dorsal/arxiv
View SchemaLimitations of scaling and universality in stock market data
| Authors | Janos Kertesz, Zoltan Eisler |
|---|---|
| Categories | |
| ArXiv ID | physics/0512193 |
| URL | https://arxiv.org/abs/physics/0512193 |
Abstract
We present evidence, that if a large enough set of high resolution stock market data is analyzed, certain analogies with physics -- such as scaling and universality -- fail to capture the full complexity of such data. Despite earlier expectations, the mean value per trade, the mean number of trades per minute and the mean trading activity do not show scaling with company capitalization, there is only a non-trivial monotonous dependence. The strength of correlations present in the time series of traded value is found to be non-universal: The Hurst exponent increases logarithmically with capitalization. A similar trend is displayed by intertrade time intervals. This is a clear indication that stylized facts need not be fully universal, but can instead have a well-defined dependence on company size.
{
"annotation_id": "e12edd3a-5c0c-4975-9303-7307504ed1c3",
"date_created": "2026-03-02T18:01:04.279000Z",
"date_modified": "2026-03-02T18:01:04.279000Z",
"file_hash": "492e8408b2329ac45de3a4e236af1d95db525c62a53ed371aa3fa74dc4ef3fa8",
"private": false,
"record": {
"abstract": "We present evidence, that if a large enough set of high resolution stock\nmarket data is analyzed, certain analogies with physics -- such as scaling and\nuniversality -- fail to capture the full complexity of such data. Despite\nearlier expectations, the mean value per trade, the mean number of trades per\nminute and the mean trading activity do not show scaling with company\ncapitalization, there is only a non-trivial monotonous dependence. The strength\nof correlations present in the time series of traded value is found to be\nnon-universal: The Hurst exponent increases logarithmically with\ncapitalization. A similar trend is displayed by intertrade time intervals. This\nis a clear indication that stylized facts need not be fully universal, but can\ninstead have a well-defined dependence on company size.",
"arxiv_id": "physics/0512193",
"authors": [
"Janos Kertesz",
"Zoltan Eisler"
],
"categories": [
"physics.soc-ph",
"physics.data-an",
"q-fin.ST"
],
"title": "Limitations of scaling and universality in stock market data",
"url": "https://arxiv.org/abs/physics/0512193"
},
"schema_id": "dorsal/arxiv",
"source": {
"execution_id": "5f48af8e-941a-409e-87ac-ea369a3e6871",
"id": "arXiv Dataset IDs",
"type": "Model",
"variant": "snapshot-2026-03-01",
"version": "0.1.0"
},
"user_id": 1000002
}