Excerpt
Can the number or type of images in a firm’s annual report tell us anything about the firm? Or is it just a marketing strategy that doesn’t hold any further information? With the help of novel machine learning techniques, the authors Azi Ben-Rephael, Joshua Ronen, Tavy Ronen, and Mi Zhou study this problem in their paper “Do Images Provide Relevant Information to Investors? An Exploratory Study”.
Firms seem to use more and more images in their annual reports. The authors ascribe this fact to the so-called ‘information theory’, meaning that by the extensive use of images, they make the content more engaging to the reader and easier to understand. The images help to do this in three ways: they support the narrative proposed in the text, they help to focus the reader’s attention on what the firm wants to emphasize, and lastly, they simply make the text more engaging, thus keeping the reader’s attention. The authors oppose this theory to the theory that explains visual readability as a marketing strategy or a way of boosting positive sentiment in the public, for which they don’t find sufficient evidence.
This study is innovative because even though the readability of the annual reports has been studied before, it was never the informativeness of the images themselves. Visual readability, the term authors choose for this feature of the reports, measures how much the usage of images supports the narrated story in the text. The Google Vision API machine learning that the authors use allows them to explore the informativeness of pictures on the page level. That way, they are able to assign individual images differential weights depending on the layout of the page and allow for a more complex insight. Each picture is analyzed and assigned words that describe it, as you can see in the figures below. This way, ‘non-informative images, such as logos, can be distinguished from ‘informative images’, which help to underline the meaning of the text itself. The analysis is run within-firm, so other factors don’t influence the results.
Once the pages containing the visual elements are identified, the authors construct two metrics. The first one is PAV, or ‘the percentage of all visuals’, which is defined as the ratio of pages with any visual elements to the number of all pages in the report. Over the whole sample, this is on average around 10,2%. The authors find a positive association between this metric and the number of institutional holdings of the firm in the next year. It also seems to be associated with a positive change in the bond ratings, an increase in the firm’s sales, and in ROA.
The other studied measure is titled RFC, which is the increase in information reinforcement. In other words, whether the images used are related to the words in the text of the report. This also seems to be associated with an increase in the bonds ratings, lower analyst disagreement, and better forecast accuracy.
Authors: Azi Ben-Rephael, Joshua Ronen, Tavy Ronen, and Mi Zhou
Title: Do Images Provide Relevant Information to Investors? An Exploratory Study
Link: https://ssrn.com/abstract=3954219
Abstract: We introduce the concept of “visual-readability” in annual reports and use novel machine-learning algorithms to construct visual-readability metrics: visual use and information-content-reinforcement (the degree to which information content in images reinforces text). We find that increased news coverage and asset growth are associated with increased visual use. In turn, increased visual use is associated with subsequent-year realizations: lower risk and cost-of-capital, increased institutional holdings, and higher bond ratings, market share, and ROA. Further, the degree of information-content-reinforcement is associated with subsequent lower analyst disagreement and greater forecast accuracy. By and large, our results are consistent with an information-based story.
Visit Quantpedia to read the full article: https://quantpedia.com/is-there-any-hidden-information-in-annual-reports-images/
Disclosure: Interactive Brokers
Information posted on IBKR Campus that is provided by third-parties does NOT constitute a recommendation that you should contract for the services of that third party. Third-party participants who contribute to IBKR Campus are independent of Interactive Brokers and Interactive Brokers does not make any representations or warranties concerning the services offered, their past or future performance, or the accuracy of the information provided by the third party. Past performance is no guarantee of future results.
This material is from Quantpedia and is being posted with its permission. The views expressed in this material are solely those of the author and/or Quantpedia and Interactive Brokers is not endorsing or recommending any investment or trading discussed in the material. This material is not and should not be construed as an offer to buy or sell any security. It should not be construed as research or investment advice or a recommendation to buy, sell or hold any security or commodity. This material does not and is not intended to take into account the particular financial conditions, investment objectives or requirements of individual customers. Before acting on this material, you should consider whether it is suitable for your particular circumstances and, as necessary, seek professional advice.