Study finds people underestimate how often systems fail

People tend to think breakdowns are less common than they are, according to research by Lauren Eskreis-Winkler and colleagues. For everyday investors, the finding is relevant because several of the largest gaps show up in areas tied to markets, including drug development, business survival, debt, patents and consumer returns.

The researchers describe the pattern as a “failure gap”: the difference between what people estimate and the reported actual rate of a negative or unsuccessful outcome. Across the questions tested, respondents often guessed that failure, rejection or harm happened far less often than the benchmark figures used by the study.

The paper, available on SSRN, covers a wide set of topics rather than one industry. That makes the result less about a single blind spot and more about a broad tendency to undercount how often systems, products, institutions and personal plans do not work as expected.

Where the gaps were largest

Some of the biggest misses came in categories that show up often in business and public-policy headlines. The study reported that respondents estimated 58% of new drugs fail, meaning they do not receive FDA approval, while the benchmark figure was 85%. For patents, respondents put the rejection rate at 64%, compared with 86% in the study.

Entrepreneurship showed a similar pattern. Respondents estimated that 60% of new businesses started by entrepreneurs close, while the study’s benchmark was 80%. For restaurants, the gap was smaller: people estimated that 56% close within the first year, compared with 60% reported by the study.

Consumer behavior also showed underestimation. Respondents guessed that 22% of online purchases are returned, while the study listed the actual rate at 30%. On household finance, respondents estimated that 67% of Americans are in debt, compared with the study’s benchmark of 80%.

Health, security and education also showed gaps

The research found underestimation outside the economy as well. Respondents estimated that 14% of weapons or explosives present at airport security are missed by the TSA, while the study listed the rate at 70%.

In health care, respondents estimated that 33% of people will be diagnosed with cancer at some point in their life, compared with a 40% benchmark. They estimated that 45% of people will be diagnosed with a mental health disorder in their lifetime, compared with 50% in the study. On hospital hand hygiene, respondents estimated that 28% of U.S. hospital personnel meet national standards, while the study listed 50%.

Education questions showed wide gaps too. Respondents estimated that 36% of community college students graduate on time, while the study’s benchmark was 80%. For adults lacking basic math skills, the estimate was 26%, compared with 40% reported by the study.

The pattern across topics

The study’s list includes environmental and social measures with large differences between estimates and benchmarks. Respondents estimated that 25% of wildlife has gone extinct since 1970, compared with 69% reported by the study. They estimated that 38% of wastewater flows into the environment untreated, compared with 80%.

On human rights, respondents estimated that violations such as modern slavery exist in 21% of countries today, while the study’s benchmark was 85%. On air pollution, respondents estimated that 54% of city dwellers are exposed to levels far above World Health Organization safety standards, compared with 85% in the study.

The takeaway from Eskreis-Winkler and her colleagues is not that every person misses every risk by the same amount. Their finding is that, across many questions, people repeatedly placed failure rates below the benchmark figures the researchers used.

This story draws on original reporting from Klement on Investing.