Thinking in Models: A Mostly Harmless Guide to the Future

In our ‘Thinking in Models’ series, we draw on insights from NITI Aayog’s Scenarios Towards Viksit Bharat and Net Zero reports. Earlier pieces explored how modelling can guide India’s development and clean energy transition, highlighting gaps such as land constraints, the links between finance, growth, and decarbonisation, and the role of critical minerals in supporting India’s clean energy future. The final instalment moves beyond the NITI Aayog reports to reflect more broadly on modelling and its role. It serves as a reminder that the true value of long-term modelling lies not in precise predictions, but in fostering dialogue across scales—connecting the global and national to the local and lived.

Douglas Adams’ 1999 essay on Predicting the Future is a lot more eloquent (and fun) version of George Box’s overly cited maxim—’All models are wrong, some are useful’. The essay opens by declaring prediction a mug’s game, then delights in cataloguing the confidently wrong forecasts collected in the compendium Experts Speak. In the essay, Adams recounts hearing one such prediction in real time. A Vice President of an American phone company, arguing against high-speed wireless data connections, declared that ‘Granted, you can use it in your car going sixty miles an hour, but I don’t think too many people are going to be doing that’. This lasted barely a decade before qualifying for Experts Speak!

Nevertheless, we cannot do away with future projections, especially for policy planning. For instance, let’s consider India’s long-term strategy currently defined by the dual mandate of achieving Viksit Bharat by 2047—a developed nation characterised by economic prosperity, social equity, and administrative efficiency—and reaching Net-Zero emissions by 2070. Historically, the discourse surrounding these objectives has been shaped largely through macroeconomic abstractions, national-level sectoral aggregates, and linear cost-optimisation models that treat the economy as a homogenous entity. While the ideal paradigm is ‘model-based policymaking’ that objectively explores options and informs target-setting or the vision, the reality frequently devolves into ‘policy-based modelling’, where exercises justify pre-determined targets. But whether models inform or justify, their findings unfold not in the aggregate but within specific geographies and sectors defined by finite (and often fragile) biophysical resources. A farmer deciding what crop to plant next season, a mine closure authority tasked with repurposing exhausted coal assets, or a district administrator balancing local development priorities are all actors within the same transformation. Yet they operate within very different information environments, time horizons, and incentive structures.

Even within the climate action ecosystem, there is often a disconnect between communities focused on long-term mitigation modelling and those engaged in adaptation or grassroots development. To many practitioners working on the ground, century-scale emission trajectories can feel abstract or speculative. Meanwhile, long-term strategy modellers often work within highly technical integrated assessment frameworks that remain distant from everyday policy realities. The result is a counterproductive, self-reinforcing separation of disciplines. The challenge is not only to define national pathways but also to create ways for these diverse actors to see themselves within them.

In a sense, long-term modelling is stuck between a rock and a hard place: the rock being that predictions are most likely wrong, and the hard place being the need for these models to do much more than they currently do or offer. So, how do we, as modellers, use the tools we have (while also improving them) to provide meaningful inputs across scales and sectors?

Models built to inform national- and global-scale long-term climate strategies predominantly function as computational tools to reduce socio-ecological transitions to optimised pathways defined by cost curves, welfare metrics, and equilibrium assumptions. They are indeed valuable for testing policy and identifying the least-cost pathways to achieve a certain goal. But do they also narrow the space for imagination?

If they do, then rewiring the role of models and modellers becomes essential for reconciling the now with the future and the sectoral with the systemic. For instance, modelling the Viksit Bharat goal usually entails only a year-on-year improvement in the gross domestic product (GDP) relative to current trends. Can modellers create the means to facilitate collective inquiry into how this GDP trajectory will be achieved? Will economic expansion primarily emerge from energy and resource-intensive digital infrastructure, such as large-scale data centres? Could alternative development trajectories, such as regenerative agricultural systems or decentralised energy and industries, generate comparable value while reshaping social and ecological outcomes?

In a complex democracy like India, collective action rarely follows the simplicity of the graphs we construct. Arguably, the more critical challenge is to scale down the broader national vision for development, in contrast to the current discourse fixated on scaling up solutions. For instance, this could mean translating the national trajectory into state- and district-level targets and making clear what they imply for energy costs, jobs, and cultivation patterns—so that participatory dialogues can lead to informed, actionable decisions. Rewiring our approaches to build such a climate democracy is the task at hand.

To open up models and their results to a broader set of actors for collective inquiry, we need models that will go beyond giving us ‘an answer’ or ‘one optimal solution’. Instead, they should help us understand this interconnected, complex system we are part of and reveal the hard constraints and potential risks in the pathways our choices today will set us upon (the least-cost pathway may not be the least-risky or most constraint-free).

Making models more universally comprehensive may require rethinking our role as modellers. We are trained to simplify complexity through internally consistent frameworks, enabling analysis and comparison. But perhaps the next frontier lies not only in refining those models but in using them as tools to facilitate dialogue across scales and help different actors interrogate pathways, question assumptions, and explore alternatives together.

Going beyond models and traditional policy tools altogether is another pillar. Emerging cultural movements, such as Solarpunk, offer a striking example. Through visual art, speculative design, architecture, and storytelling, Solarpunk imagines futures where technological progress coexists with ecological restoration and community-centred development. Its landscapes are filled with solar infrastructure woven into everyday life, cities integrated with green ecosystems, and decentralised systems that blend innovation with local knowledge. While these visions are not policy blueprints, they perform an important function: they expand the realm of what societies perceive as possible. Art, in this sense, becomes a companion to modelling, but by helping societies visualise and emotionally engage with alternative futures that models alone struggle to convey.

• Shift focus from models to modellers and create forums for healthy debate there. Because after a point, models simply recreate the modellers’ narratives or ideas for the future (a modeller once inputted the climatic conditions of Venus into a popular model, and it still showed strong economic growth and prosperity!). So, we need to be able to contest these narratives in a participatory manner and enable consensus-building.

• Use modelling tools to scale down the national vision and as the base for diverse actors to come together and visualise their possible futures. Create spaces where farmers, district officials, industry leaders, and policymakers can examine pathways together.
• Refine models to reveal what is risky or unsustainable, too, not just what is optimal.
• Collaborate beyond models. Work with artists, storytellers, and speculative designers to broaden collective imagination. The technical and the creative need not be strangers (after all, an economist once told one of us, ‘Projecting GDP growth is more of an art than a science’).

Like any good model, this article may have raised more questions than it answered. We hope and believe that these are the right questions to ponder over. Reach out to us if you have the (non-42) answers!

So long and thanks for all the fish!

Links to the previous blogs in the series