Measuring the unmeasurable
Ahmed Tahoun’s research is dedicated to one aim: using data to reveal risk and its impact on firms and countries
In 30 Seconds
- LBS Associate Professor is driven by the desire to utilise data to reveal human behaviour
- Research focuses on deconstructing real-world problems in a way that can point to solutions
- Tool developed with colleagues adapts methods from computational linguistics to construct quantifiable insights for risk analysis
- Approach has many applications in academia and industry, including measuring the impact of new technologies and exposure to cyber threat
If you did not dig too deep into Ahmed Tahoun's CV, you might assume that he has always been an academic. As Associate Professor of Accounting at London Business School, his work has been widely published in many leading academic journals, and he was the youngest and most junior accounting professor when he joined the editorial board of the Journal of Accounting Research . He also has been the recipient of no less than eight consecutive research grants from the prestigious Institute of New Economic Thinking.
The fact that his work has been so widely referenced in the non-academic press around the world, however – including The Economist , The New York Times , The Wall Street Journal and the Financial Times – possibly speaks to the fact that he did his doctoral training at six institutions in five countries: the Wharton School and Chicago Booth in the US, the London School of Economics and the University of Manchester in the UK, Tilburg University in the Netherlands, the University of Valencia In Spain, and Cairo University in his native Egypt, where he was also a banker at HSBC (and where he gained first-hand insight into the world of finance – warts and all.)
The move into academia was not the dramatic career switch it may seem. His PhD dissertation had analysed why politicians invest in the stock market; he subsequently developed the thesis for a paper in The Journal of Financial Economics in 2014 entitled ‘The role of stock ownership by US members of Congress on the market for political favors’.
Common theme
The stark title of the research hints at its findings: “The ownership by US Congress members in firms contributing to their election campaigns is higher than in noncontributors. This bias toward contributors depends on the financial incentives of politicians and the relation’s value. Firms with a stronger ownership-contribution association receive more government contracts.”
If there is one common theme in Dr Tahoun’s work – really his North Star – it is about “trying to measure the unmeasurable” and use the findings to quantify the hitherto unquantifiable. Measuring the unmeasurable begins with data. He joined LBS in 2012 and the impressive body of work he has published since exemplifies his highly innovative, data-driven approach; one that has culminated in the development with fellow academics Tarek Hassan, Stephan Hollander, Markus Schwedeler and Laurence van Lent of a proprietary tool that can be used to predict the impact of systemic shocks; whether in the form of political events or global pandemics.
Turning raw text into decision-making data
The tool, NL Analytics, adapts simple methods from computational linguistics to construct measures of risk. It has key applications for policymakers – not least for central banks, who can leverage its insights to monitor country risk and sentiment as perceived by different sets of firms; analyse contagion and transmission of risk during crises; measure risks, costs and opportunities that firms associate with specific shocks, policies and regulations; and compare impacts and risks that are attributed to different sources.
The original methodology was developed with Tarek Hassan, Stephan Hollander and Laurence van Lent. It collates raw text transcripts and turns them into quantifiable insights.
First revealed in a 2019 Quarterly Journal of Economics article, ‘Firm-level political risk: measurement and Effects’ constructs a new measure of political risk faced by individual US firms – the share of their quarterly earnings conference calls that they devote to political risks – and finds that firms exposed to political risk respond by retrenching hiring and investment, and actively lobbying and donating to politicians.
There is another thread running through all Dr Tahoun’s research: topicality, allied to utility. All his projects begin with focusing on a real problem “out there” and attempting to deconstruct it in a way that can point to solutions.
This topicality is evident in building new methodologies for measuring and analysing the international propagation of uncertainty shocks at the firm level (see ‘The global impact of Brexit uncertainty’; accepted for publication in The Journal of Finance ) and for studying sudden and fast disruptive events as they unfold in real time (see ‘Firm-level exposure to epidemic diseases: COVID-19, SARS, and H1N1’).
Discover fresh perspectives and research insights from LBS
‘‘I always wanted to develop ways to capture behaviour; to measure it in numbers so you can clearly see patterns"
Identifying sources and transmission of risk
In another article, ‘Sources and transmission of country risk’, accepted at The Review of Economic Studies , Dr Tahoun (with Tarek Hassan, Jesse Schreger, and Markus Schwedeler) developed a methodology that measures the amount of risk managers and investors at individual firms associate with each country at each point in time. Flexibly aggregating this firm-country-quarter-level data allows for systematically identifying spikes in perceived country risk (“crises”) and › documenting their source and pattern of transmission to foreign firms. The methodology can be applied to show the transmission of risk due to the Russian invasion of Ukraine to firms around the world. (See Figure 1).
Ways of seeing
The Russia risk analysis is a comparatively straightforward application of the methodology, but it can also be used to analyse the apparently more nebulous effects of shocks such as Brexit and Covid-19 – not that, as Dr Tahoun says, he ever set out with the intention of developing such a tool.
“The starting point was never to create an application for Covid and Brexit, he reveals. “I come from Egypt and – I don’t want to label it corruption – but let’s say it was impossible for me not to notice that some politicians and firms often behave in very self-interested ways.
“When I left Egypt and went to live in the US and the UK I thought it might be different, but the truth is, it’s exactly the same! It’s really the same thing everywhere – it’s just that not everyone can see it. So the people who can see these things need ways to help others see them. And if you are one of those who do see it, what do you do about it?”
The determination to help that process of enlightenment was a profound motivation: “That always shaped the way I wanted to think about my research. I always wanted to develop ways to capture actual behaviour; to measure it in numbers so you can clearly see patterns. Opinions on people’s behaviour are highly subjective: numbers are not.”
The big picture
The methodology was put to effective use in determining risk stemming from diverse events in Italy by analysing foreigners’ perception of the Italian country risk (see Figure 2). Findings showed that financial firms were highly sensitive to Italian risk during the European sovereign debt crisis and the 2019 political instability.
The NL Analytics approach can also be used to compare the impact of different risks stemming from different (and unrelated) events to create a ‘big picture’ of country exposure, which is vital for markets and policymakers.
For example, Figure 3 shows how the tool produced a graphical comparison of the level of risk that US firms attributed to inflation, trade policy, Covid-19, the supply chain and Russia.
Analysing the risk of global cyberattacks
Another key area of research is the risk around cybersecurity. The World Economic Forum identifies systemic cyber risk as one of the most likely and impactful risks for firms: in the decade from 2011 to 2020, research shows that major institutions lost nearly $500 billion from operational risk events that were predominantly due to cyberattacks.
This explains why cyber-security ranks as the most challenging risk for firms, second only to political risk, in surveys of financial-market participants, and cyber-attacks pose particularly serious threats to trading and banking systems, as attacks on individual firms can develop into systemic crises if unchecked.
Co-authored with LBS’s Hélène Rey and Rustam Jamilov of the University of Oxford, ‘The anatomy of cyber risk’ uses data from more than 12,000 firms in 85 countries over the past 20 years to construct measures of firm-level cyber risk.
The study again uses the breakthrough methodology in the NL Analytics approach; namely, finding a way to capture the impact of messy, seemingly haphazard events by collating information from text transcripts. The research used quarterly earnings calls of firms and ‘sorted’ each cyber-related reference into category topics to capture sentiment, monetary loss, country names, and so on.
And it is not merely a diagnostic tool, as Dr Tahoun reveals, “Cyber-risk exposure has significant direct and contagion effects on stock returns. The research documents new facts concerning the worldwide rise of cyber risk and its industrial and geographical composition. The tool can characterise those firms most affected by cyber risk. We also believe our indices can predict future cyber-attacks.”
The research findings had such immediate significance for policymakers and corporates that the study soon found itself on the front page of The Economist .
Measuring disruptive technology
The NL Analytics tool has been developed and refined such that it has many broader applications in academia and industry.
The 2021 paper ‘The diffusion of disruptive technologies’, co-authored with Nicholas Bloom, Tarek Alexander Hassan, Aakash Kalyani and Josh Lerner, used the tool to try to quantify the impact of new technologies in the US.
Here, the ‘text-as-data’ methodology was used to track firms’ patents, job postings and earnings calls regarding disruptive technologies such as artificial intelligence.
The analysis enabled the team to document not only the development and spread of novel technologies, but also to illustrate the speed at which they spread across regions, firms, and industries – key elements in pinpointing economic growth, entrepreneurship and firm dynamics, but also crucial in helping to highlight economic inequality.
If the nature and complexity of Dr Tahoun’s work creates the impression of a solemn, data-driven academic who lives in the world of facts and figures, feedback from the students who attend his classes quickly dispels this illusion.
“Dr Tahoun injects humour, energy and real-world examples into what is a pretty dry subject,” attested one student.
“Exceptionally vibrant and a breath of fresh air,” declared another.
In a testament to the practical impact his classes often have, another said, “Since I started running my own business, I understand the value of a lot of things taught so passionately by Dr Tahoun.”
Student feedback also bears witness to an always-grounded approach: “Dr Tahoun provides simple messages that will stay with me for the rest of my career and have changed my perspective on how I view business operations and structures.”
Such acclaim helps explain why, among his many professional achievements to date, he cites his 2018 nomination in the Poets&Quants list of Top 40 Business Professors Under 40 as his proudest.
For further reading see the following two Wheeler Institute projects:
The diffusion of disruptive technologies – New research will help understand how innovations are spread across regions, industries, occupations and firms to generate key data insights for businesses
Systemic cyber risk – Establishing an interdisciplinary community of practice
Ahmed Tahoun is Associate Professor of Accounting at London Business School.
