Can Capital Market Information Predict Corporate Risk Before Financial Deterioration? Evidence from Sci-Tech Firms

In the world of finance, companies are often judged by their past performance, much like a student is graded on their last report card. For decades, lenders and investors have relied on standard financial numbers—how much debt a company owes, how much profit it made, and how quickly it collects money from customers—to decide if a business is safe to fund. However, this approach has a significant blind spot when it comes to technology and science firms. These companies often operate with very little physical assets, spend heavily on research that may not pay off for years, and face unpredictable future cash flows. By the time their traditional financial numbers start to look bad, it is often too late to prevent a crisis. The question researchers have long asked is whether the stock market itself, with its constant stream of price changes and trading activity, can act as an early warning system, spotting trouble before the official financial reports even show a problem.

A team of researchers set out to answer this question by looking at a specific group of high-tech companies listed on China's Science and Technology Innovation Board, often called the STAR Market. They gathered data from 454 of these firms over a five-year period, from 2019 to 2024. Instead of just looking at the final numbers, they built a sophisticated prediction system that combined traditional financial data with a wide array of market signals. This included how much the stock price jumped around, the specific patterns of that movement, how connected a company was to others in its industry, and crucially, how consistently the companies shared information about their research and development. They used a powerful computer algorithm, a type of machine learning that can find complex, non-linear patterns in data, to see if these market signals could predict which companies would face financial distress, regulatory trouble, or audit issues in the following quarter.

The study began by testing a specific idea: that the way a stock price moves over time contains more information than just its past volatility. The researchers found that a measure called conditional volatility, which captures how the market's expectation of risk changes from day to day, provided a significant boost to their ability to predict trouble. When they added this dynamic market measure to their model, their ability to correctly identify risky firms improved noticeably. However, they also discovered that the relationship between stock market jitteriness and company risk is not a simple straight line. It is not true that the more a stock price swings, the more dangerous the company is. Instead, the data showed a specific pattern where the risk of financial trouble was highest when volatility was in a medium-to-high range. When volatility became extremely high, the risk of immediate financial collapse actually dropped slightly, suggesting that extreme market panic might be driven by short-term shocks rather than the slow, grinding deterioration of a company's health.

Beyond the stock market movements, the researchers examined what the companies were actually doing and how they were talking about it. They found that the companies' own behavior regarding information disclosure was the single most powerful predictor of risk. Specifically, the gaps in a company's reporting on its research and development were a major red flag. When a company stopped sharing updates on its research projects or had consecutive periods of missing information, it signaled a high probability of future trouble. This was a stronger signal than the company's network connections to other firms, which surprisingly did not add much extra value to the prediction once the company's own market behavior was already accounted for. The study also highlighted that while high investment in research is generally seen as positive, the consistency of that investment and the quality of the company's cash flow were far more important indicators of safety than the sheer amount of money spent on innovation.

The final model, which combined all these different types of information, proved to be highly effective at sorting companies into risk categories. When the researchers tested the system on data it had never seen before, it successfully identified the companies that would face trouble with a high degree of accuracy. The system was particularly good at catching the companies that were most likely to fail, ensuring that very few high-risk firms slipped through the cracks. This suggests that for technology companies, the path to financial trouble is not always visible in the balance sheet until it is too late. Instead, the warning signs are often hidden in the way the market reacts to the company and in the consistency with which the company shares its story.

The implications of these findings are clear for anyone involved in lending to or investing in technology firms. Relying solely on traditional financial reports is insufficient for these dynamic businesses. A better approach involves watching the market's pulse and paying close attention to the continuity of a company's disclosures. If a high-tech firm suddenly stops talking about its research or if its stock price begins to show a specific pattern of unstable movement, these are likely early signals of trouble that appear long before the company's profits turn negative. By integrating these market signals and disclosure patterns into their risk assessments, lenders and investors can move from a reactive stance, waiting for financial numbers to deteriorate, to a proactive one, spotting the subtle shifts that precede a crisis. The study confirms that the market is indeed a powerful tool for discovery, but only if we know how to read its complex, non-linear language.