Can AI be used to forecast economic conditions?

AI can do many things, such as your school assignments, compose work emails, or even code, but can artificial intelligence be used to predict macroeconomic trends, the stock market, and the next recession? 

Right now, the answer is no. However, it may not be as far in the future as you may think, with artificial intelligence models becoming more sophisticated day by day, accelerating. Deep learning, a method used to train artificial intelligence, can consider possible political, human activity, social, and environmental factors to combine them in a single model to give predictions. Traditional economic forecasting is affected a lot by human bias, thus being more subjective, therefore resulting in forecasts being often widely divergent. The advantage of using AI models is that they do not consider human bias, which can make mistakes, but more objectively pay attention to economic cause and effect, making them more comprehensive and precise. 

For example, recently, Yale University devised an AI model that could use texts from the "Wall Street Journal" on different topics, quantifying them and creating indicators to predict macroeconomic trends. This type of model is called sentiment analysis. The model showed a correlation between the prevalence of recession-related articles in "The Wall Street Journal" and the troubling economic times ahead. Furthermore, the model derived that articles related to "recession" could explain 25% of the variation in stock market returns. To analyze if the model can be used to predict the future, however, the research team tried to find a correlation between news articles and industrial production and employment in the next three years, and the model actually did pretty well; it found a link between articles mentioning "recession" and industrial production falling by 1.99% 17 months later and a 0.92% drop in employment 20 months later. Policymakers can use such a model to derive better solutions for controlling unemployment and economic slowdowns.

Another type of model is called high-frequency data analysis, where AI can process and analyze high-frequency data, such as real-time financial market data or transactional data, to identify short-term patterns and anticipate market movements. This can be particularly useful in forecasting economic events that may immediately impact financial markets. High-frequency data analysis accuracy, however, depends on many factors, such as:

Data quality: The accuracy of analysis heavily relies on the quality and reliability of the data being used. High-frequency data can be noisy and subject to errors, so data accuracy is crucial for obtaining reliable results.

Data relevance: The relevance of the high-frequency data being analyzed is important. It should be aligned with the specific economic or market variables of interest. Using irrelevant or unrelated data can lead to inaccurate predictions.

Modeling techniques: The choice of modeling techniques and algorithms plays a significant role in the accuracy of high-frequency data analysis. Different approaches, such as time series analysis, machine learning algorithms, or statistical models, may yield varying levels of accuracy depending on the specific context and the quality of data available.

Market efficiency: High-frequency data analysis assumes that markets are efficient and that relevant information is quickly incorporated into prices. However, in certain situations, such as during periods of extreme volatility or when there is a lack of liquidity, market inefficiencies can arise, making predictions more challenging.

It's important to note that while AI can enhance macroeconomic forecasting, it's not infallible. Economic forecasting is inherently challenging due to the complexity and unpredictability of various factors that influence the economy. AI models are only as good as the data they are trained on, and they may struggle to account for unforeseen events or structural changes in the economy. Human judgment and expertise are still crucial in interpreting AI-generated forecasts and making informed decisions.

Written by Islam Buleshov | Proofread by Yasmin Uzykanova