How AI Will Redefine Economics
AI will bring economics closer to being a true science.
Tealfeed Guest Blog
Artificial intelligence is changing the world in unprecedented ways and redefining almost all fields of human endeavor. Economics, which has so far been resistant to the adoption of AI, is poised to undergo dramatic upheavals — arguably the greatest changes since Adam Smith’s inception of the science — and thereby help humanity better solve its problems.
AI, put simply, is a set of techniques that allows us to analyze data and make inferences, predictions, and, in general, make better decisions. Machine learning is a subset of AI that is capable of continually learning on its own, without having to be preprogrammed by humans.
Understanding how AI and machine learning are changing the field is essential to setting the right priorities and addressing existing issues that do not allow economics to be considered true science.
WHY ECONOMICS IS NOT LIKE NATURAL SCIENCE — AND HOW AI WILL CORRECT THAT
The main problem with economics and politics is that they do not constitute true “sciences”, such as physics or mathematics. In other words, there exists an apparent distinction between natural sciences and social disciplines that makes conventional scientific method inapplicable in social disciplines.
As the winner of the Nobel Prize in economics Paul Samuelson said, “Economics has never been a science — and it is even less now than a few years ago.”
Economists stick to their theoretical models the majority of which are of little to no value in the real world because they just can’t take account of the impact unpredictable and inherently irraitonal human behavior on the state of affairs.
Humans are irrational; their biases, prejudices, misconceptions and fallacies determine their actions. In social events, thus, people make decisions based on flawed knowledge.
Most economists, however, have been firm in their belief that markets reflect people’s rational expectations (leading to the creation of theories like the theory of rational expectations and perfect competition), despite the fact that social events are characterized by inherent indeterminacy. By the way, it is the inability to recognize our irrationality that led to the advent of neoliberal paradigm in economics and the 2008 crisis.
Economists stick to their theoretical models the majority of which are of little to no value in the real world because they just can’t take account of the impact human perception on the state of affairs and subsequent uncertainty.
Unlike physics, economics deals with people, their perceptions and irrationalities. A model that worked one day will not work the next time because of the rapidly changing moods of players.
Behavioral economics already recognizes that conventional economic theory is based on false assumptions. However, “With the lack of replication of data from its studies and a lack of understanding science and the science of behavior (behavior analysis), economic professors are just making guesses based on their experiences.” Clearly, behavioral economics, though aim to better perceive the world, has so far been unable to do so scientifically.
The implementation of AI and information technologies in general in economics, nevertheless, might change this situation for better. By modeling human behavior we can potentially estimate its impact on the real world. This will make it easier to identify financial bubbles and potential vulnerabilities in the economy, as well as to make better predictions.
(DE)POLITIZATION OF ECONOMICS
Due to the involvement of human thinking, economics can’t claim to live up to the standards of natural sciences. But the problem with economics goes beyond the flawed understanding of the world.
Scientists’ primary objective is the pursuit of truth. They endeavour to embrace an impartial attitude that allows them to analyze the real world in the best way possible.In theory, scholars of social sciences aim to attain the truth as well. In practice, however, this rarely holds.
Politicians’ main goal is not the achievement of truth by virtue of a better understanding of reality, but the pursuit of power, which is often attained thanks to the manipulation of reality. Economics, although far from being a true science, is nevertheless a source of clarity and objectivity in today’s increasingly partisan environment.
However, politics often interfere with economics.
As The Atlantic reports,
Since the 2016 election, the partisan economic expectations gap — that is, the difference between Republicans’ and Democrats’ assessment of the economy’s direction — has widened to an unprecedented level, going from roughly 20 points during the presidencies of Ronald Reagan, George W. Bush, and Barack Obama to 56 points today. Democrats now expect an imminent recession, whereas Republicans anticipate a robust spell of growth to just keep going. “The gap is just huge,” Richard Curtin of the Institute for Social Research at the University of Michigan told me.
Very often economic debates are turning into political, and economists, instead of dedicating themselves to the pursuit of truth, become advocates of their party’s platform, distorting or deliberately misinterpreting and ignoring available evidence.
In order to avoid the politization of economics, we will likely utilize AI to analyze all of the information about the economy and give impartial and objective evidence-based and data-driven suggestions. AI will put an end to endless politically charged economic debates and will provide us with balancd suggestions on how to tackle economic challenges.
ECONOMIC PREDICTIONS — NOW WITH AI
When it comes to predicting economic conditions, the only certainty is uncertainty. Economists are terrible at forecasting.
Prakash Loungani’s (IMF) analysis, for instance, revealed that economists had failed to predict 148 of the past 150 recessions.
There are many impediments to efficient economic forecasting. One is that it almost impossible to predict large swings in the moods of consumers.
Increased complexity is not the only problem — forecasts are also made less trustworthy because of a feedback loop. So if a meteorologist says it will rain, the fact that you take an umbrella out with you does not affect the weather. But if an economist forecasts that inflation will rise by 3% and we react by asking for at least a 3% rise in wages, we have changed the basis on which the forecast was made. Inflation is now likely to rise by more than 3%. The fact that the forecast exists changes the reality it is trying to predict.
Why economic forecasting has always been a flawed science, The Guardian.
AI-powered predictive analytics will overcome some of the challenges present in economic forecasting. By implementing AI techniques in behavioral economics, economists will be able to more precisely estimate the impact of human perceptions and behavior on the actual state of affairs. The use of predictive analytics will combine classical statistical analysis and the new world of AI.
AI algorithms can also analyze how media headlines influence sentiments about the economy. In fact, JPMorgan already uses an algorithm that tracks the effects of President Trump’s tweets on financial markets.
Central banks and fiscal authorities, by knowing when a recession hits, will be more effective and rapid in enacting monetary and fiscal tools, thereby mitigating the effects of business cycles.
We might even predict changes in supply and demand to implement necessary changes in order to avoid economic downturns.
As John Thornell has written in Financial Times,
All our connected devices are pumping out oceans of data about our real-time economic activities, demands and desires. Properly harnessed, that data could mimic the price mechanism as a means of matching demand and supply.
“The federal government is only scratching the surface in its use of information technologies to collect, analyze, and use data in innovative ways” write D. Esty and R. Rushing in Issues.
As I wrote in my article in Human Events, “instead of relying on personal experience, irrational instincts, misguided dogma, or harmful biases and prejudices, decision-makers could rely on data and A.I. algorithms.”
In theory, scholars of social sciences aim to attain the truth as well. In practice, however, this rarely holds.
AI will empower government officials to stop relying on traditional flawed methods of data collection and will instead base their decisions on data. Information technologies will make it possible to pursue data-driven monitoring, policy evaluation and analysis.
Thanks to the emergence of evidence-based economic policymaking, we will be able to put an end to endless economic debates and find common ground — based on data.
AI will make better predictions and factor the effects of human feelings (that is, reflexivity) on the actual state of affairs. Together, these changes will bring economics closer to a true science.
The advent of artificial intelligence is a single most important event in the history of economics. And our aim should be to effectively harness the power of AI to enhance the economic conditions in which we all can flourish.
This article was originally published by Sukhayl Niyazov on medium.
Tealfeed Guest Blog