{"paper_id":"2510.07444","title":"Minimizing the Value-at-Risk of Loan Portfolio via Deep Neural Networks","authors":["Albert Di Wang","Ye Du"],"paper_date":"2025-10-08","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2510.07444v1","page_url":"https://thequant.space/flowcharts/minimizing-the-value-at-risk-of-loan-portfolio-via-deep-neur/","scores":{"empirical_rigor":6,"math_complexity":7,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes two deep neural network models, DeNN (low degree of freedom) and DSNN (high degree of freedom), to predict loan default probabilities and times, and uses Monte‑Carlo simulations to minimize portfolio VaR and CVaR.","quote":"We propose a low degree of freedom deep neural network model, DeNN, as well as a high degree of freedom model, DSNN, to tackle the problem.","summary":"Risk management is a prominent issue in peer-to-peer lending. An investor may naturally reduce his risk exposure by diversifying instead of putting all his money on one loan. In that case, an investor may want to minimize the Value-at-Risk (VaR) or Conditional Value-at-Risk (CVaR) of his loan…","strategy":"DeNN"},"assumptions":{"status":"available","items":[{"assumption":"Loan returns are independent of each other.","type":"model","stated":"explicit","quote":"assuming loans’ returns are independent of each other."},{"assumption":"Portfolio consists of a fixed number of 40 loans.","type":"execution","stated":"explicit","quote":"We set the portfolio size 𝑛=40, and randomly sample 2,000 portfolios with replacement from the test set."},{"assumption":"Monte‑Carlo simulation of returns approximates the true VaR/CVaR distribution.","type":"statistical","stated":"implied","quote":"we use Monte‑Carlo methods to simulate possible realized returns for loans in a portfolio"},{"assumption":"The cleaned Lending Club data accurately represents loan characteristics.","type":"data","stated":"implied","quote":"After thoroughly cleansing and preprocessing, 244,720 loans with 104 features and 2 labels (lifetime and default indicator) are left."}]},"datasets":{"status":"available","items":[{"name":"Lending Club loan dataset","provider":"Lending Club","kind":"fundamentals","access":"public","period":null,"universe":"244,720 loans","quote":"The dataset was downloaded from Lending Club, one of the largest P2P online lending platforms in the world. After thoroughly cleansing and preprocessing, 244,720 loans with 104 features and 2 labels (lifetime and default indicator) are left."}],"known":[],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"weekly-monthly"},"transaction_costs":{"status":"available","modeled":"no","headline_net_of_costs":false,"detail":null,"quote":null,"source":"front_matter","rule":"R1","candidates":{"front_matter":"no","evidence_card":"no","evidence_card_quoted":false},"meaning":"a trading or execution result is reported without costs"},"results":{"status":"available","evaluation_type":"trading-pnl","findings":[{"finding":"Both DeNN and DSNN significantly reduce portfolio VaR at various confidence levels compared to equal‑weight and random‑weight baselines.","kind":"performance","quote":"both DeNN and DSNN are significantly outperforming the Equal weights and Random weights baselines in VaRs minimization."},{"finding":"DeNN outperforms DSNN in most scenarios despite having a lower degree of freedom.","kind":"performance","quote":"the low degree of freedom model, DeNN, outperforms DSNN in most scenarios."},{"finding":"DeNN reduces VaR at 95% confidence by 17.45% when optimizing CVaR' compared to the LRM benchmark.","kind":"performance","quote":"comparing to LRM, the DeNN reduces the VaR at 95% when optimizing CVaREF% by 17.45%, which corresponds to a decrease of"},{"finding":"DSNN reduces VaR at 95% confidence by 6.26% when optimizing VaR' compared to a simple SNN benchmark.","kind":"performance","quote":"comparing to simple SNN, the DSNN reduces the VaR at 95% when optimizing VaREF% by 6.26%, which corresponds to a decrease of annualized loss from 44.40% to 41.62%."}],"reported_performance":{"strategy":"DeNN","sharpe":null,"sharpe_benchmark":null,"annual_return_pct":null,"total_return_pct":null,"max_drawdown_pct":null},"out_of_sample":{"method":"holdout","headline_out_of_sample":true}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":45,"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":1,"extracted_at":"2026-11-14T17:25:11Z"},"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/"}