Abstract:
This paper presents a novel approach to construct a weekly and monthly Twitter-based Economic Policy Uncertainty (TEPU) index for Lebanon, covering the period from January 1, 2011, till January 18, 2023. We have developed a unique and distinctive methodology that was specifically designed to overcome the challenges posed by the unavailability and lack of reliability of data in Lebanon. Our methodology enabled us to create a reliable and effective TEPU index for Lebanon. In this paper, we created and employed Python scripts that interact with the official Twitter API to fetch and transform tweets into actionable data. This curated data was the basis for generating a TEPU index for Lebanon.
Moreover, we employed two scaling methods to construct our TEPU index. The first was proposed by Baker et al., (2021), and the second was also proposed by Baker et al., (2021) as a variant of their main index and adopted by Lee et al., (2023) as their primary scaling method. We showed that despite some notable differences, the two TEPU indexes shared many common points, such as the general trend of economic policy uncertainty over time in Lebanon. We concluded that our TEPU index using the first scaling method was more reliable than the index using the second scaling method since the latter index tends to be inaccurate when outliers are present.
Finally, we conducted an event analysis to demonstrate how our TEPU index links to significant political and economic events that occurred in Lebanon throughout our sample period. We observed that our TEPU index significantly spikes during major political and economic events that occurred in Lebanon, effectively tracking the evolution of economic policy uncertainty.
