Why policymakers need economic historians
Historical comparisons matter
In early May, I hosted a workshop on how economic history can help policymakers navigate the AI revolution. This is the final post in a three-part series inspired by those discussions. Here are the first and second posts.
Imagine you are a policymaker facing the AI revolution. The forecasts look grim: seismic shifts in the labour market, entire categories of jobs disappearing within a decade. What do you do? How do you prepare?
Perhaps you think of the introduction of the internet. If you are over forty, like me, you might even remember what it felt like. I can still hear the static symphony of the dial-up modem from my high school years, and recall the hours spent lost on Geocities, wandering through strange and wonderful early websites and wondering how I could use this incredible new world to make money.1
Yet, as Ben Schneider and Hillary Vipond argue in a recent working paper, the introduction of the internet might not be the best analogy for the AI revolution.2 Econ history postdoc Vipond, who attended our early-May Stellenbosch workshop, captured it neatly: ‘Having the long run to look at really highlights some of the limitations of our current thinking.’
At the workshop, the discussion repeatedly returned to economic history as a practical tool for policymakers. Used thoughtfully, it can support policy in several crucial ways. Firstly, economic history generates essential context, explaining the deeper origins of our present-day economic circumstances. Thanks to powerful new computational tools and expansive historical datasets, we can now systematically examine how past decisions shaped today’s conditions. Secondly, history provides policymakers with a rich repertoire of case studies – what Berkeley economic history professor Brad DeLong called ‘a huge filing cabinet of potential analogies’. And lastly, history guards against presentist thinking, reminding us our current challenges may not be as unique as we assume.
But choosing the right historical analogy matters deeply. Vipond cautioned against relying exclusively on recent history: ‘Looking at very recent events might frame our expectations and set them to a very inaccurate benchmark’. Her argument, elaborated in her paper with Schneider, is that the AI revolution more closely mirrors the labour-replacing disruptions of the First Industrial Revolution than it does the more recent digital and ICT revolutions. Those older shifts are distant enough that we can more fully grasp their long-term impacts.
Yet, what if the analogy you pick is the wrong one? DeLong shared a telling example at our workshop, recalling how former US Treasury Secretary Larry Summers invoked the stagflation of the 1970s when forecasting post-Covid inflation. Summers predicted persistently high inflation, but DeLong and others challenged his reliance on the 1970s analogy, arguing instead that post-pandemic inflation would be transitory. As DeLong put it bluntly:
We should be planning for policies so that we can deal with all reasonable scenarios rather than focusing on the one we think is most likely, especially if the reason we think it’s most likely is because it happened when we were there and at an impressionable age.
Summers’ analogy was not only influenced by recency bias; it was pessimistic, too. Princeton historian Bronwen Everill pointed out our frequent attraction to dramatic historical examples – hyperinflation in Weimar Germany, for instance – even when these extreme analogies are neither relevant nor probable. Everill noted how pessimism is often mistaken for realism, while optimism is easily dismissed as naïve. In crisis moments, particularly when theoretical consensus is absent, policymakers often exaggerate the severity of events, leaping to the most extreme historical comparisons available.
The analogies we draw are also shaped by the ways in which events are framed in public discourse; the media often has a much greater influence than academics. The Economist’s Africa correspondent, John McDermott, reminded us at Stellenbosch about the impact media narratives have on policymaking. He referenced two famous Economist covers – one from 2000 declaring Africa ‘The Hopeless Continent’, and a later one heralding ‘Africa Rising’. McDermott explained how these images continue to frame perceptions of Africa in business schools and policy seminars, often overshadowing nuanced reporting. Such stories shape the international narratives African policymakers must navigate.
While media framing often compresses complex realities into compelling but oversimplified narratives, consciously diversifying the analogies policymakers draw upon – academic as well as popular – can lead to better-informed decisions. Barry Eichengreen has described this as widening policymakers’ ‘decision set’. DeLong captured this succinctly at our workshop:
All we can really do is hope that we produce enough ideas and talk to each other about them enough, and judge them well enough. The answer is that we do that through discussion, through debate, through reasonable persuasion.
It is clear that most policymakers – like most of us – are not trained economic historians. And few could easily tell you what bootmakers in Victorian England have to do with ChatGPT today. Nevertheless, research shows that policymakers typically seize upon the first historical analogy that comes to mind, even if it is not particularly relevant or helpful.
This is precisely why QUB economic historian Chris Colvin and I are currently working on a paper we will present at the World Economic History Congress in Lund at the end of July. We argue that economic history, by embracing narrative and rigorous analysis, can better translate deep historical insights into actionable policy advice. Economic theory, as Colvin and Paul Winfree note, is most helpful to policymakers when universal. But a truly general theory is, by definition, only theoretically true. Economic history’s real strength lies in its acceptance of context and messiness: social, political, and institutional realities shape what’s possible.3
Comparative politics scholar Ken Opalo summed up this pragmatic approach perfectly at the workshop:
We must accept that our first, best solutions are often not the most practical solutions in context. Science doesn’t make decisions – people do. We may have the first, best solution, but we should be humble enough to accept that maybe the third-best solution is what is practical.
Such pragmatic, historically-informed thinking is precisely why economic historians deserve a seat at the policymaking table.
‘Why policymakers need economic historians’ was first published on Our Long Walk. Thank you for supporting my writing. The images were created with Midjourney v7.
I did try a few things, including building an online mall, something like the Yellow Pages but with fewer, larger shops, each with its own digital storefront. Books, too, of course. For a few months, I managed to convince an American telecoms company to become my first tenant. (They sent a cheque to my parents’ house every month by snail mail.) But a year later, someone with more smarts and tenacity launched an online bookshop that did rather well. He called it Amazon. Well done, Jeff.
Schneider, Benjamin, and Hillary Vipond. "The Past and Future of Work: How history can inform the age of automation." (2023).
Colvin, Christopher L., and Paul Winfree. "Applied history, applied economics, and economic history." Journal of Applied History 1, no. 1-2 (2019): 28-41.