Estonia's Ministry of Finance has released its economic forecast, which will be used to prepare next year's budget. Eesti Pank also produces its own economic forecast, and major commercial banks publish their future outlooks as well. Therefore, I will focus more broadly on the topic of economic forecasting.

We live in complicated times, when looking into the future has become extremely difficult. If earlier it was possible to forecast economic trends based on past events, now this has become increasingly complicated — if not impossible.

This type of past‑based forecasting is called adaptive expectations in economics. For example, if last year (and the two or three years before that) inflation has hypothetically been a stable two to three percent and no major shocks are expected, then one can quite calmly assume that prices will rise by roughly the same amount next year. Some say this is like driving a car using the rear‑view mirror, but in a society with stable development, such an approach works.

Looking at the picture more broadly, one can assume that several major global trends that began 15–20 years ago will continue for the next 10–15 years. This is the time period for which we intuitively believe we can make rational forecasts. This is the so‑called perceived future.

For example, we know that automated systems take over jobs that involve a lot of routine. Do we still have old‑fashioned milking machines in barns? Do young people know what a taxi dispatcher used to do? The concept of a bookkeeper is fading into the mists of time, replaced by the concepts of financial specialist or CFO.

Many people believe that at some point streets will be filled with self‑driving cars, that 4D printing will appear alongside 3D printing, and robots will take over cleaning jobs. Even the word "printing" no longer means traditional printing.

On the other hand, history offers many examples of how betting on trends led nowhere. Forty‑five years ago (in the early 1980s), people sincerely believed that the Earth had enough oil for only 25 more years. Another legend claims that in 1894 the London Times published an article predicting that in 50 years London's streets would be covered with several meters of horse manure, making movement difficult. In reality, cars appeared instead of manure. There is still debate about whether such an article actually existed, but the problem was clearly real.

Both examples illustrate that if we extend existing trends linearly into the future, the result may be wrong. New technologies emerge, people's behavior changes, institutions change.

In the year 2000, no one could have predicted that smartphones would change the world — the first Apple smartphone appeared only in 2007. Thus, extending linear trends into the future has two problems: a new technology may replace the old (cars instead of horses), or something completely new may emerge that we had no prior idea about.

One more important nuance. If something has existed for thousands of years — monasteries and monks, for example — then we can assume with high probability that such a phenomenon will exist for another 2,000 years. If a trend has lasted 20 years, then perhaps it will last another 20 years; beyond that, it is difficult to say.

Roy Amara formulated a regularity that still works well today: we tend to overestimate small technological changes and underestimate large technological changes. Many believe that streets will soon be filled with self‑driving cars — but they won't be filled immediately. Meanwhile, AI applications and large language models appeared practically out of nowhere, at least from the perspective of ordinary people.

What the impact of artificial intelligence will be on our daily lives may still be too early to say. We are only seeing the blossoms; the fruits are still ripening.

Earlier, it was thought that robots would replace humans in routine jobs, such as the bookkeepers or assembly‑line workers mentioned above. Now it is understood that AI is pushing into the domain of creative people: AI is used to write music, create art, make films, etc. With its help, it is possible to write books, poems, draft contracts, perform statistical analysis, visualize results, compile literature reviews, analyze texts, and so on. These are the "simple" tasks anyone can give to AI.

In addition, AI controls automated systems — other robots — composes complex logistical models for supply chains, analyzes big data, programs, deals with cybersecurity, invests, and more. AI has entered fields that we have considered the domain of humans, associated with adaptability, quick action, empathy, and creativity.

But that is not all. Alongside the "ordinary" megatrends — such as population aging, urbanization, the rise of social media, the shift of economic power centers to Asia, and increased human mobility — we are often hit by completely unexpected events, things we were not prepared for and did not foresee. Such "black swans" include the coronavirus crisis, the war in Europe, the Israel–Hamas war, and the blockade of the Strait of Hormuz. All of these have greatly affected our lives.

The world becoming increasingly unpredictable has made forecasting the future more difficult. We no longer have stable time series on which to build mathematical models, because there has been no stability.

Anyone who has worked with time‑series modeling understands that 15–20‑year time series no longer predict anything in economics, because for the last five years we have been riding a roller coaster.

In addition to looking at the past, the near future can be forecast through the expectations of economic agents. If one person or a group of friends believes that prices will rise next year, or some political parties say that the economy will improve, this may not affect the actual picture. But if the vast majority of society believes that prices will rise or, conversely, that the economy will improve, then that is what tends to happen.

The mechanism itself is simple. If everyone believes that for some reason (oil crisis, VAT increase) inflation will accelerate next year, then people start spending money "in advance." This increases demand, and when demand grows, companies raise prices. Simple economic logic.

If people believe the economy will grow, they start making larger expenditures and taking loans more boldly. Companies invest more, and when consumption and investment grow, GDP grows. This can happen even now — growth may come faster than initially expected.

This behavior is known in economics as rational expectations. Simply put: if most people believe something will happen in the economy, then it usually does. An individual may be wrong in their future outlook; all people cannot be wrong, because their collective opinion shows the direction in which the economy is moving. They constitute the entire economy.

The theory of rational expectations was established in the early 1960s in the United States. Two scholars, Robert Lucas and Thomas Sargent, made this approach a major part of macroeconomics in the 1970s. Both have received the Nobel Prize in Economics, though at different times.

Rational expectations work when we have some vision or opinion about the future. When unexpected events occur, these expectations are thrown into the trash. Would U.S. President Donald Trump have launched an attack against Iran if he had known that Iran could shut down oil traffic in the Red Sea? We cannot say in hindsight.

To what extent are human behavior and social processes predictable at all? We can observe certain behavioral patterns, find recurring expressions or calls in word clouds that may lead to uprisings, as was later acknowledged in the case of the Arab Spring. Does such an approach work for the future? Perhaps, but in an uncertain world likely no longer.

We arrive at the conclusion that the only reasonable weather‑forecasting models can look three to five days ahead; everything beyond that is a dark future. Translated into economics, everything beyond two quarters is uncertainty.

A harsh conclusion, but we live in difficult times. If someone gets economic forecasts exactly right, then they should also be advised to play the lottery — the person simply has a lucky hand. All of the above should be kept in mind when new economic forecasts soon arrive, both for this year and the years ahead.

--

Follow ERR News on Facebook and X and never miss an update!