The data files contain seven low-dimensional financial research data (in .txt format) and four high-dimensional daily stock prices data (in .csv format). The low-dimensional data sets are provided by Lorenzo Garlappi on his website, while the high-dimensional data sets are downloaded from Yahoo!Finance by the contributor's own efforts. The description of the low-dimensional data sets can be found in DeMiguel et al. (2009, RFS).

Finance Data Fusion Market Outlook

The global finance data fusion market size is projected to grow at a CAGR of 12.5% from 2024 to 2032, with the market value increasing from USD 2.5 billion in 2023 to an estimated USD 7.4 billion by 2032. This impressive growth is driven by an intensifying demand for real-time analytics, the increasing complexity of financial transactions, and the need for improved risk management and fraud detection mechanisms in the financial sector.

One of the primary growth factors propelling the finance data fusion market is the rising necessity for robust risk management solutions. Financial institutions are increasingly recognizing the importance of integrating diverse data sources to gain comprehensive insights into potential risks. With the advent of big data and advanced analytics, data fusion technologies enable organizations to synthesize information from multiple datasets, including market data, transactional data, and social media feeds, thereby enhancing their ability to predict and manage risks in a dynamic market environment. This capability is particularly critical in an era where financial stability and regulatory compliance are paramount.

Another significant driver of market growth is the surging demand for enhanced fraud detection systems. Financial fraud has become increasingly sophisticated, necessitating the adoption of advanced technologies that can detect and mitigate fraudulent activities in real-time. Data fusion solutions allow for the integration of diverse data points, providing a holistic view of customer behavior and transaction patterns. This multi-dimensional analysis significantly improves the accuracy of fraud detection systems, enabling financial institutions to safeguard their assets and maintain customer trust. The growing reliance on digital payment systems further underscores the need for advanced fraud detection technologies.

Furthermore, the growing importance of customer analytics in the financial sector is contributing to the market's expansion. Financial institutions are leveraging data fusion technologies to gain deeper insights into customer preferences, behavior, and needs. By integrating data from various sources, such as transaction histories, social media interactions, and demographic information, organizations can create detailed customer profiles that drive personalized marketing strategies and improve customer satisfaction. The ability to deliver tailored financial products and services based on comprehensive data analysis is a key competitive advantage in the financial industry.

Regionally, North America is expected to dominate the finance data fusion market, owing to its advanced financial infrastructure and the early adoption of innovative technologies. The presence of major financial institutions and a highly developed regulatory framework further supports market growth in this region. Europe and Asia Pacific are also anticipated to witness substantial growth, driven by increasing investments in financial technology and the rising demand for advanced data analytics solutions. In contrast, Latin America and the Middle East & Africa are projected to experience moderate growth, influenced by varying levels of technological adoption and economic development.

Component Analysis

The finance data fusion market can be segmented by component into software, hardware, and services. The software segment is expected to hold the largest market share, driven by the increasing adoption of advanced analytic tools and platforms that enable the integration and analysis of diverse data sources. Financial institutions are investing heavily in software solutions that provide real-time insights and predictive analytics, facilitating more informed decision-making and enhancing operational efficiency. The proliferation of cloud-based software solutions is also contributing to the segment's growth, as they offer scalable and cost-effective alternatives to traditional on-premises systems.

The hardware segment, although smaller in comparison to software, plays a crucial role in supporting data fusion activities. High-performance computing systems, storage solutions, and network infrastructure are essential for managing and processing the vast amounts of data generated in the financial sector. As financial institutions continue to expand their data capabilities, the demand for robust and scalable hardware solutions is expected to rise. Innovations in hardware technology, such as advanced processors and high-speed storage devices, are further driving the segment's growth.

In this study, we introduce a novel approach, called the quasi maximum likelihood estimation (Q-MLE), for estimating large-dimensional matrix factor models. In contrast to the principal component analysis based approach, Q-MLE considers the heteroscedasticity of the idiosyncratic error term, the heteroscedasticity of which is simultaneously estimated with other parameters. Interestingly, under the homoscedasticity assumption of the idiosyncratic error, the Q-MLE estimator encompassed the projected estimator (PE) as a special case. We provide the convergence rates and asymptotic distributions of the Q-MLE estimators under mild conditions. Extensive numerical experiments demonstrate that the Q-MLE method performs better, especially when heteroscedasticity exists. Furthermore, two real examples in finance and macroeconomics reveal factor patterns across rows and columns, which coincide with financial, economic, or geographical interpretations.

PhD thesis Hao Li

Credit risk assessment remains a critical function within financial services, influencing lending decisions, portfolio risk management, and regulatory compliance. It integrates multiple categories of financial, transactional, and behavioral data to enable advanced machine learning applications in the domain of financial risk modeling.

Data Composition and Structure

The dataset comprises a total of 1,212 distinct features, systematically grouped into four principal categories, alongside a binary target variable. Each feature category represents a specific dimension of credit risk assessment, reflecting both internal transactional data and externally sourced credit bureau information.

Target Variable

The dependent variable, denoted as bad_flag, represents a binary risk classification outcome associated with each customer account. The variable takes the following values:

• 0: Denotes a low-risk, creditworthy customer
• 1: Denotes a high-risk, default-prone customer

This variable serves as the target for binary classification models aimed at predicting credit risk propensity.

Feature Groups

Each feature within a category follows a systematic sequential naming convention (e.g., transaction_attribute_1, bureau_1), facilitating programmatic identification and group-level analysis.

Data Characteristics

The dataset exhibits several characteristics that mirror operational credit risk data environments:

• High Dimensionality: The feature space exceeds 1,200 variables
• Mixed Data Types: Numerical values (continuous and discrete), binary indicators
• High Sparsity: A substantial proportion of features contain zero values or missing entries
• Value Range Disparity: Feature values exhibit significant variance, with magnitudes ranging from small ratios (0.001) to large transaction amounts (288,500)

Methodological Rationale

The dataset was constructed by simulating data generation processes typical within financial services institutions. Transactional behaviors, bureau records, and inquiry histories were aggregated and engineered into derivative features.

Sliced inverse regression (SIR) is the most widely used sufficient dimension reduction method due to its simplicity, generality and computational efficiency. However, when the distribution of covariates deviates from multivariate normal distribution, the estimation efficiency of SIR gets rather low, and the SIR estimator may be inconsistent and misleading, especially in the high-dimensional setting. In this article, we propose a robust alternative to SIR—called elliptical sliced inverse regression (ESIR), to analysis high-dimensional, elliptically distributed data. There are wide applications of elliptically distributed data, especially in finance and economics where the distribution of the data is often heavy-tailed. To tackle the heavy-tailed elliptically distributed covariates, we novelly use the multivariate Kendall’s tau matrix in a framework of generalized eigenvalue problem in sufficient dimension reduction. Methodologically, we present a practical algorithm for our method. Theoretically, we investigate the asymptotic behavior of the ESIR estimator under the high-dimensional setting. Extensive simulation results show ESIR significantly improves the estimation efficiency in heavy-tailed scenarios, compared with other robust SIR methods. Analysis of the Istanbul stock exchange dataset also demonstrates the effectiveness of our proposed method. Moreover, ESIR can be easily extended to other sufficient dimension reduction methods and applied to nonelliptical heavy-tailed distributions.

This is the replication file for the paper "Global spillovers from multi-dimensional US monetary policy" by G. Georgiadis and M. Jarocinski.

Supplementary Information Files for A three-dimensional asymmetric power HEAVY modelThis article proposes the three‐dimensional HEAVY system of daily, intra‐daily, and range‐based volatility equations. We augment the bivariate model with a third volatility metric, the Garman–Klass estimator, and enrich the trivariate system with power transformations and asymmetries. Most importantly, we derive the theoretical properties of the multivariate asymmetric power model and explore its finite‐sample performance through a simulation experiment on the size and power properties of the diagnostic tests employed. Our empirical application shows that all three power transformed conditional variances are found to be significantly affected by the powers of squared returns, realized measure, and range‐based volatility as well. We demonstrate that the augmentation of the HEAVY framework with the range‐based volatility estimator, leverage and power effects improves remarkably its forecasting accuracy. Finally, our results reveal interesting insights for investments, market risk measurement, and policymaking.

Our research is about robust analysis for high dimensional factor model in present of heavy-tailed data. We propose novel methods by integrating the modified Huber loss function and the common Principal Component Analysis. The methods are superior or comparable to others in numerical studies and the estimated factor number is more aligned with financial practice.

The real data in finance is from Kenneth R. French's website: http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/data_library.html. We use three portfolio pools: Pool A, Pool B, and Pool C to do factor analysis. Each pool contains 100 portfolios with complete monthly average value-weighted returns data from July 2016 to June 2024. The Portfolios in each pool are influenced by two primary factors. The authors have no permission to share the data or make the data public available.

We give the R codes for data generating, parameter setting and computational details in simulations.

NOTE: The EU Budget for Hungary of EUR 21.7 billion presented in this dataset reflects the allocations to Hungary in decisions adopted on 22 December 2022. As of 1 January 2025, the total EU budget effectively allocated to Hungary is EUR 20.7 billion. Once the detailed mid term decisions are adopted during 2025 the dataset will reflect the corrected amounts. Find out more in the FAQ - question 2.10 - https://cohesiondata.ec.europa.eu/stories/s/jxcd-m4vy. 2021-2027 financial details broken down by Fund / MS / programme / Policy objective / specific objective / Categorisation dimension. This dataset provides information on planned total and EU financing for 11 different EU Funds (2021-2027) in current prices. The data is taken from multiple categorisation tables in adopted programmes and is broken down by fund, programme, priority, policy objective objective, specific objective, category of region (more developed, less developed, etc. where available) and by the dimensions in the categorisation system. The values from the differnent dimensions should not be aggregated as this would lead to doublecounting. Find out more about the cohesion policy categorisation system here: https://cohesiondata.ec.europa.eu/stories/s/2021-2027-categorisation-information-system/hhu3-atyz. It is updated daily to reflect any modifications (i.e. thematic reallocations) agreed between the Member States and the Commission. The data covers the more than 480 programmes and includes the EU and national co-financing covered by the adoption decision. Financial allocations in the adopted programmes may change over time (i.e. transfers between themes, between funds). For ERDF, ESF and CF the "priority" columns relate financing envelopes in the programmes; the policy objectives and specific objectives relate to the fixed lists of defined objectives set in the specific fund Regulations. If downloading, check the format of financial amounts and where necessary change your regional settings and default formatting in your chosen software. You can check the results of your aggregation of the financial data by comparing results either the charts on the public platform. Please refer to the user guide - https://cohesiondata.ec.europa.eu/stories/s/Cohesion-Open-Data-User-Guide/cf5w-2b26 - for information on how to download data, etc.

This study focused on how financial difficulties may hinder or facilitate sound financial management. A survey of 1,000 adults aged 18 years and older from the general population was conducted by Knowledge Networks on behalf of the National Center for Family and Marriage Research. The survey was completed by 1,014 respondents out of 1,517 cases (66.8 percent response rate). Although financial behavior research is common in the literature, no financial behavior scale exists that is both multi-dimensional and psychometrically validated. Using data from a national sample, this study developed and examined the psychometric properties of a new scale of financial management behaviors. The Financial Behavior Scale (FBS) displayed adequate reliability (alpha = .81). Further, it was highly associated with other measures of financial behavior and discriminated between financial behaviors and time use behaviors. Finally, the scale was highly predictive of savings, consumer debt, and investments. Thus, the FBS appears to be a reliable and valid scale of financial behaviors.

This analysis presents a rigorous exploration of financial data, incorporating a diverse range of statistical features. By providing a robust foundation, it facilitates advanced research and innovative modeling techniques within the field of finance.

Xometry (XMTR): A New Dimension in Manufacturing?

Financial data:

• Historical daily stock prices (open, high, low, close, volume)

• Fundamental data (e.g., market capitalization, price to earnings P/E ratio, dividend yield, earnings per share EPS, price to earnings growth, debt-to-equity ratio, price-to-book ratio, current ratio, free cash flow, projected earnings growth, return on equity, dividend payout ratio, price to sales ratio, credit rating)

• Technical indicators (e.g., moving averages, RSI, MACD, average directional index, aroon oscillator, stochastic oscillator, on-balance volume, accumulation/distribution A/D line, parabolic SAR indicator, bollinger bands indicators, fibonacci, williams percent range, commodity channel index)

Machine learning features:

• Feature engineering based on financial data and technical indicators

• Sentiment analysis data from social media and news articles

• Macroeconomic data (e.g., GDP, unemployment rate, interest rates, consumer spending, building permits, consumer confidence, inflation, producer price index, money supply, home sales, retail sales, bond yields)

Potential Applications:

• Stock price prediction

• Portfolio optimization

• Algorithmic trading

• Market sentiment analysis

• Risk management

Use Cases:

• Researchers investigating the effectiveness of machine learning in stock market prediction

• Analysts developing quantitative trading Buy/Sell strategies

• Individuals interested in building their own stock market prediction models

• Students learning about machine learning and financial applications

Additional Notes:

• The dataset may include different levels of granularity (e.g., daily, hourly)

• Data cleaning and preprocessing are essential before model training

• Regular updates are recommended to maintain the accuracy and relevance of the data

This analysis presents a rigorous exploration of financial data, incorporating a diverse range of statistical features. By providing a robust foundation, it facilitates advanced research and innovative modeling techniques within the field of finance.

Nano Dimension (NNDM) Stock Forecast: Printing Money with 3D Electronics

Financial data:

• Historical daily stock prices (open, high, low, close, volume)

• Fundamental data (e.g., market capitalization, price to earnings P/E ratio, dividend yield, earnings per share EPS, price to earnings growth, debt-to-equity ratio, price-to-book ratio, current ratio, free cash flow, projected earnings growth, return on equity, dividend payout ratio, price to sales ratio, credit rating)

• Technical indicators (e.g., moving averages, RSI, MACD, average directional index, aroon oscillator, stochastic oscillator, on-balance volume, accumulation/distribution A/D line, parabolic SAR indicator, bollinger bands indicators, fibonacci, williams percent range, commodity channel index)

Machine learning features:

• Feature engineering based on financial data and technical indicators

• Sentiment analysis data from social media and news articles

• Macroeconomic data (e.g., GDP, unemployment rate, interest rates, consumer spending, building permits, consumer confidence, inflation, producer price index, money supply, home sales, retail sales, bond yields)

Potential Applications:

• Stock price prediction

• Portfolio optimization

• Algorithmic trading

• Market sentiment analysis

• Risk management

Use Cases:

• Researchers investigating the effectiveness of machine learning in stock market prediction

• Analysts developing quantitative trading Buy/Sell strategies

• Individuals interested in building their own stock market prediction models

• Students learning about machine learning and financial applications

Additional Notes:

• The dataset may include different levels of granularity (e.g., daily, hourly)

• Data cleaning and preprocessing are essential before model training

• Regular updates are recommended to maintain the accuracy and relevance of the data

Snapshot img

Expense Management Software Market Size 2025-2029

The expense management software market size is forecast to increase by USD 7.5 billion at a CAGR of 16.2% between 2024 and 2029.

The market is witnessing significant growth, driven by the increasing adoption of cloud-based solutions. Companies are increasingly turning to these solutions to streamline their expense management processes and improve operational efficiency. Another key trend is the integration of artificial intelligence (AI) and machine learning (ML) technologies, which enable automated expense categorization and approval workflows. However, the market also faces challenges. Security and privacy concerns continue to be a major obstacle, as companies must ensure the protection of sensitive financial data.
Ensuring compliance with data protection regulations, such as GDPR and HIPAA, is crucial for maintaining customer trust and avoiding potential legal issues. Additionally, the implementation of these advanced technologies requires significant investment in IT infrastructure and expertise, which may be a barrier for smaller organizations. Companies seeking to capitalize on market opportunities and navigate these challenges effectively must prioritize data security, invest in IT capabilities, and offer user-friendly, cost-effective solutions. Multi-dimensional analysis, staff communications, and workflow and data integration are essential features for modern expense management systems.

What will be the Size of the Expense Management Software Market during the forecast period?

Explore in-depth regional segment analysis with market size data - historical 2019-2023 and forecasts 2025-2029 - in the full report.
Request Free Sample

In the travel expense management market, a centralized system with multi-platform software and cloud-based deployments is becoming increasingly popular among buyers and users. This shift towards cloud-based expenditure enables staff communications and spending visibility, enhancing organizational effectiveness and reducing operating costs. File attachments, such as receipts, are seamlessly integrated into these systems, ensuring audit trails and streamlining the expense reporting process. Artificial intelligence (AI) and machine learning technologies are revolutionizing the market, offering multidimensional analysis capabilities that go beyond simple expense categorization. Travel technology companies are partnering with software providers to offer integrated solutions, enhancing workflow efficiency and minimizing manual data entry.

Employee hard drives are no longer the primary storage solution for expense reports, as cloud-based systems provide secure access to information from any mobile device. Automation of expense reporting processes and real-time data analysis enable quicker decision-making and improved financial management. Integration with other business systems, such as accounting software and HR platforms, further enhances the value proposition of travel expense management solutions. The market is expected to continue growing, driven by the need for streamlined expense management and increased operational efficiency.

How is this Expense Management Software Industry segmented?

The expense management software industry research report provides comprehensive data (region-wise segment analysis), with forecasts and estimates in 'USD million' for the period 2025-2029, as well as historical data from 2019-2023 for the following segments.

Component

  Solution
  Service


Application

  Large enterprises
  SMEs


Deployment

  Cloud-based
  On-premises


Geography

  North America

    US
    Canada


  Europe

    France
    Germany
    Italy
    Spain
    UK


  APAC

    China
    India
    Japan


  Rest of World (ROW)

By Component Insights

The Solution segment is estimated to witness significant growth during the forecast period. The market is experiencing significant growth due to its ability to streamline business-related activities for organizations. This market encompasses solutions like expense reporting software, travel and expense management software, invoice management software, and more, collectively referred to as expense management software-as-a-service. These solutions cater to both on-premises and cloud-based deployments. While cloud-based solutions are hosted externally and accessed via the Internet, on-premises solutions are installed locally. Phishing and cybercriminals pose threats to financial data security, making predictive analytics a crucial component in expense management software. Mobile devices are increasingly used for business activities, necessitating mobile terminal support in expense management solutions.

Client responses and employee hard drives are also targeted by malware and social engineering attacks, emphasizing the importance of robust security measures. Large enterprises

This dataset contains data for 'Measuring Financial Inclusion: A State-Wise Impact Analysis of India’s National Mission on Financial Inclusion'. This dataset contains 3 files where the individual dimensional scores are calculated, and a FIS Calculation file which combines the three dimensions into one comprehensive value.

Purpose: This study aims to calculate the level of financial inclusion across India’s states/ UTs and Union Territories from 2011 to 2019, to quantify the impact of India’s National Mission on Financial Inclusion (Pradhan Mantri Jan Dhan Yojana) which was launched in 2014. Approach: We utilise the Index of Financial Inclusion, which has 3 dimensions- Banking Penetration, Banking Services Availability and Banking Services Usage, comprised of sub-dimensions to capture various individual indicators of financial inclusion. This allows us to factor in the various aspects of financial inclusion into one score, which can then be used for empirical studies. Findings: We find that the National Mission on Financial Inclusion (PMJDY) has been able to accelerate the growth of financial inclusion in India, and as of 2019, 24 of India’s 31 states/UTs, considered in this paper, have been able to achieve High Financial Inclusion, with the remaining 7 achieving Medium Financial Inclusion. This represents a significant improvement over the state of financial inclusion in 2011. Originality: We have enhanced the Index of Financial Inclusion used in previous studies to include additional indicators of financial inclusion based on existing literature. The application of the measurement framework to quantify the impact of India’s National Mission of Financial Inclusion is original. The financial inclusion scores have been calculated over an extended period from 2011 to 2019.

Energy plays a crucial role in global economic development, but it also contributes significantly to CO2 emissions. China has proposed a “dual-carbon” goal, and a key aspect to achieving this objective is finding effective ways to promote the decarbonization of the energy consumption structure (DECS). Compared with traditional finance, green finance is pivotal in advancing green and low-carbon development. However, the mechanism through which green finance impacts DECS has not been thoroughly explored. This study employs an enhanced weighted multi-dimensional vector angle method, which is more systematic and scientific, to measure DECS. Then, dynamic panel data from 30 provinces in China spanning the years 2003 to 2020 are used. A double fixed-effects model is applied to investigate the impact of green finance on the DECS and identify potential pathways. Results reveal that green finance significantly enhances DECS, primarily by reinforcing green development. The critical impact pathway involves the promotion of green technology innovation and green industry development. Moreover, the enhancing effect of green finance on the DECS is considerably significant in regions with relatively low government spending on science and technology (S&T), and where the focus is not on the “Atmospheric Ten” policy. The measurement of DECS is innovative, and the conclusions derived from it can offer compelling evidence for various social stakeholders. The government has the opportunity to establish a green financial system, supporting green technological innovation and the development of green industries. This approach can accelerate the DECS and work toward achieving the “double carbon” goal at an earlier date.

Regression results for different dimensional benchmark models.

Context Predicting stock market movements is a classic challenge in machine learning. While raw Open, High, Low, Close, and Volume (OHLCV) data is the standard starting point, its predictive power is often limited. To build robust models, data scientists require a much richer feature set that captures different aspects of market dynamics, from technical patterns to sentiment hidden in financial news.

This dataset was created to bridge that gap. It provides a highly-enriched, pre-processed collection of features for the Dow Jones Industrial Average (DJIA), designed to accelerate research and modeling for stock price prediction.

Content The dataset is organized into several files, each representing a distinct category of engineered features. This modular structure allows you to easily select, combine, or test the importance of different feature types.

• final_daily_news_graph_embeddings.npy This is a 3D NumPy tensor with the shape (Number of Days, 25, 128).

Description: Each day's top 25 news headlines have been transformed into a sophisticated knowledge graph. These graphs, enriched with data from Wikidata, are then encoded into 128-dimensional vectors using a Graph Convolutional Network (GCN). This file captures the semantic meaning and relationships within the news, providing a powerful non-price-based feature.

• DJIA_engineered_features_1.csv

Description: Contains fundamental features derived directly from OHLCV data. These are crucial for capturing intraday volatility and price action.

Example Features: intraday_range, body_size, price_change, simple_return, log_return, price_volume_interaction.

• DJIA_technical_indicators_2.csv

Description: A wide array of popular technical indicators calculated using the pandas-ta library. These features are staples of financial analysis and help identify trends, momentum, and volatility.

Example Features: Simple Moving Averages (SMA_20, SMA_50, SMA_200), Exponential Moving Averages (EMA_12, EMA_26), MACD, RSI, Bollinger Bands (BBL, BBM, BBU), On-Balance Volume (OBV), and more.

• DJIA_statistical_time_features_3.csv

Description: This file includes features based on the statistical properties of returns over an optimized rolling window, as well as cyclical time-based features. The optimal window was determined by finding the period with the highest correlation to future returns.

Example Features: rolling_mean, rolling_std (volatility), rolling_skew, rolling_kurt, day_of_week_sin, day_of_week_cos, is_month_end.

• DJIA_advanced_features_4.csv

Description: More complex and transformational features designed to capture deeper market dynamics.

Example Features: Lagged returns and RSI, quantitative candlestick pattern features, wavelet transform coefficients (to decompose price signals into different frequencies), and the Hurst Exponent (to measure long-term memory in the time series).

Methodology The features were systematically generated using a series of Python scripts.

News Embeddings: Headlines were processed to extract named entities. These entities were used to build knowledge subgraphs from Wikidata. Finally, a Graph Convolutional Network (GCN) model encoded these graphs into dense vectors.

Tabular Features: All other features were generated from the raw DJIA price and volume data. The process involved several stages, from basic price calculations to advanced transformations. For features requiring a lookback period (e.g., rolling statistics, Hurst exponent), an optimal window length was programmatically determined to maximize its correlation with the target variable.

Acknowledgements The raw OHLCV and news data was originally sourced from: https://www.kaggle.com/datasets/aaron7sun/stocknews. We thank them for making the data available.

Inspiration This dataset is perfect for a variety of financial machine learning tasks:

Can you build a model to predict the next day's market direction (Up/Down)?

Which feature set is the most powerful? The technical indicators, the news embeddings, or a combination of all?

How do advanced features like the Hurst exponent or wavelet coefficients contribute to model performance?

Can you use these features to build a profitable trading strategy (backtesting required)?

Climate change-induced pan-financial market and the contagion of systemic financial risks are becoming important issues in the financial sector. The paper measures the temperature difference in terms of the degree and direction of deviation of the actual temperature relative to the average temperature of the same historical period. Based on the high-dimensional time-series variable LASSO-VAR-DY framework, we construct a pan-financial market volatility correlation network consisting of 112 Chinese listed companies in banking, insurance, securities, real estate, traditional energy, and new energy, use eigenvector centrality to measure the systematic risk of each firm, and then empirically test the effect of temperature difference on systematic risk under pan-financial market scenario. The results of the study show that (ⅰ) There is a significant difference among the systemic risk of financial sectors such as banking, insurance, and securities in the financial market pan-financial market scenario and the systemic risk when the financial market pan-financial market is not taken into account;(ⅱ) Higher temperature significantly exacerbates systemic financial risk, while colder temperature significantly mitigates systemic risk, but both have an asymmetric effect on systemic risk, and there is sectoral heterogeneity.(ⅲ) From the dynamic evolutionary characteristics, there are significant differences in the response of systemic financial risk to positive and negative temperature shocks;(iv) The results of the systemic risk variance decomposition indicate that the temperature change contributes more to the variance of systemic risk in the banking and securities sectors in pan-financial market;(ⅴ) The contagion source of financial systemic risk shows an obvious path of leaping and changing characteristics, and the contagion source of systemic risk (source of impact) shows the evolution law of "bank → real estate → new energy → temperature difference," which means that the temperature difference has become the contagion source of systemic financial risk. This study provides a reference for preventing and resolving systemic risks under pan-financial market scenario and provides a basis for improving the current macroprudential regulatory framework.

An increase in a currency internationalization levels can positively impact its credibility in international economic activities, and expand the effective demand and optimize the supply structure for the country’s financial service trade. In this way, a state can improve its financial service trade competitiveness in the international market. This study builds a vector autoregressive model based on time-series data of China-US financial services trade from 2010 to 2021, analyzes the impact of different quantitative indicators of RMB internationalization on this trade from the impulse response results, and validates the conclusions using various inspection methods. The results show that the increase in RMB internationalization helps to narrow the China-US financial services trade balance, but with a significant lag. And this effect is heterogeneous in different dimensions, demonstrated by the fact that the development of overseas RMB securities business is more important for the level of RMB internationalization to narrow the China-US financial services trade balance. Finally, among the specific measures to improve its financial services trade, China should focus on developing the international competitiveness of the traditional RMB deposit and loan financial sector, while the competition in the overseas market for high value-added financial businesses must also not be neglected. Furthermore, China needs to implement more targeted RMB internationalization development policies at different levels in the future to provide high-quality financial services to the rest of the world and aid in the economic recovery of the world in the "post-pandemic" era.