India v. Pakistan: Who Would Win?

Charting the Last Wars of the Human Era

Prediction markets are giving a 24% chance of an Indian invasion of Pakistan before July. I don’t care to bet on this. (No alpha). The level of escalation so far is modest relative to previous crises. However, one novelty that hasn’t happened before is that India declared it will now divert the Indus River headwaters to for its own benefit. Long-term, this is an existential risk for Pakistan, most of whose 250M people subsist on the primarily meltwater-fed Indus. On a broader note, if some version of the “anthropic shadow” is true, we may expect weirder and weirder things to start happening as we approach the Singularity and the range of universes that don’t make it gets selected away. So it’s perhaps worthwhile considering the most potentially bloody pre-Singularity conventional conflict. (A China-US war will be primarily aeronaval, so while more military capital will be destroyed, it’s likely that much fewer soldiers will die).

India will win against Pakistan almost by default whether it’s a short, limited war (like all the four previous ones) or a long, total war. At this point, there’s just no comparison on any relevant industrial metric. India has a 6x population advantage (1.45B to 250M), GDP advantage ($4T v. $350B), or manufacturing output ($450B v. $45B). These are Russia v. Ukraine ratios c.2022 and that war Russia would have won there and then if not for Western aid regardless of how incompetent Putin and his generals were.

India now spends 10x more on its military than Pakistan - $84B to $8B in 2024 according to SIPRI - up from the 5x differential when its four previous short wars happened. For that matter, as we know, India won all of them even then, despite Pakistan having qualitative superiority in 1965. Nowadays India has better kit, though Pakistan has acquired 20 J-10s as well as HQ-9 and HQ-16 air defense systems from China.

Pakistan used to be richer than India. However, Pakistan is mired in semi-feudalism and patriarchy, its military controls large sectors of the economy, and it insists on pursuing Great Power pretensions that its economy cannot cash. India overtook it in the 2000s and never looked back. Even Bangladesh, an unassuming country that quietly focuses on light manufacturing, overtook Pakistan about a decade ago. So it has slipped in relative terms despite the higher population growth borne of its own persisting backwardness.

So there’s little point in commenting further if not for one big outside factor: The possibility of Chinese intervention. India’s MIC crushes Pakistan’s one on one, but 10% of China’s MIC working for Pakistan evens out the disbalance. China has a 5x bigger GDP than India, a 10x bigger manufacturing sector, and spends 4x as much on the military. The chart above shows how big the shifts in the balance of power are when just 5% of China’s GDP, manufacturing, military spending, and Nature Index score (proxy for science prowess) are added to the Pakistani side. If China made its industry preferentially available to Pakistan, that would even out the disbalance and conceivably degrade such a war into a long one of attrition.

This would be functionally similar to NATO devoting ~0.2% of its GDP to Ukraine and putting American space-based ISR capabilities at its disposal. This contribution, modest in absolute terms relative to what was spent on COVID stimulus and green energy, was sufficient to keep Ukraine in the game. China is at the same scale in orders of magnitude. Similar GDP to the US ($20T to $30T; already higher in PPP terms); 2x the manufacturing output; similar number of ISR satellites (both ~250) or 20-30x higher than both Russia’s and India’s ~dozen. Even more relevantly, China dominates the civilian drone market; both Ukraine’s and Russia’s drone industries are largely downstream of it. About 70% of Russian casualties in Ukraine now accrue to drones and this upwards trend will never stop. Drones also massively privilege the defense. Consequently, in a prolonged war in which China supports Pakistan, India’s armored and air superiority will become progressively less relevant.

Manpower will not be a limiting factor for either country for the indefinite future. It hasn’t been even in aged and post-nationalist Eastern Europe - Ukraine mans its armies with press gangs, and Russia with massive bribery. But here you have much younger societies with 10x the population and much higher levels of nationalism (nationalism peaking at the lower middle-income levels after universal schooling does its job but before higher wealth makes society smart/cynical and individualistic). In a prolonged war, Pakistan will get an inflow of jihadists from Afghanistan and the wider Muslim world. They won’t be of much use, but the main job of a soldier is to be drone fodder anyway, so that’s not all that relevant. Conversely, what India will enjoy in a longer war are the services of its moneyed and technically talented diaspora. They are strongly nationalistic, and I would expect them to play a central role in US tech transfers to what will in 2-3 years become the world’s second largest drone industry after China. This will prevent Pakistan from developing an overawning technological lead in drones through Chinese imports.

One disadvantage Pakistan has relative to Ukraine is that it is a relatively narrow strip and this makes it vulnerable to Indian air strikes and multi-pronged offensives. This is true with respect to stand off strikes, though this would mainly be relevant until prewar missile stocks are expended and before both sides learn to disperse. The Ukraine War suggests that will take just a few months. India’s air superiority wouldn’t be game-breaking in the way it would have been in the past; as we see in Ukraine, even last generation air defenses that are sufficiently thick outcompete air power these days, with the Russian Air Force relegated to dropping FABs on Ukrainian positions from within their own territory; and on the first days, we see Indian Rafales being shut down within their own territory, having been detected by Pakistani AWACS deep behind their border. The FABs are still ruinous for Ukraine, because it doesn’t have much manpower; but that notion would be laughable for Pakistan.

As regards land offensives, Pakistani vulnerability is something of an optical illusion. The southernmost border is sealed by saltwater marshes called the Rann of Kutch, which are near totally impenetrable to vehicles. The middle portion is defined by the Thar Desert, which doesn’t offer cover and is also logistically challenging to cross. Kashmir itself is mountainous and heavily fortified on both sides, and offensive actions in this environment are famously near unfeasible. So the epicenter of serious fighting, as in most of the previous wars, will be the narrow 250 km plain between the desert and the Himalayan foothills with their many population centers and dense road and rail networks. Already 200,000 Indian soldiers face off against 150,000 Pakistani ones there. I would expect this area to become the locus of the fighting, with India advancing on account of its manpower superiority, but at a snail’s pace and with heavy casualties as with Russia’s crawl through the western Donbass.

Will China support Pakistan? That’s obviously the million dollar question and one I don’t care to bet on. Supporting Pakistan would make sense. On the other hand, for all the talk about “wolf warrior” diplomacy, China’s actual conduct has often been surprisingly passive. It is internally repressive, but quite cautious and some (such as Chinese nationalists) might even say cowardly on the international arena. Most notably, it has aided Russia much less than I expected it to, instead preferring to focus on exploiting Russia’s lack of options to lock in favorable deals on natural resources. (This seems to be a pattern; African geopolitics is far more obscure, but China even failed to speak out in defense of Congo, with which it had mineral deals, against Rwandan aggression; the result may have been that the US has taken over as Congo’s main sponsor while France champions Rwanda). No Chinese support - Pakistan loses. But if China does support Pakistan with munitions, drones, and ISR, then I don’t see why this couldn’t become a long attritional conflict. Logistics will be harder than in Ukraine. China wouldn’t be able to supply Pakistan by sea, since the Indian Navy is much stronger and will blockade Pakistan; and even overland transport through the Karakoram highway will be difficult because of its extreme isolation, tortuous winding routes through the mountains, and vulnerability to blockages. However, China does have dozens of Y-20 military cargo planes it can fly directly to Pakistan, or (safer) staging areas in Afghanistan. China’s relations with the Taliban are as warm as they possibly can be between Islamic fundamentalists and atheists who otherwise send them to reeducation camps. And though the Taliban currently have frosty relations with Pakistan due to ongoing border disputes, I can’t see that continuing in the event of a “holy war” against India. I don’t think the Taliban will join the war directly, but they will likely allow China to set up logistics hubs and for mujahideen of whom there will be many tens of thousands to leave for the front.

To be sure India will also likely get support from the US, and to a lesser extent from the EU (it is more focused on Russia), as well as from Israel (technologically relevant). Although JD Vance has said that war between India and Pakistan is “none of our business”, this is unlikely to last in the event of Chinese intervention. Pakistan is extremely unpopular - it is an unprestigious country even within the Muslim world - so it can’t expect much organic sympathy. Russia will maintain studied neutrality; it is preoccupied with its own problems, and although friendly with India - reflecting ties that go back to Soviet days - it will not want step on China’s toes. Iran has good relations with India and will remain neutral; so will Bangladesh, which fought an independence war against Pakistan. However, by and large, it is only the US response that is relevant; and regardless of its precise extent, it will not cardinally change the big picture of a war defined by the struggle of Indian manpower and matériel against Pakistani manpower and Chinese matériel.

How do nukes factor into this? A 2007 study projected 12 million fatalities in India and 7 million in Pakistan - or 1% and 3% of their populations, respectively - in a countervalue exchange of 50 15-kt nuclear weapons. This is about the most catastrophic outcome. Though both sides have ~150 nukes in total, India needs to keep some in reserve to deter China, while many of Pakistan’s are tactical nukes intended to blunt Indian armored breakthroughs (i.e., have become a much less relevant consideration in the drone age). The idea of “nuclear winter” is more of an environmentalist myth and post-apocalyptic trope than anything scientifically concrete, and certainly wouldn’t apply here. I don’t expect either India, nor even Pakistan - again, contingent on Chinese support - to collapse on account of this. However, what I would expect to happen if nukes are used in anger is for the war to become truly total. In particular, I wouldn’t expect Pakistan to survive as a state in the event of its defeat. As I noted, Indians are very nationalist to an extent that liberal Western commenters are constitutionally incapable of comprehending. If things get that far, I expect currently minority (but not marginal) rhetoric about reconversion and expulsions to become practical policy.