Global Merger-Arbitrage Forecasting with Language Models

基于语言模型的全球并购套利预测

Abstract: We present a language-model forecasting system for merger arbitrage, a specialized high-stakes financial setting in which the task is to predict the outcome of announced M&A deals.

摘要： 我们提出了一种用于并购套利的语言模型预测系统。并购套利是一个专业且高风险的金融领域，其任务是预测已宣布的并购交易结果。

Unlike prior work on judgmental forecasting with LLMs, which has focused on broad mixed-topic benchmarks and short context such as news snippets, we study a setting that requires long-context reasoning over hundreds of pages of technical documents.

与以往利用大语言模型（LLM）进行判断性预测的研究不同——那些研究多集中于广泛的混合主题基准测试和新闻片段等短上下文——我们研究的场景需要对数百页的技术文档进行长上下文推理。

Our system combines expert-guided context engineering with finetuning on hindsight-guided reasoning traces derived from historical deals.

我们的系统将专家引导的上下文工程与基于历史交易推导出的“事后引导推理轨迹”（hindsight-guided reasoning traces）微调相结合。

Given an announced deal, it outputs a probability distribution over three mutually exclusive outcomes: closing at announced terms, a higher bid, or deal termination.

针对一项已宣布的交易，该系统会输出三种互斥结果的概率分布：按宣布条款成交、出现更高报价，或交易终止。

On an out-of-sample set of more than 400 large deals spanning 42 countries, our finetuned system achieves the best performance of any method we evaluate, reducing class-balanced Brier score to 0.151.

在涵盖 42 个国家、超过 400 笔大型交易的样本外测试集中，我们微调后的系统表现优于所有评估方法，将类别平衡的布里尔分数（Brier score）降低至 0.151。

This is 24% below calibrated market-implied probabilities, 19% below XGBoost, and 25-42% below frontier language models.

这一结果比校准后的市场隐含概率低 24%，比 XGBoost 低 19%，比前沿大语言模型低 25% 至 42%。

These results, together with ablation studies, show that LLM-based forecasting can succeed in specialized, long-context financial workflows, with hindsight-based supervision and expert-designed context playing a critical role.

这些结果连同消融研究表明，基于大语言模型的预测可以在专业的长上下文金融工作流中取得成功，其中基于事后的监督和专家设计的上下文起到了至关重要的作用。