Size matters: some stylized facts of the stock market revisited

Zoltan Eisler,Janos Kertesz

DOI: https://doi.org/10.1140/epjb/e2006-00189-6

2006-05-17

Abstract:We reanalyze high resolution data from the New York Stock Exchange and find a monotonic (but not power law) variation of the mean value per trade, the mean number of trades per minute and the mean trading activity with company capitalization. We show that the second moment of the traded value distribution is finite. Consequently, the Hurst exponents for the corresponding time series can be calculated. These are, however, non-universal: The persistence grows with larger capitalization and this results in a logarithmically increasing Hurst exponent. A similar trend is displayed by intertrade time intervals. Finally, we demonstrate that the distribution of the intertrade times is better described by a multiscaling ansatz than by simple gap scaling.

Physics and Society,Statistical Mechanics,Statistical Finance

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to re - analyze the high - resolution data of the New York Stock Exchange in order to study the impact of a company's market capitalization on its trading activities. Specifically, the author focuses on the following aspects:

1. **The impact of a company's market capitalization on trading activities**: The author has studied the relationship between the value of each average transaction, the average number of transactions per minute, and the average trading activity and the company's market capitalization. They found that these indicators increase monotonically with the increase of the company's market capitalization, but do not follow a power - law distribution.

2. **Characteristics of the distribution of trading values**: The author has re - examined previous research on the distribution of trading values, especially the so - called "inverse cube law". They found that the second moment of the trading value distribution is finite, which means that the Hurst exponent of the relevant time series can be calculated, and these exponents are not universal, but increase with the increase of the company's market capitalization.

3. **The non - universality of the Hurst exponent**: The author has shown that there are significant differences in the correlation of trading activities among companies with different market capitalizations. For companies with larger market capitalizations, the time series of trading activities show stronger persistence, while for companies with smaller market capitalizations, there is almost no persistence. This difference causes the Hurst exponent to increase with the logarithm of the average trading value.

4. **Multi - scale distribution of trading interval times**: The author has studied the distribution of trading interval times and found that it does not conform to the simple gap - scaling model, but is more suitable to be described by the multi - scale hypothesis. This indicates that the distribution tail of trading interval times has a greater weight for smaller companies.

In summary, this paper reveals the impact of a company's market capitalization on stock market trading activities at multiple levels by re - analyzing high - resolution market data, and points out some limitations and misunderstandings in previous research. These findings are helpful for a more in - depth understanding of the complexity and adaptability of financial markets.

### Formula Summary

- The relationship between the value of each average transaction, the average number of transactions per minute, and the average trading activity and the company's market capitalization \(C_i\):
\[
\langle V_i\rangle, \langle N_i\rangle, \langle f_i\rangle \text{ vs } C_i
\]
where \(\langle f_i\rangle\propto C_i^{0.98\pm0.06}\), while \(\langle V_i\rangle\) and \(\langle N_i\rangle\) cannot be fitted with a single power law across the entire market capitalization range.

- The tail index \(\lambda\) of the trading value distribution:
\[
P_{\Delta t}(f)\propto f^{-(\lambda + 1)}
\]

- The definition of the Hurst exponent \(H(i)\):
\[
\sigma_i^2(\Delta t)=\left\langle\left(f_i^{\Delta t}(t)-\langle f_i^{\Delta t}(t)\rangle\right)^2\right\rangle\propto(\Delta t)^{2H(i)}
\]

- The relationship between the Hurst exponent and the average trading value:
\[
H(i)=H(i = 1)+\gamma\log\langle f_i\rangle
\]

- The distribution of trading interval times and its multi - scale characteristics:
\[
P(T,\langle T\rangle)=\frac{1}{\langle T\rangle}F\left(\frac{T}{\langle T\rangle}\right)
\]
\[
F(x)=\delta\left(\frac{x}{X}\right)^{\delta - 1}\exp\left[-\left(\frac{x}{X}\right)^\delta\right]
\]