Using AI to Forecast Economic Data - Sevin Karabulut

 

Sevin Karabulut

(Image: OpenAI)

Artificial Intelligence (AI) has vastly increased productivity within business activities and has thus created a plethora of opportunities for various industries and corporations. The use of AI in economics is expansive. One could discuss how AI is becoming monopolised due to the small number of vendors relative to the majority of users who become more dependent on the technology each day. The numerous economic opportunities due to productivity and innovation, as well as the introduction of many new markets, could also be discussed. AI has also become widely discussed in the legal sector due to the legal and ethical challenges it presents; protecting intellectual property and the governance of AI. These, along with many more, are lengthy topics that cannot be discussed in a short blog, thus this blog will largely focus on the use of AI in economic forecasting. As mentioned, while AI does indeed bring forth many challenges, this blog will centre around how AI can be used by economists, policymakers and businesses for increased accuracy and efficiency.

Before delving into the discussion, it is important to first define AI and discuss how it works. AI programs are advanced language models, where an extensive amount of data is input during the training process,  the program then learns and picks up patterns from this, and combined with the intuitive processing algorithms, it predicts what will come next. Therefore, AI can not only mimic human intelligence, as it can solve problems, answer questions, analyse complex patterns, and use the information to effectively make predictions about the economic future, but can also do so in a fraction of the time a human can. In fact, the image at the top of the page is AI-generated and was done so in a matter of seconds, after I asked the program to generate an image that portrays advanced and futuristic technology in the business and economic world. 

AI programs can thus be immensely valuable in handling non-linear relationships. Linear regression is a statistical tool used to predict future values based on past values, often to determine trends in prices. While linear regression relates two variables (X and Y) with just a straight line (y=bx+a), non-linear regression accommodates more complex and non-linear (curved) relationships and can provide a better fit for real-world data. Widely used economic models, such as demand regression, which was the focus point of Week 2’s lecture, intertwined with other economic models, can often become far too complex. Determining the accurate analytical solutions of non-linear regressions would thus be too difficult. Alternatively, AI programs can calculate intricate patterns and relationships once extensive data is input into its training program, paving the way for accurate forecasts.

Moreover,  AI allows for more frequent forecasts to be achievable, hence providing another advantage to businesses and policymakers. This is because the economic forecasts are generally updated quarterly or monthly, but the AI program could perhaps generate real-time forecasts whenever data is provided. Large businesses could greatly benefit from this given times of uncertainty or economic volatility. For example, financial institutions could use this more productive method to forecast interest rate changes and adjust their strategies as such. Additionally, manufacturers can utilise AI programmes to better manage supply chains and avoid possible disruptions. On the other hand, smaller businesses in hospitality or tourism can also utilise AI-generated forecasts to manage seasonal demand fluctuations, and thus make decisions on inventory as such. Startups can also benefit from using AI economic forecasts as they can better predict cash flow trends and manage their budgets accordingly.

This blog has taken an optimistic view of how AI can be used in economic forecasting. However, this optimism must be adjusted with a critical look at the very possible issues AI brings forth. The following blog will discuss the legal and ethical issues of the use of AI in economic forecasting. For example, if inaccurate or biased data is used in the training phase of the AI program, the forecasts generated could be greatly flawed, resulting in businesses and policymakers making misguided decisions. This raises liability concerns – who is liable if AI-based decisions cause losses, the creator of the program or the user who willingly chose to act on such data? Issues regarding the sources of data input into the program will also be discussed, due to plagiarism and intellectual property rights.

As already mentioned, the topic of AI in economics is vast, and thus this blog is the start of a series that will further explore AI in economics, such as issues concerning the monopolisation of AI, the costs, and its impact on employment.


About the author

Sevin is a Business with Law student in her second year. With an interest in law, economics and business operations, she will aim to explore ways the law influences economic policies and current affairs.

Contact details for any questions or collaboration: bs23254@qmul.ac.uk

Comments

Post a Comment

Popular posts from this blog

How the 1973 Oil Crisis Changed the Way America Drives

Image
By Leandro Wiggett Introduction & Overview: What Happened In late 1973, members of OAPEC (Arab oil‑exporting countries) imposed an oil embargo on nations supporting Israel during the Yom Kippur War. The global oil supply was sharply cut, causing oil prices to 4x almost overnight. This was a major negative supply shock which rippled through economies: energy costs soared, inflation surged, growth slowed, and consumers and industries were forced to adjust to a new reality of more expensive fuel. The automotive sector, heavily dependent on imported, cheap oil, was among the hardest hit. Consumer Choices and the Shift to Fuel-Efficient Cars American households faced a straightforward budget problem: they couldn't afford both their existing vehicles and the now-expensive gasoline to run them. Their response shows the substitution and income effects working together. The substitution effect was immediate and dramatic. As petrol became more expensive, the relative cost of operatin...
Read more

The Hidden Costs of Vaping: Negative Externalities

Image
  Tamanna Suri (Edited by Sevin Karabulut) Despite UK regulations limiting the amount of nicotine in vapes, consumption of this persists due to smuggling from countries with looser laws. This blog will explore the negative externalities of vaping; it will discuss its impact on public health and NHS spending, thus government spending and the opportunity costs associated with this. Moreover, the blog will consider the UK’s rising national debt to discuss whether the negative externalities justify stricter governance of vaping.   Before delving into the case study, it is important to understand what negative externalities are and their impact on society. Negative externalities refer to the costs on third parties due to the actions of producers and consumers. This blog will focus on the negative externalities of the consumption of vapes in the UK: The graph above demonstrates the negative externalities of vaping consumption.  Here, the marginal private benefit (MPB) e...
Read more

London’s Rent Crisis

Image
Sevin Karabulut London’s rental market has experienced further rises in prices over the past year, due to two main issues: inflation and an increase in demand relative to supply. As a result, many tenants are unable to find affordable homes, leaving them worried and on edge.  The Office for National Statistics illustrates truly how severe the rental crisis in London is. It reveals that private rental prices in the UK rose by 6.2% from January 2023 to January 2024 (ONS, 2024).  Moreover, it highlights that London had the highest annual percentage change as opposed to other parts of England, such as the Northeast, which saw the lowest at just 4.7% (ONS, 2024). Figure: Private rental price percentage change (ONS, 2024) The continued rise in London's rental rate is the result of multitude of factors that increased demand for rental homes. This is illustrated in the diagram as a shift from D to D1 moves the initial equilibrium (E1) output from  P1Q1 to P2Q2...
Read more