The AI Economist
Predicting Prosperity in the Digital Age
For those of you who know, I am an economist and also a huge nerd about a fun little nit bit of things happening in the field which is always fun to see and gather the latest gossip or trends and stuff and of course, I think the intersection of science and the use of AI tools is in my neighborhood so of course something like the usage of Machine Learning (ML) tools in economics is no doubt a game-changer. Traditional economic models, once confined to the limitations of human analysis and historical data, are now being transformed by AI-driven predictive analytics. This shift is not merely an incremental improvement but a fundamental revolution in how economic forecasting is approached, and we gotta be very much aware of the impacts and implications this might have for our lives in the coming years.
Economics! huh! yeah, what is it good for?
AI's foray into economic forecasting represents a significant leap from traditional econometric models(which I suck at). Traditionally, economic forecasting relied heavily on past historical data and linear models to do their work and give out somewhat relevant insights to economists and then use that for policymaking or influence the public’s opinion about something – you know economists stuff...
However, given the complexity and dynamic nature of global economies, we are left falling short in being able to model and comprehend all the variables and their dynamics and relationships on a large scale to be able to forecast effectively certain phenomena in our economies. But, it seems that AI and machine learning algorithms, will be capable of handling this task in a better way since their capacity to work with vast datasets and uncovering intricate patterns, offer a more nuanced and accurate approach to predicting economic trends better than “old-school” economics trying to fit the complexities of the world into somewhat simple equations. So we might be seeing a very interesting shift in the field.
The Evolution of AI in Economic Forecasting
In recent years we have witnessed a surge in AI applications within economics. For instance, Wharton Professor Jules van Binsbergen co-authored a study using a machine learning algorithm to generate more accurate and detailed forecasts for GDP growth, employment, and interest rates. This advancement is part of a broader trend where AI economists are predicted to eventually replace human economists in many areas, leveraging unsupervised or reinforcement learning algorithms to process an almost infinite set of relations and variables. This goes back to the question old economic models had the incapability of working with all the relations and variables the REAL economy has so maybe AI economists will have a better capacity to work than human economists in these incredibly complex designs.
AI-driven predictive analytics are revolutionizing econometric analysis, transforming the economic landscape from forecasting and data analysis to decision-making processes and so much more a recently published research done by goddamned McKinsey's research team suggests that the global economic impact of AI could be monumental, potentially delivering an additional $13 trillion in global economic activity by 2030. This equates to approximately 1.2 percent additional GDP growth per year, which is incredible to think about we are already in 2023 which means one piece of technology could add to the global economic activity an entire China’s worth of GDP in LESS THAN A DECADE!
The Evolution of Economics from Credibility to Bots
Economic forecasting has its roots in traditional econometric methods. Initially, these models relied heavily on historical data and were often constrained by the limitations of human analysis and the finite nature of data sets also these models were constrained by the mathematical capabilities of economists in pre-computational eras or even during the beginnings of computation which helped but still were very much archaic compared to what we can do today. Due to human constraints, traditional economic models often had low prediction accuracy. The limited data available in these traditional methods made it a pain in the ass to control and forecast macroeconomic and development trends comprehensively.
Also, another revolution was happening in the field of economics before the meteor strike of AI and ML, a tectonic shift in the field that strode away from needlessly modeling overly complex things that resemble reality and then retroactively trying to fit the economy into them but actually using natural experiments and from those results derive insights, this is known as the credibility revolution. This revolution came from a movement of economists that started using data from phenomena happening in the real world and using those as sources to draw conclusions and then replicate or advise in policymaking based on previous results. This movement has garnered more and more support, especially after the 2021 Nobel of Economics to David Card, Joshua Angrist, and Guido Imbens for their work in fostering the credibility revolution. Also recently 2023’s recipient Claudia Goldin is also one of the spearheads of the credibility revolution researching labor differences among men and women in the labor force worldwide.
Thanks to the advent of the digital economy and big data these phenomena have significantly transformed the traditional means of economic analysis. This era marked the beginning of the transition from a purely statistical economy to what can be described as an “intelligent economy”, thanks to the enhanced connectivity and precise data sharing facilitated by the digitalization of most economic processes and companies. AI tools have significantly improved the accuracy and efficiency of economic forecasting. They are capable of processing and analyzing data at unprecedented speeds, allowing economists to develop more accurate models that capture complex relationships between various economic variables
These AI tools as well as other software (thank you Python, R, and fuck you Matlab) have automated many repetitive tasks like data cleaning, processing, and model optimization. This automation has freed up economists' time for more strategic decision-making and higher-level analysis, allowing them to explore complex economic relationships and conduct more nuanced policy analysis.
AI tools will continue to evolve, and economists are expected to play a crucial role in reshaping economic policy-making with these tools. AI-generated econometric models will likely aid policymakers in understanding the potential impacts of different policy decisions and in designing more effective interventions in complex economic systems. This means that there is a growing need for economists who are proficient in both econometric analysis and AI technologies(ooh I wonder where we could find economists who have in-depth experience in both… OH WOW LOOKIE HERE). Continuous learning and upskilling are essential to staying abreast of advancements in AI and machine learning, ensuring the effective utilization of AI tools in economic analysis and also in all fields, I would say if you are not making the most and learning to work alongside AI tools for your Job you will most likely be out of that job in the short to medium run...
AI-Driven Predictive Analytics in Action:
AI can and will be implemented in many different areas of economics and certainly, these applications will lead to a more in-depth and insightful overview and results coming from the field in the coming years. Maybe even some world-changing ideas about how we manage and deal with our economies now that we are approaching an ever-changing world.
It is a very interesting thing to see some of the scenarios done by a group of economists describing the transition from the digital economy to the “intelligent economy” and the repercussions this will have on the field of study and in our lives as well.
Evolution to an Intelligent Economy: The paper discusses how the integration of big data and AI is transforming traditional statistical economics into an intelligent economy. Traditional economic models, limited by human consciousness and small sample sizes, are being replaced by AI methods that offer more objective and comprehensive analysis this will inevitably lead to better conclusions and in-depth analysis, and even some emergent properties we are not yet aware might arise from this level of data analysis. .
Applications and Advantages: AI-based big data analysis methods are being applied in various economic sectors including stock analysis, industry analysis, capitalist economic development, and more. These methods enable rapid and efficient analysis of vast datasets, guiding policymakers in formulating more effective economic policies. This could be a great incentive for governments to rely more on AI-economists which could provide better results than purely human ones.
Advanced Modeling Techniques: The paper also proposes advanced economic modeling and forecasting methods using big data and AI. These methods account for the multiple factors influencing economic activity, using graph network structures and Long short-term memory (LSTM) models for comprehensive forecasting.
Stock Forecasting: A case study on stock forecasting using LSTM shows that the AI method effectively fuses multimodal data for accurate predictions, demonstrating a significant improvement in prediction accuracy. And, this for me is very interesting and something that could be a game-changer. The development of a sufficiently complex AI that could “beat” the stock market and turn into a constant consistent profit would be a “cheat code” for capitalism and whoever is able to have that tool and provide that to its investors and stakeholders kinda “wins” the game so this is something to be very much aware for in future (I might do an entire article just for this in the future)
But we can see the implementation of AI-driven forecasting has shown tremendous performance improvements across the board and soon we will be seeing more and more cases not only in economics but of course in other fields of study, technological development, and innovation. Leveraging AI techniques suitable for different data environments has enabled companies to significantly enhance their operational efficiency and decision-making processes of economists in their fields.
So, When do we get our EconoBots?
We can see that the application, and integration of AI are inevitable in the field of economics, the benefits on a technical level and the output increase that this could bring are incredibly positive and could lead to great policymaking for the global economy which would increase global welfare and enhance the economic prospects of the future greatly. This is always a noble endeavor that economists should pursue with their work helping us to better understand and predict the future of economic development and growth.
But we have to be careful because this trend of relying on and using AI economists does open the door for our policies to be driven by AI-biased economic forecasts and insights which puts a lot of pressure on how to design these systems to avoid abuses and reduce biases because if not we could be going into the dark side of automating analysis and giving us biased results that might lead us towards unfavorable economic decisions and not putting human welfare maximization at the center of it all and that is a key concern that we all should be aware.
This is not exclusive to economics, but, given the high regard we as societies put into economists (as we should) it is extremely important to be aware of the potential risks when designing these tools before full implementation and adoption by academics to prevent any unintended consequences.