Free Corporation Limited’ stock which is listed on Shanghai Stock Exchange, it can be said that QQ plots looks more and more
- Heavy tails are present in most markets; tail index less than 2 is observed only for Turkey, while tail indices greater than 5 are observed in all markets except Chile, South Africa, and Turkey. statistical
Source sentence
Tail index less than 2 has been observed only in case of Turkey whereas Tail indices greater than 5 have been observed in all the markets except Chile,South Africa and Turkey.However, more than 75% of the observed tail indices lie in the interval 2 to 5 in all the markets as expected.
- Volume‑volatility correlation is generally positive as expected, with only a few markets (Brazil, Chile, Indonesia) showing a small negative correlation (absolute value < 0.02). statistical
Source sentence
4.0.6 Volume-Volatility Correlation One negative value of correlation coefficient between volume and squared returns whose absolute value is much small (less than 0.02) has been obserevd in markets of Brazil,Chile and Indonesia.All the other values of observed correlation coefficients are positive as expected.
- Volatility clustering is validated across the data, indicated by high proportions of stocks rejecting null hypotheses in portmanteau tests. statistical
Source sentence
Overall, the data validates the presence of volatility clustering in financial time series data.
- Absence of autocorrelations is contradicted in most markets, being confirmed only in Brazil, Canada, and Turkey. statistical
Source sentence
Hence, it can be said that this stylized fact has been contradicted except in case of Brazil, Canada and Turkey.
Checked by others
Supported by
- Extreme Value Analysis for Finite, Multivariate and Correlated Systems with Finance as an Example
Compare the findings
This paper: Heavy tails are present in most markets; tail index less than 2 is observed only for Turkey, while tail indices greater than 5 are observed in all markets except Chile, South Africa, and Turkey.
That paper: The estimated tail indices for all modes are mostly of the Fréchet type, confirming heavy‑tailed behavior of high‑frequency financial returns.
- Dynamical properties of volume at the spread in the Bitcoin/USD market
Compare the findings
This paper: Aggregational Gaussianity is observed: the distribution of returns becomes more normal as the time scale increases from daily to quarterly.
That paper: Aggregational Gaussianity: as the time scale of volume returns increases, their distribution approaches a Gaussian, with significant normality at 8‑hour intervals.
Qualified by
- International Financial Markets Through 150 Years: Evaluating Stylized Facts
Compare the findings
This paper: Absence of autocorrelations is contradicted in most markets, being confirmed only in Brazil, Canada, and Turkey.
That paper: Mixed evidence is observed for the absence of log‑return autocorrelation, differing between traditional markets and cryptocurrencies.
2 papers in the archive : MarS: a Financial Market Simulation Engine Powered by …
Code
No repository by the authors found in the paper.
Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.