Cause-Specific Smooth Transition Hazards for Stock Exchange Delisting: Estimation, Calibrated Testing, and an Application to Bursa Malaysia
This paper introduces a flexible cause-specific smooth transition hazard model to distinguish between voluntary and involuntary stock exchange delistings, applying it to Bursa Malaysia data to reveal distinct risk factors and timing patterns for each exit type while highlighting significant estimation challenges and test size distortions through simulation.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Every year, companies disappear from the stock market. For the most part, they leave in one of two ways. Sometimes, a company is bought out or decides to go private, a voluntary exit that usually happens when the business is healthy and an investor wants to take control. Other times, a company is forced off the list because it is in financial trouble, has broken the rules, or simply cannot pay its debts. For a long time, statisticians studying these events treated all departures as the same kind of occurrence, looking for a single set of reasons why a firm might vanish. But this approach misses a crucial detail: the reasons a healthy company leaves are often the exact opposite of the reasons a failing one is removed. If you mix these two groups together, the data becomes a confusing blur that tells you nothing about either process.
This is the puzzle that researchers set out to solve using data from Bursa Malaysia, the country's main stock exchange. They wanted to understand not just when companies leave, but why, and whether the rules governing these departures have changed over time. To do this, they built a new kind of statistical model that treats voluntary and involuntary exits as two separate stories happening at the same time. They also introduced a way to test if the rules of the game shifted during the study period, specifically looking at how the pandemic and the subsequent return to normal economic conditions might have altered the landscape for struggling firms. The goal was to see if the factors that predict a company's failure are different from those that predict a company's decision to leave, and if the timing of these events has changed in recent years.
The researchers gathered a massive amount of information on 836 companies listed on the exchange between 2016 and 2025. They tracked every single month these companies were on the market, recording their size, their share price, how often their stock was traded, and how much debt they carried. When a company finally left the exchange, the team did not just note the date; they read the official announcements to determine exactly why it happened. Was it a takeover? A court order? A failure to meet regulatory requirements? This careful classification allowed them to separate the 116 exits into two distinct groups: 62 voluntary departures and 46 involuntary removals.
Once the data was sorted, the patterns became clear. The two types of exits were driven by completely different forces. Involuntary removals, the kind that happen when a company is in distress, were strongly linked to small size, low share prices, and a lack of trading activity. These firms were often ignored by the market, with their shares sitting untouched for long periods. In contrast, voluntary exits happened among companies that were actually doing well. These firms had healthy prices and positive momentum, but they too suffered from thin trading. The researchers found that these healthy companies often went quiet for about seven months before a buyer made an offer. This suggests that voluntary exits are not signs of failure, but rather a result of investor neglect, where a good company is simply overlooked until a private buyer steps in to rescue it.
The study also uncovered a significant shift in how the market handles financial distress. For most of the decade, the risk of a company being forcibly removed remained relatively steady. However, the data pointed to a change occurring around December 2024. After this date, the overall risk of involuntary removal increased, and a new factor became critical: debt. Before this shift, having a high level of debt did not seem to make a company much more likely to be removed. After the shift, debt became a major warning sign. This change likely reflects the end of temporary relief measures put in place during the pandemic, which had previously delayed the removal of struggling firms. Once that relief ended, the market began to act on the underlying weaknesses of these companies, particularly those with heavy debt loads.
To ensure these findings were real and not just a result of the math, the researchers ran thousands of computer simulations. They discovered that standard statistical tests, which are usually trusted to confirm results, were prone to false alarms in this specific type of data. When the researchers used the standard test, it suggested the December 2024 shift was a strong, statistically proven fact. However, when they used a more careful, custom-built test designed for this specific situation, the evidence became less certain. The new test showed that while the shift was economically meaningful and the pattern was clear, the statistical proof was not as ironclad as the standard test claimed. The result is a finding that is highly suggestive and practically important, but one that requires a degree of caution in how it is interpreted.
The researchers also tested whether their conclusions depended on the specific mathematical shape they chose for their model. They tried four different mathematical frameworks, ranging from simple to complex, and found that the results remained the same regardless of the tool used. This consistency gives confidence that the patterns they found are real features of the market, not artifacts of a particular calculation method. Furthermore, their simulations revealed a hidden danger in similar studies: if all the companies in a dataset start their "life" on the exchange at the same time, it becomes nearly impossible to tell the difference between a company getting older and the calendar time changing. The researchers found that to see a gradual shift over time, you need a mix of companies that have been listed for different lengths of time. Without this mix, the data can create a misleading illusion of change where none exists.
In the end, this work provides a clearer picture of the stock market's exit doors. It shows that the market does not treat all departures equally. A small, neglected company is likely to be pushed out by financial distress, while a healthy, well-priced company might leave because it has been quietly ignored by investors until a buyer arrives. It also highlights that the rules of the game can change, as seen with the rise of debt as a risk factor in late 2024. By separating these stories and using more careful testing methods, the study offers a more accurate way to understand the life and death of public companies, reminding us that in the world of finance, context is everything.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.