The NLA Corporate Country Risk Index

Methodology

This Curated Measure is based on the approach in Sources and Transmission of Country Risk published in the Review of Economic Studies in 2023 by Hassan, Schreger, Schwedeler, Tahoun.

1. Country dictionaries with weights
   For each country, we build a large country-specific keyword dictionary. Each keyword receives a weight reflecting how strongly it identifies the country. This is based on a tf‑idf based methodology using a comprehensive training library.

2. Sentence and transcript-level scoring
   Each transcript is split into sentences. A sentence contributes to the score when it contains (a) a country keyword (from the country dictionary), and (b) risk synonym. That sentence is assigned the weight of the matched country keyword(s). Each transcript score is the sum of all weighted country-risk sentences in the call.

>Please note: In contrast to Hassan et al. (2023), we use sentence level scoring (they use a window of +/- 10 words).

Data

• 400k+ earnings-call transcripts, ~14,000 public firms based 80+ countries (~60% based in the United States), 2002–today
• Language: English
• Source: LSEG
• Coverage varies by market depth and transcript availability