10-K Risk Factor Report methodology
The 10-K Risk Factor Report utilizes data from publicly available sources, including the risk section of 10-K reports, which are annual documents that publicly traded companies must file with the Securities and Exchange Commission (SEC). The 10-K report analyzes 10-K data to surface helpful visuals, which can be used to communicate your company’s risk factors to stakeholders.
The 10-K report uses a fictional company, called an Average Industry Peer, created from aggregated data of a peer group, to compare with individual companies. The scope of individual companies includes the Russel 3000: the largest 3,000 public companies in the US.
Terminology
Here are common terms used in the 10-K report methodology and their definition.
-
Risk factor – a short summary of the risk disclosed in a 10-K by a company. There are three levels of risk in a 10-K:
-
Risk category – a general topic that contains several, more detailed risk factors.
-
Risk factor – a specific description of a risk.
-
Risk detail – the most detailed description of a risk. It has a one-to-one connection to each company risk factor.
-
-
Average Industry Peer (peer groups) – a peer group is identified by their Global Industry Classification Standard (GICS) Sector and Industry codes. GICS is an industry analysis framework that helps investors understand the key business activities for companies around the world. Some examples include Energy, Health Care, Communication Services, etc.
Topic classification methodology
To understand and compare risk factors, we categorize them into topics. Using a machine learning algorithm, the first two sentences of each risk factor is analyzed, which generates topic categories.
These categories are grouped into 33 broader topics using AI-driven topic classification. This helps you identify which areas are of most concern across different sectors and industries.
Average Industry Peer: An imaginary benchmark
The Average Industry Peer represents the average of all the companies in a peer group. This fictional benchmark helps us see how individual companies compare to their industry peers.
Creating the Average Industry Peer
To create the Average Industry Peer, we count all unique risk factors referring to one specific topic, and then we divide it by the number of companies in the peer group. We disregard all values below one, and round the leftover numbers. This ensures that only topics relevant to each company are present in the peer group. The number of risk factors for the Average Industry Peer are then counted by the weighted scores of all topics included.
-
Counting topics in risk factors: Each company lists various risk factors in their reports which contain different topics. If a topic is important, the company will mention it multiple times. We count these unique risk factors for each topic across the peer group.
-
Averaging the count: Next, we divide the total risk factor count by the number of companies. For example, if there are 50 unique risk factors across 10 companies, we divide 50 by 10 for an average of 5.
-
Rounding and finalizing: Any average below one is disregarded because not every company in the peer group considers the risk significant. We then round the remaining numbers to determine how many risk factors the Average Industry Peer should have for each topic.
Weighted overlap
This section explains how we compare individual companies to the Average Industry Peer, using weighted overlap.
We only calculate a score for matching topics, otherwise the value is 0. Scores are calculated individually by topic, and then we sum up all values. For example, if a topic matches, then the value is calculated as x divided by the number of risk factors in the Average Industry Peer, where x is the number of times a topic is mentioned.
-
Matching topics: We only consider matching topics.
-
Scoring: The score is calculated as x number of risk factors divided by the number of risk factors in the Average Industry Peer, where x is the number of times a topic was mentioned by a company. However, x cannot exceed the number of times the topic is mentioned in the peer group.