Paper: arXiv 2507.01989
Authors: Yazdan Babazadeh Maghsoodlo, Amin Safaeesirat
Abstract
Understanding the stochastic behavior of currency exchange rates is critical for assessing financial stability and anticipating market transitions. In this study, we investigate the empirical dynamics of the USD exchange rate in three economies, including Iran, Turkey, and Sri Lanka, through the lens of the Kramers-Moyal expansion and Fokker-Planck formalism. Using log-return data, we confirm the Markovian nature of the exchange rate fluctuations, enabling us to model the system with a second-order Fokker-Planck equation. The inferred Langevin coefficients reveal a stabilizing linear drift and a nonlinear, return-dependent diffusion term, reflecting both regulatory effects and underlying volatility. A rolling-window estimation of these coefficients, paired with structural breakpoint detection, uncovers regime shifts that align with major political and economic events, offering insight into the hidden dynamics of currency instability. This framework provides a robust foundation for detecting latent transitions and modeling risk in complex financial systems.
Complexity vs Empirical Score
- Math Complexity: 8.0/10
- Empirical Rigor: 7.0/10
- Quadrant: Holy Grail — high math complexity, high empirical rigor
Why this score: The paper employs advanced stochastic calculus including Kramers-Moyal expansion and Fokker-Planck equations, demonstrating high mathematical density. It uses real-world currency data with rolling-window estimation and structural break detection, indicating substantial empirical implementation and backtest-ready methodology.
Research Flowchart
flowchart TD
A["Research Goal:<br>Model Exchange Rate Instability"] --> B["Data Acquisition<br>USD Exchange Rate Data: Iran, Turkey, Sri Lanka"]
B --> C["Methodology:<br>Kramers-Moyal Expansion & Fokker-Planck Formalism"]
C --> D{"Markovian Analysis"}
D -- Verified --> E["Computational Process:<br>Infer Langevin Coefficients via Rolling-Window Estimation"]
E --> F["Outcome 1:<br>Stabilizing Linear Drift & Nonlinear Diffusion"]
E --> G["Outcome 2:<br>Regime Shifts via Structural Breakpoint Detection"]
F --> H["Final Insight:<br>Framework for Latent Transition Detection & Risk Modeling"]
G --> H
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 proposes a data‑driven stochastic framework for USD exchange rates using the Kramers‑Moyal expansion and a second‑order Fokker‑Planck equation.
Source sentence
In this study, we investigate the empirical dynamics of the USD exchange rate in three economies, including Iran, Turkey, and Sri Lanka, through the lens of the Kramers–Moyal expansion and Fokker–Planck formalism.
Assumes 5
- The exchange‑rate log‑return series are realizations of a continuous Markov process. model
Source sentence
We use a methodology to extract stochastic characteristics from empirical time series under the assumption that the time series are realizations of a continuous Markov process.
- The series are locally stationary and ergodic. statistical
Source sentence
Assuming local stationarity and ergodicity [16], we estimate the Kramers–Moyal coefficients from a single time series
- Higher‑order Kramers‑Moyal coefficients (D4) are zero, justifying truncation at second order. model
Source sentence
In practice, this expansion is cut off at the second order—a reduction justified by Pawula’s theorem when D4(x,t) = 0 leading to the Fokker–Planck equation.
- Log‑returns provide a locally stationary series under mild assumptions. data
Source sentence
To consider its logarithm, logS(t), to facilitate mathematical treatment and stabilize variance. However, logS(t) typically exhibits nonstationary behavior
Show 1 more
- The Markov length is short (≈0.8 days), allowing daily resolution modeling. statistical
Source sentence
We estimated that the Markov time scales for the US dollar exchange rate for Iran, Turkey, and Sri Lanka is approximately 0.78, 0.80, and 0.78 days, respectively
Data
- Iran USD exchange rate (unofficial market) market-prices unknown
Source sentence
Iran, over the period from March 2002 to February 2025 (unofficial US currency exchange market)
- Turkey USD exchange rate market-prices unknown
Source sentence
Turkey, from January 2010 to May 2025
- Sri Lanka USD exchange rate market-prices unknown
Source sentence
Sri Lanka, from January 2005 to May 2025
Costs
Not applicable
Finds 5
- The estimated drift coefficients are linear and stabilizing across all three countries. statistical
Source sentence
The inferred Langevin coefficients consistently point to a stabilizing drift
- The diffusion coefficient shows a clear nonlinear, return‑dependent (quadratic) shape. statistical
Source sentence
The diffusion coefficient D2(r) demonstrates a clear nonlinear dependence on r, captured by a second‑order polynomial fit
- Higher‑order coefficient D4 is negligible, validating the second‑order truncation. theoretical
Source sentence
In all three data sets, D4(r) remains negligible compared to D2(r), satisfying the conditions of the Pawula’s theorem
- Structural breakpoints in drift and diffusion align with major political and economic events. statistical
Source sentence
Using the ruptures Python library, we detect structural breakpoints in these coefficients and observe that abrupt changes often coincide with major political or economic events
- The log‑return series exhibit Markov lengths of roughly 0.78–0.80 days, confirming daily Markovian behavior. statistical
Source sentence
We estimated that the Markov time scales for the US dollar exchange rate for Iran, Turkey, and Sri Lanka is approximately 0.78, 0.80, and 0.78 days, respectively
Checked by others
No other paper in the archive has been found to support or contradict these findings (compared with 27 related papers). Coverage is partial, so this is neither replication nor refutation.
Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.
Cited References (42)
- Kramers–Moyal Expansion and Fokker–Planck Equation (2019) — link
- The ergodic theorem (1939)
- Approaching complexity by stochastic methods: From biological systems to turbulence (2011) — link
- Official exchange rate (LCU per US, period average) - Iran, Islamic Rep — link
- Approximation of the linear Boltzmann equation by the Fokker-Planck equation (1967)
- Influence of macroeconomic variables on exchange rates (2015)
- Forecasting the Romanian financial system stability using a stochastic simulation model (2010)
- Evidence of Markov properties of high frequency exchange rate data (2001)
and 34 more
- Dynamical stochastic processes of returns in financial markets (2007)
- Modelling the Wealth and Probability of Default of Unsecured Loans: The Stochastic Partial Differential Equation Approach (2025)
- PRICING ASIAN OPTIONS IN UNCERTAIN STOCHASTIC MARKETS WITH JUMPS (2024)
- Topics in rough and stochastic analysis with applications in finance (2025)
- New numerical method via RBF approach to the price of fixed-rate mortgages (2025)
- The Hull-White Model for Interest Rates: A Comprehensive Overview
- Stable Multilevel Deep Neural Networks for Option Pricing and xVAs Using Forward-Backward Stochastic Differential Equations (2024)
- A Langevin equation for the rates of currency exchange based on the Markov analysis (2007)
- Geopolitical risks and exchange rates (2025)
- Fokker–Planck equation of distributions of financial returns and power laws (2001)
- Medium and small-scale analysis of financial data (2007)
- Modelling high-frequency economic time series (2000)
- Stochastic modeling of fat-tailed probabilities of foreign exchange rates (2002)
- A theory ofnon-Gaussian option pricing (2002)
- Financial market dynamics (2003)
- Nonstationary increments, scaling distributions, and variable diffusion processes in financial markets (2007)
- Early warning signals for critical transitions in complex systems (2023)
- Critical slowing down as an early warning signal for financial crises? (2019)
- Early warning signals for critical transitions: a generalized modeling approach (2012)
- Slowing down as an early warning signal for abrupt climate change (2008)
- Generic indicators for loss of resilience before a tipping point leading to population collapse (2012)
- Deep learning for early warning signals of tipping points (2021)
- Early warning signals for stock market crashes: empirical and analytical insights utilizing nonlinear methods (2024)
- A Kramers-Moyal Approach to the Analysis of Third-order noise with applications in option valuation (2015)
- Disentangling the stochastic behavior of complex time series (2016)
- The impact of macroeconomic news on exchange rate volatility (2004)
- Markov analysis and Kramers-Moyal expansion of nonstationary stochastic processes with application to the fluctuations in the oil price (2007)
- Applications of Stochastic Markov Models to Financial Markets (2024)
- The Volterra Stein-Stein model with stochastic interest rates — → 2503.01716
- Regime shifts in the behaviour of international currency and equity markets: A Markov-switching analysis (2021)
- Control and system theory of discrete-time stochastic systems (2021)
- Fokker-planck equation (1996)
- Smooth approximation of stochastic differential equations (2016)
- Analysis of sliding window techniques for exponentiation (1995)
1 of 42 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.