Once upon a time, or so the story goes, Sri Lanka was a self-sufficient agricultural nation. We grew our own rice, we fed millions, we didn't need fertilizer, we didn’t need open economies, and we certainly didn’t need these new-fangled technologies and expensive imports to pull it off.

If you’re a Sri Lankan, you’ve heard this story at some point.

As talking points from a politician, as off-hand references from journalists, as ‘dad wisdom’ repeated at the kadey, or after a few drinks and bites. It’s ubiquitous.

Agricultural self-sufficiency (and thus, food security) is an important goal. On a basic, common-sense level, it prevents our food supply from being at the mercy of global supply chain disruptions.

The problem is that the story we tell ourselves doesn't quite fit the data.

Firstly, rice is our most important crop, with the majority of the agricultural workforce dedicated to its production. However, we have been importing milled rice at least since 1960 [1].

Lentils, sugar, fruit, milk and milk products - while there is some amount of local production, we rely on imports on all of these things. In the past sixty years, the closest we’ve ever come to being completely self-sufficient is 2006, and that was only in rice.

Think about what this indicates: every person alive right now has consumed (if not lived on) imported food, regardless of whether we ran a closed economy (pre-1977) or an open, liberalised economy (post-1977). We are not, and have not been, agriculturally self-sufficient. Such a time has not existed in recent memory.

What about ancient history?

Sri Lanka has been cultivating and trading agricultural goods for centuries. That much we know: maritime trade routes between Europe, the Middle East, and Asia used this place as a hub.

Often referred to as the ‘granary of the East’ or peradiga dhaanyaagaaraya in Sinhala, rice and other locally grown produce fed the residents of the island, while also forming a component of goods traded with other countries. At this stage in history, Sri Lankan farmers are believed to have used techniques such as animal manure, crop residue, crop rotations, shifting cultivation, and crop diversification.

These points are coupled with another: the idea that during the reign of Parakramabahu the Great, Sri Lanka — operating on these traditional farming practices — fed a population larger than the existing population of the country.

This is where things get complicated.

Estimating these population numbers are incredibly difficult. From what I know, we have no proper records from these times. Our most valued historical documents (like the Mahavamsa, Chulavamsa, and the Pujaveliya) focus on monarchs and religious affairs, not the numbers.

There are three texts that I’m roughly familiar with that have some numbers: the Pujaveliya and the the Nikaya Sangrahaya, which wrote about Parakramabahu's army (1170 A.D); and the Rajaveliya (1300 A.D), which shows up in early statistician’s attempts to make sense of the story of Sri Lanka.

The Pujaveliya states that “10 out of 100” men were selected for the army, and that the resulting army was “ek wisi laksha paswisi dahasak thora gena” - 2,125,000 soldiers. The Nikaya Sangrahaya, which potentially uses the same sources, backs this up.

Assuming the Pujaveliya army figures are correct, the “10 out of 100” figure would put the population at 21.25 million people - almost what modern Sri Lanka’s population is right now. [2]

The problem is that these numbers are wildly out of whack for their times. Here’s an easy visualisation by artist Martin Vargic of the largest armies in the world over time:

So assuming that the numbers are correct, this means that the single Sri Lankan army assembled by Parakramabahu was, for its time, far larger than the army of the Roman Empire (450,000), Genghis Khan’s entire operation (>900,000) and was almost twice the size of the army of China’s Ming dynasty (1,300,000).

Two million soldiers is almost the equivalent of the modern US army, one of the most devastatingly powerful empires of our times. It’s a number so large that it kind of defies comparison for that time period.

So here’s the question: if it existed in these numbers, what did this colossal army do, and what did it achieve?

Parakramabahu’s forces captured the Pandyan kingdom, and advanced deeper into Chola territory. The problem is that these exploits don’t mirror a strength of two million soldiers. The Sri Lankan army, if the Pujaveliya numbers are anywhere near correct, should have been one of greatest forces in the old world.

For example, the empire of Chandragupta Maurya stands as the largest political entity to exist on the Indian subcontinent - larger than the British Raj. Megasthenes (the Greek historian and sometime Indian ethnographer) and Pliny (the Roman author and military commander) both put Chandragupta’s force at around 600,000 infantry - with Pliny adding 30,000 cavalry, and 9,000 elephants [3]. Magesthenes regarded this as the largest army assembled in the ancient world.

Rome built an empire with 450,000 soldiers. The Mongols terrorised China so much that the Chinese built a wall that can be seen from space to keep them out. The Ming army of 1.3 million was raised from an empire of 103 million people. A Sri Lankan army of two million should have been so absurdly large that it could conceivably overrun all ancient armies of record. Yet the Sri Lankan invasions ultimately failed [4] and the larger post-Parakramabahu era was a tale of fracture and demise.

It is simpler to admit that the authors of those ancient texts were not statisticians; they were storytellers, and they exaggerated. Early historians often did. Pliny certainly did [5]. If they were, indeed, exaggerating, this blows a hole in the math of a 21.25 million estimated population for Sri Lanka.

But assume for a second that we did, by some strange turn of events, have a population approximately that of the largest empire in India.

What happened to these millions of people?

They certainly don’t show up in estimates around the 1000 AD mark. Avakov’s Two Thousand Years of Economic Statistics [6] estimates Song China to have had around 80 million people; the Chola Empire to have around 27– 28 million people; both the Holy Roman Empire and the Eastern Roman Empire combined to be around 24 million people. By 1500 AD China, under the Ming dynasty, had 103 million people; the Holy Roman Empire had 23 million; the Delhi Sultanate and the Kingdom of Mali both seem to have had around 20 million people.

Going by the mythic estimates, Sri Lanka in Parakaramabahu’s times should have been large enough to compete with Rome. Not just trade with everyone - we would have been in a position to colonise significant chunks of the world, to dominate history, art and thought. Sinhala should have been as widespread as Latin. The sandakadapahana should dot the ancient world.

And yet, by 1871, when census efforts began in earnest, population figures are around the 2.4 million mark. CICRED (the Committee for International Cooperation in National Research in Demography) monograph on the population of Sri Lanka from 1974 showed the increase.

This is almost a ten-fold reduction in population over 500 years, starting from an estimate of 21.25 million people.

To put it this way: imagine 8 out of 10 people you know dying.

So either a) the population numbers we can derive from ancient sources are grossly exaggerated, or b) pre-colonisation Sri Lanka failed so hard in its population support that we actually killed off most of our own population - a 500-year extinction that should have gone noticed.

Sir Emerson Tennent estimated that Polonnaruwa’s population under Parakramabahu would have been three million. Some historians think his numbers an overestimation [6], but even as such, they seem closer to the truth than a state the size of modern Sri Lanka.

What does this mean for agriculture?

It means that the evidence for this conversation is thin to nonexistent.

We’re working on a series on fertilizer, to which this topic is inextricably tied, but let’s look at farming practices. Ancient farmers had organic methods of farming. They may have kept ~three million people alive. However, we have no evidence for claiming that the same methods can scale to meet the needs of modern Sri Lanka. Our needs are far larger. Decades of different types of politics, despite the rhetoric, have not solved this problem.

Sri Lanka is no doubt ideal for so many types of agriculture. But the idea of a mythical state of affairs that could solve all our issues today - that is a story. It’s a Garden of Eden myth. It’s a story that makes us feel good about ourselves, but that doesn’t mean it applies today.

We don’t actually have a lot of data on how exactly the techniques of an ancient kingdom can map to a modern economy. Political speeches and the words of a few writers (including mine) aren’t enough.

Globally, the methods of the ancients are limited by the amount of nitrogen in the soil. As population starts growing, it becomes increasingly more difficult to scale up production without throwing in vast amounts of land and labour into the effort. We must remember that these methods worked for an age of spears and kings, a primitive economy powered by rajakariya (the right of a king to call up a population to labor on public works), with a significant portion of its population bent over a rice field.

While we don’t have accurate population breakdowns of Sri Lanka for those times, it seems highly unlikely that you can have a modern economy of lawyers, software engineers, CIMA-certified accountants, tuk drivers, call centre operators, and the like in any great number. Those who harp on these glory days today may not understand the role they will play in such a system [8].

I should note here that I appreciate stories; my personal belief is we are not Homo sapiens sapiens, the wise man, but Homo narrator, the storyteller. Stories are how we interact. Where they enable connection, they let us come together and conduct commerce, form kingdoms, build religions, and add to the fabric of history. But at some point, the math has to make sense. We need to look at these stories and ask:

But how?

The data for this piece

Rice imports by year, Sri Lanka, 1960-2021

Wheat imports by year, Sri Lanka, 1960-2021

CICRED 1974 monograph on Sri Lanka.pdf 6291125

[1] I used the USDA because while this data ultimately originates from Sri Lankan government authorities, the Sri Lankan sources lock data in PDFs that require an enormous amount of work to make usable in tabular form. The USDA data is far more convenient for this type of brief overview. Looking through USDA reports, they rely on organizations like the Sri Lankan Department of Census and Statistics and the Sri Lankan Department of Agriculture, with the US Foreign Agricultural Service (FAS) Colombo branch (operating out of New Delhi) providing synthesis and reports.

[2] This calculation is a very common one, often used to justify all sorts of strange claims. For example, consider this article on Sinhalanet titled “Sri Lanka only country in the world where “Genocide of Tamils” increases Tamil population!” or this recent paper from the University of Sri Jayawadenepura that imputes populations using the same technique and follows the rabbit hole to a claim that Sri Lanka had a population of 5 billion people at some point

[3] Mookerji, R. (1966). Chandragupta Maurya and his times. Motilal Banarsidass

[4] Mendis, G. C. (1996). The Early History of Ceylon and Its Relations with India and Other Foreign Countries. Asian Educational Services.

[5] Carrier, R. (2017). The scientist in the early roman empire. Pitchstone Publishing (US&CA).

[6] Avakov, A. V. (2010). Two Thousand Years of Economic Statistics: World Population, GDP and PPP. Algora Publishing

[7] Ross, R. R., Savada, A. M., & Nyrop, R. F. (1990). Sri Lanka, a country study.

[8] This is conjecture on my part, but odds are high that I will not be writing this article (unless I was a monk or a minister), and you certainly wouldn’t have the time to read it. We will both be in the fields, squabbling over an udella.