Philip Tetlock is a Canadian psychologist who, in 1987, asked a question that we all often ask ourselves: to what extent are expert predictions reliable? To try to answer it, he started a fascinating scientific experiment: over 18 years, he collected predictions about the political and economic future; by the end of the process, he had 27,500 predictions from nearly 300 of those experts. In 2005, he looked back, compared the predictions with what had actually happened during that long period, and came to a conclusion: experts are often wrong. But he was not satisfied with the result.
Consequently, he decided to take his experiment further and called it the Good Judgment Project. He gathered other psychologists and together they contacted 20,000 experts on political and economic issues, asking them to make precise predictions expressed as probability percentages about very specific questions, such as: what was the likelihood that a particular country would declare bankruptcy? Or that there would be a coup d’état in another? Some of these experts were given precise instructions on what was expected of them, and others were not. Some made predictions alone; others chose to work in teams, deliberate among themselves, and agree on their predictions.
After years of work, Tetlock and his team reached three logical and, at the same time, fascinating conclusions. First, people who received some training on the art of making predictions—such as how to neutralize their biases or how to use probability percentages—tended to be more accurate than those who did not. Second, they discovered that there are people, whom Tetlock called “superforecasters,” who have an extraordinary ability to correctly predict what will happen in the future, with a much higher and more sustained accuracy rate than others. But thirdly, Tetlock and his team discovered that, when making predictions, teamwork works.
Gather people with some talent for forecasting, tell them what you expect from them, invite them to talk, share information, and discuss, and generally their predictions will be much better than those who work alone. The future, indeed, is a conversation.
The Rise of Foresight
Tetlock has had a huge impact on the field of foresight. This is not a new discipline. But in the 1970s, understood as a tool for businesses and governments, it became more systematic and ubiquitous, incorporating scientific elements. Both the private and public sectors wanted future scenarios on issues such as fossil fuel reserves, lifestyle changes and consumer preferences, or the possibility of a nuclear war breaking out among the major Cold War powers. However, in the years following the fall of communism, largely due to the political optimism of the time, which assumed relative stability, the dominant branch of foresight was economic and focused on macro scenarios, market trends, and the potential impact of both on a particular sector or brand.
In recent years, however, with the return of geopolitical instability, fragmentation, and polarization, companies are paying more and more attention to political foresight. “The volatility and complexity of the current geoeconomic environment force both banks and the entire business fabric to strengthen and broaden the angles of prospective analysis,” says Alicia Coronil Jonsson, Chief Economist at Singlar Bank and member of LLYC’s Advisory Board. “We are moving in a context characterized by a historic combination of structural changes and new paradigms that define a new era, in which the global governance rules in place since World War II are progressively losing relevance.” For this reason, political risk analyses, which measure the impact that electoral cycles, government coalition health, or even more serious crises can have on an investment, are increasingly used. “Traditionally, companies primarily evaluated macroeconomic risks (cycles, inflation, interest rates, demand) because these were the main variables affecting their bottom line. However, growing political fragmentation, social polarization, trade tensions, and the disintegration of the world order have expanded the perimeter of monitoring,” says Coronil Jonsson.
To this end, some companies have their own foresight departments or hire consulting firms that offer these kinds of services. In many cases, risks are assessed through open-source analysis (facts and data that are public but require interpretation) or through contact with political insiders. Sometimes, this foresight literally requires a conversation: the human interaction between a decision-maker and an expert, or preferably a group of experts, is sometimes essential for the former to get a real sense of the most probable future scenarios and how they overlap with their own interests.
“It’s not only about managing risks and detecting challenges, but also identifying the opportunities offered by a world in full transformation,” says Coronil Jonsson.
When it comes to aggregating many opinions, as Tetlock recommended, surveys remain a relatively common tool, although in the economic and political foresight field they are usually conducted among experts rather than the general population. But new tools are also emerging to understand and design future scenarios. Recently, prediction markets that aggregate the forecasts of hundreds or thousands of people betting on a specific event have gained importance. “Aggregating the predictions of many people systematically beats the vast majority of individual predictions, even when we use ‘superforecasters,’” says Kiko Llaneras, chief editor of visual narratives and data at El País, and author of the book Think Clearly: Eight Rules to Decipher the World and Succeed in the Data Age. “The logic for this to work is intuitive: each person has slightly different information and also different biases. By aggregating predictions, information is combined and biases are smoothed out. The result, on average, is judgments better than almost anyone’s.” And in many cases, especially in recent years, this translates into a figure (or call, in industry jargon) that encapsulates the estimated probability that a given event will happen. This increases the clarity of the prediction and can help decision-makers who must choose among various investment or strategic options.
However, the unprecedented rise of social networks and the growing influence of digital conversation have led to the creation of tools that allow aggregating an even larger number of opinions. These are not from experts and therefore may lack reliability, but they are so massive that they provide real clues about the dominant ideas about the future in society, or about major consumption trends or political opinion. One such tool is LLYC’s Data Analytics Suite, which uses Big Data and Artificial Intelligence to identify conversation topics on social networks, which actors dominate that conversation, their relationships, the volume of conversation, and how and when it peaks or generates opinion shifts. These realities add a new layer to companies’ foresight exercises.
However, the new foresight, which uses these kinds of tools as well as the knowledge of expert predictors, and sometimes “superforecasters,” is no longer only a private sector tool. In recent years, many public and private institutions have also adopted foresight to advance geopolitical scenarios, consumption trends, or climate and demographic evolutions. Since 1997, the Office of the Director of National Intelligence of the United States government has published a report every four years analyzing long-term global trends; its latest edition, for example, focuses on 2040. In 2020, the Spanish Government created the National Office of Foresight and Strategy, which aims to anticipate future scenarios so that the country’s governance and legislation align with major trends. One of its first projects was Spain 2050, which brought together about a hundred academics from various specialties to conduct a “strategic foresight” exercise, in the document’s own words, to anticipate the “social, economic, and environmental” challenges Spain will face, it said, “in the coming decades” and to get ahead of them. In several Latin American countries, such as Chile, there is growing talk about foresight, or “anticipatory governance,” as a tool governments should adopt.
Reducing Uncertainty
Foresight is not the art of fortune-telling. As Coronil Jonsson says, its goal is to foresee multiple scenarios, “not as predictions but as tools to understand possible trajectories and prepare agile responses.” The only honest way to do foresight, Llaneras adds, “is to say things like: ‘I see a 20% chance that GDP will contract in 2026’ or ‘Candidate X will win the election with an 87% probability.’ Anything else,” he says, “can lead us to predict with overconfidence or to give up making predictions.” Ultimately, it is about reducing uncertainty, not guessing the future; about narrowing possible scenarios, not knowing exactly which one will occur. About adding technological, statistical, and methodological tools to peek, in a moment of uncertainty, at what is coming. And thus be able to contribute to leaving our mark on it.
To understand the future, reduce uncertainty, and help shape the major future trends, it is necessary to aggregate many voices.
There are many ways to do this. But Tetlock’s way not only seems intuitive, it has been scientifically proven to be the best: gather talented people, give them a clear methodology to compensate for their biases and think in terms of probability, clearly ask for what we need, and put them to talk. It is the recipe to increase the chances of success. And it is, in a certain sense, what the conversation for the future that LLYC proposes intends.