How Singapore investors can profit from unstructured data

Data that is collected in the business environment can be structured or unstructured. In general, structured data refers to information which is highly organised and which can easily be stored in rows and columns within database systems. On the other hand, unstructured data does not have a strict data structure, and is also not organised in a pre-defined manner.

Whilst data used in the business context is often well-structured and financial in nature, unstructured data represents a valuable but often neglected source of information that investors can use in their analysis of firms. In particular, a variety of unstructured data can be used to supplement and enhance their analysis of financial information traditionally reported by firms when making investment decisions.

Certainly, there has been a growing focus globally on using technology to gain insights from unstructured data. In a recent report titled “Finance reimagined: Finding long term value in a digital age” published by financial services company State Street, 2,000 investors and 500 investment providers were surveyed globally.

A key finding in the report is that the ability to combine both structured and unstructured data is set to become one of most important aspects of data that will define stewardship for the digital age.

In Singapore, there have also been efforts to leverage on unstructured data to gain insights. In a press release titled “DBS bank engages IBM”s Watson to achieve next generation customer experience,” the bank highlighted how it was using technology to deliver innovation anchored around analytics and the use of big data, both structured and unstructured.

In this article, I highlight two types of unstructured data – audio and text data - that exists in many firms, and which can be leveraged on by investors to gain insights into these firms.

Audio data
Potential sources of audio data include quarterly earnings calls, presentations to shareholders, and other audio recordings of public disclosures that executives routinely make. Such audio data can be analysed by investors to generate important insights beyond what is traditionally reported by a company in its financial statements.

For example, recent academic research suggests that a CEO’s affective state, as detected in his or her speech patterns (for instance, on earnings calls), may be effective in predicting the likelihood of a firm’s future stock price movements.

Further, there is also evidence to indicate that vocal and linguistic cues in a CEO’s voice may be predictive of possible irregularities in a company’s reported financial statements, which could eventually lead to restatements of reported financial results.

Text data
Text data represents another category of unstructured data that investors can leverage on when evaluating firms. Sources of text data include annual reports, regulatory filings, and other written disclosures made by a firm.

As with audio data, academic research has examined how text data can be incrementally informative in predicting future firm performance and events. For example, recent research has found evidence that the linguistic tone of disclosures made in the Management Discussion & Analysis (MD&A) section of a firm’s regulatory filings can be predictive of a firm’s possible bankruptcy in the near future and also of future firm earnings.

Incorporating the analysis of relevant audio and text data can be tremendously helpful. It can provide analysts and investors with a more holistic evaluation of a firm, and can provide valuable insights that would ordinarily not be available to them through traditional financial analysis techniques.

Therefore, even as traditional quantitative methods of financial analysis that leverage on structured data continue to form the basis of investors’ evaluations of firms, unstructured data – including audio and text data - can also be an important, complementary area of focus that can provide important insights and information.