{"paper_id":"2504.19050","title":"Phase Transitions in Financial Markets Using the Ising Model: A Statistical Mechanics Perspective","authors":["Bruno Giorgio"],"paper_date":"2025-04-26","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2504.19050v1","page_url":"https://thequant.space/flowcharts/phase-transitions-in-financial-markets-using-the-ising-model/","scores":{"empirical_rigor":6,"math_complexity":7.5,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes using the Ising model, simulated via Monte Carlo methods, to replicate the main stylized facts observed in financial assets such as the S&P 500.","quote":"This dissertation investigates the ability of the Ising model to replicate statistical characteristics, or stylized facts, commonly observed in financial assets. The results demonstrate that the Ising model is capable of replicating the majority of the statistical features analyzed.","summary":"This dissertation investigates the ability of the Ising model to replicate statistical characteristics, or stylized facts, commonly observed in financial assets. The study specifically examines in the S&P500 index the following features: volatility clustering, negative skewness, heavy tails, the…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"Each agent is represented by a spin variable Si(t) that can take values ±1.","type":"model","stated":"explicit","quote":"Let us consider Si(t) = ±1 as the spin orientation for each i-th agent in the lattice.","support":"full"},{"assumption":"For simplicity, the dynamics of each agent are updated according to the probabilistic rule Si(t+1)=+1 with p = 1/(1+e^{−2βhi(t)}) and Si(t+1)=−1 with 1−p.","type":"model","stated":"explicit","quote":"For simplicity, we assume that the dynamics of each agent are updated according to the following equations: Si(t + 1) = +1 with p = 1 1 + e−2βhi(t) Si(t + 1) = −1 with 1 −p In the equations, β is a coupling coefficient, hi(t) rep-resents the local field acting on agent i at time t, and p is the probability.","support":"full"},{"assumption":"β is a coupling coefficient that determines the influence of the local field on agents.","type":"model","stated":"explicit","quote":"For simplicity, we assume that the dynamics of each agent are updated according to the following equations: Si(t + 1) = +1 with p = 1 1 + e−2βhi(t) Si(t + 1) = −1 with 1 −p In the equations, β is a coupling coefficient, hi(t) rep-resents the local field acting on agent i at time t, and p is the probability.","support":"full"},{"assumption":"When β = 0 (infinite temperature) all spin configurations occur with equal probability and spins behave independently.","type":"model","stated":"explicit","quote":"In con-trast, when β = 0 (i.e., infinite temperature), all configurations occur with equal probability, and the spins behave independently.","support":"full"},{"assumption":"The lack of an analytical or deterministic solution for large systems necessitates the use of Monte Carlo simulation.","type":"model","stated":"explicit","quote":"Due to the lack of an analytical or deterministic solution, the Monte Carlo method was employed to explore the model's statistical properties.","support":"full"},{"assumption":"Daily adjusted‑close prices of the S&P 500 index are an appropriate proxy for market returns.","type":"data","stated":"explicit","quote":"We collected daily financial data spanning the past forty years, specifically from February 1982 to Febru-ary 2022. The dataset consists of the Adjusted Close prices of the S&P 500 index.","support":"full"}]},"datasets":{"status":"available","items":[{"name":"S&P 500 Adjusted Close prices","provider":null,"kind":"market-prices","access":"public","period":"Feb 1982 - Feb 2022","universe":"S&P 500 index","quote":"We collected daily financial data spanning the past forty years, specifically from February 1982 to Febru-ary 2022. The dataset consists of the Adjusted Close prices of the S&P 500 index.","support":"full"}],"known":[],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"daily"},"transaction_costs":{"status":"not_applicable","modeled":"n/a","headline_net_of_costs":null,"detail":null,"quote":null,"source":"front_matter","rule":"R3","candidates":{"front_matter":"n/a","evidence_card":"n/a","evidence_card_quoted":false},"meaning":"the paper reports no trading or execution result"},"results":{"status":"available","evaluation_type":"none","findings":[{"finding":"The Ising model is capable of replicating the majority of the statistical features analyzed, including volatility clustering, heavy tails, and negative skewness.","kind":"statistical","quote":"The results demonstrate that the Ising model is capable of replicating the majority of the statistical features analyzed.","support":"full"},{"finding":"The simulated returns exhibit volatility clustering, with clusters of large fluctuations followed by other large changes, mirroring the empirical S&P 500 behavior.","kind":"statistical","quote":"Analysis of Results Volatility Clustering Volatility clustering was the first criterion used to evaluate whether the model exhibited behavior con-sistent with real financial markets. The model exhibits a memory effect, characterized by clusters of large fluctuations followed by other large changes.","support":"partial"},{"finding":"Autocorrelation of absolute returns in the Ising simulation shows long‑memory behavior that decays according to a power‑law, similar to the real market data.","kind":"statistical","quote":"The autocorrelation of absolute returns demon-strated long-memory behavior, decaying according to a power-law—a significant departure from the inde-pendence assumed in random walk models.","support":"full"},{"finding":"Both the Ising model returns and the S&P 500 returns display significant skewness and leptokurtic (heavy‑tailed) distributions, rejecting normality in Shapiro‑Wilk and Jarque‑Bera tests.","kind":"statistical","quote":"Statistical analysis further confirmed the presence of leptokurtic distributions with heavy tails and nega-tive skewness in both the Ising simulations and actual market data. These findings were supported by infer-ential tests (Jarque-Bera and Shapiro-Wilk), which consistently rejected the null hypothesis of normality, aligning with the heavy-tailed, asymmetric nature of financial return distributions.","support":"partial"},{"finding":"The Monte Carlo simulation was performed on a 32×32 lattice for 1,000,000 iterations, generating 10,000 return observations with lag Δt = 100.","kind":"theoretical","quote":"In our initial simulation, we employ a two-dimensional lattice composed of 32×32 agents. Specifically, the system runs through a loop of 1, 000, 000 iterations.","support":"partial"}],"reported_performance":null,"out_of_sample":{"method":"none","headline_out_of_sample":null}},"independent_evidence":{"status":"available","supports":[{"paper":{"paper_id":"2504.18960","title":"Impact of the COVID-19 pandemic on the financial market efficiency of price returns, absolute returns, and volatility increment: Evidence from stock and cryptocurrency markets","paper_date":"2025-04-26","authors":["Tetsuya Takaishi"],"empirical_rigor":5,"page_url":"https://thequant.space/flowcharts/impact-of-the-covid-19-pandemic-on-the-financial-market-effi/"},"question":"Do absolute returns in financial markets display long‑memory (Hurst exponent >0.5) and power‑law autocorrelation decay?","this_paper_finding":"Autocorrelation of absolute returns in the Ising simulation shows long‑memory behavior that decays according to a power‑law, similar to the real market data.","other_paper_finding":"The Hurst exponents of stock returns remain around 0.5 (except VIX), while absolute returns have exponents >0.5, reflecting long‑memory.","other_paper_quote":"The Hurst exponents of the returns time series remain approximately 0.5, except for VIX, for which h(2) is less than 0.5, exhibiting anti-persistence. However, the Hurst exponents of the absolute returns time series are greater than 0.5, reﬂecting the long-memory property of the absolute returns time series (Ding et al., 1993).","note":"A reports Hurst exponents >0.5 for absolute returns; B observes power‑law decay of absolute‑return autocorrelation, confirming long‑memory.","confidence":0.783},{"paper":{"paper_id":"2310.12333","title":"Black-Litterman Asset Allocation under Hidden Truncation Distribution","paper_date":"2023-10-18","authors":["Jungjun Park","Andrew L. Nguyen"],"empirical_rigor":3,"page_url":"https://thequant.space/flowcharts/black-litterman-asset-allocation-under-hidden-truncation-dis/"},"question":"Do asset return distributions deviate from normality, showing skewness and heavy tails?","this_paper_finding":"Both the Ising model returns and the S&P 500 returns display significant skewness and leptokurtic (heavy‑tailed) distributions, rejecting normality in Shapiro‑Wilk and Jarque‑Bera tests.","other_paper_finding":"Likelihood‑ratio test strongly rejects normality, supporting the skew‑normal assumption for the data (p‑value = 0.0015).","other_paper_quote":"Observed–data Log-likelihood normal 3652.19 Log-likelihood skew-normal 3668.80 2 × ln(LR) 33.212 P-value 0.0015","note":"Finding A rejects normality (p=0.0015) favoring skew‑normal; Finding B also rejects normality via Shapiro‑Wilk and Jarque‑Bera, showing skewness and leptokurtosis.","confidence":0.511},{"paper":{"paper_id":"2402.11930","title":"Stylized Facts of High-Frequency Bitcoin Time Series","paper_date":"2024-02-19","authors":["Yaoyue Tang","Karina Arias-Calluari","M. N. Najafi","Michael S. Harré","Fernando Alonso-Marroquin"],"empirical_rigor":7,"page_url":"https://thequant.space/flowcharts/stylized-facts-of-high-frequency-bitcoin-time-series/"},"question":"Do absolute‑return autocorrelations exhibit power‑law decay?","this_paper_finding":"Autocorrelation of absolute returns in the Ising simulation shows long‑memory behavior that decays according to a power‑law, similar to the real market data.","other_paper_finding":"The autocorrelation of absolute returns follows a power‑law initially, with Period 2 showing a slightly higher decay rate than Period 1.","other_paper_quote":"When we sample the autocorrelation of absolute returns, we observe a power-law relationship, indicating time dependence in both periods initially. We fitted the autocorrelation with a power law to capture the decay and found that the second period experienced a slightly higher decay rate.","note":"Both A and B find power‑law decay; B’s simulation confirms A’s real‑data pattern. In A, Period 2 decays slightly faster than Period 1.","confidence":0.472},{"paper":{"paper_id":"2507.01983","title":"Comparing Bitcoin and Ethereum tail behavior via Q-Q analysis of cryptocurrency returns","paper_date":"2025-06-26","authors":["A. H. Nzokem"],"empirical_rigor":6.5,"page_url":"https://thequant.space/flowcharts/comparing-bitcoin-and-ethereum-tail-behavior-via-q-q-analysi/"},"question":"Do the returns of the studied assets exhibit heavy‑tailed, non‑normal distributions?","this_paper_finding":"Both the Ising model returns and the S&P 500 returns display significant skewness and leptokurtic (heavy‑tailed) distributions, rejecting normality in Shapiro‑Wilk and Jarque‑Bera tests.","other_paper_finding":"Both Bitcoin and Ethereum return distributions are heavy‑tailed, deviating significantly from normality.","other_paper_quote":"Therefore, Bitcoin and Ethereum are heavy-tailed distributions. Through rigorous statistical modeling and quan-tile–quantile (Q–Q) plot analysis, we demonstrate that both assets exhibit significant deviations from normality, charac-terized by heavy-tailed distributions.","note":"Both findings report heavy‑tailed returns and rejection of normality (A: significant skewness, leptokurtosis, Shapiro‑Wilk and Jarque‑Bera; B: heavy‑tailed Bitcoin and Ethereum).","confidence":0.442}],"contradicts":[],"qualifies":[],"compared_with":26,"citations":{"cited_by_in_corpus":0,"by_independent_authors":0,"how_cited":{"used_as_baseline":0,"uses_method_or_data":0,"extends":0,"discusses_findings":0,"mentions":0,"unclassified":0},"independent_papers":[],"scope":"Citations between papers in this corpus only; citations from outside it are not counted."}},"implementation":{"status":"available","public":false,"code_url":null,"note":"No repository presented by the authors as their own was found in the paper. Absence here is not proof that no code exists."},"replication":{"status":"none_found","card":null,"note":"No replication of this paper has been run by thequant.space yet."},"extraction":{"text_source":"fulltext","model":"openai/gpt-oss-120b","version":2,"extracted_at":"2026-11-14T19:50:33Z"},"provenance":{"claims":"Everything here is what the authors state in the paper. Nothing was re-run or reproduced by the platform.","quotes":"Assumptions, datasets, cost detail and findings were extracted by a language model. Every quote is verbatim paper text. Where an item has `support`, \"full\" means every figure in it appears in the quoted sentences and \"partial\" means a figure is stated elsewhere in the paper; items with a figure found nowhere in the paper were discarded.","datasets_known":"datasets.known lists well-known vendors and public datasets by a stable id, found by a dictionary over the full text where the sentence says the data is used; it is independent of the model and covers only names in the dictionary.","flags":"transaction_costs.modeled, results.evaluation_type and results.out_of_sample.method are platform classifications (structured-decision models); see https://thequant.space/score-guide/#penalty-flags","comparisons":"supports / contradicts / qualifies are a language model's reading of two papers' quoted findings, restricted to closely related papers that share no author. They are pointers to read, not verified replications, and coverage is partial.","implementation":"code_url is a repository the paper itself presents as its own (rule-scored, ambiguous cases judged by a model), checked for liveness when it was recorded."},"docs":"https://thequant.space/agents/"}