On this page
Loams has a line we use a lot: "Don't rent a cloud. Loam it." It borrows an old idea, that renting and owning are different relationships to the same thing. Before we lean on that idea, it deserves a fair look. This post is a short history of land ownership, what the evidence says about why it matters, the parts of the story that are not flattering, and then the comparison to software. It is a blog post, not a treatise, and we point to the sources so you can check us.
Land as the first family asset
For most of recorded history, land was the main form of wealth, and who could hold it was a political question. In much of medieval Europe, most people worked land they did not own, under obligations to a lord. Over the following centuries, rights to hold, sell and bequeath land spread unevenly: through common law, through registries, and, in some places, through deliberate policy.
The United States is a well-documented example of policy. The Homestead Act of 1862 let a settler claim up to 160 acres of public land and gain title after living on it and improving it for five years. Over the decades the act was in force, many families became owners for the first time. That is the part that is usually told.
Land rarely stays plain land. Over the twentieth century, in rich countries, a family's home became the biggest asset most households owned. The Federal Reserve's Survey of Consumer Finances has tracked this in the United States since 1983, and it consistently shows housing as a large part of middle-class wealth. Thomas Piketty's Capital in the Twenty-First Century (2014) documents how much of national wealth, historically, has been land and housing, and how much wealth passes through inheritance. A house or a farm that stays in a family is one of the simplest ways wealth moves from one generation to the next, and children of owners tend to start with more.
What the evidence says about secure property rights
The claim that secure ownership helps people invest is old, and the evidence is mixed but real. Three studies are worth reading.
- Ghana. Timothy Besley's "Property Rights and Investment Incentives: Theory and Evidence from Ghana" (Journal of Political Economy, 1995) found that in some regions, stronger rights over land were associated with more investment in it. The link was not uniform across the regions he studied.
- Argentina. Sebastian Galiani and Ernesto Schargrodsky studied a case where squatters on neighbouring plots got legal title at different times, because of how a court case unfolded. Their paper, "Property rights for the poor: Effects of land titling" (Journal of Public Economics, 2010), found that titled households invested more in their homes. It did not find a large rise in access to credit, a result that cuts against a popular story.
- Peru. Erica Field's "Entitled to Work: Urban Property Rights and Labor Supply in Peru" (Quarterly Journal of Economics, 2007) found that titled households worked more hours outside the home, since they no longer had to guard their property.
The popular version of this idea comes from Hernando de Soto's The Mystery of Capital (2000): that informal ownership is "dead capital" until it is written down. Economists have tested it and found parts of it hold and parts do not. The fair summary is that secure rights tend to raise investment, and that titles alone do not fix everything else.
The part that is not a success story
A history of ownership that stops there would be dishonest. Ownership has often been denied to people, and taken from them, by law and by force. A few examples, briefly.
Enclosure. In England, between the sixteenth and nineteenth centuries, common land that villagers used for grazing and fuel was fenced and consolidated. Historians still argue about how much harm it did to whom, but J. M. Neeson's Commoners: Common Right, Enclosure and Social Change in England, 1700-1820 (1993) shows how many people lost use rights they had held for generations, and received little in return.
Colonial dispossession. In North America, the same Homestead Act that created owners opened land that was already home to Native nations. The Dawes Act of 1887 split tribal land into individual allotments and opened the remainder to sale, and Native landholding shrank sharply as a result. In Australia, the legal doctrine of terra nullius, which treated the continent as owned by no one, stood until the High Court's decision in Mabo v Queensland (No 2) in 1992.
Redlining and discriminatory law. Richard Rothstein's The Color of Law (2017) documents how federal, state and local policy in the United States steered Black families away from homeownership and from the neighbourhoods where homes gained value: through lending rules, covenants and zoning. The University of Richmond's Mapping Inequality project hosts the original 1930s maps. Economists Aaronson, Hartley and Mazumder, in "The Effects of the 1930s HOLC 'Redlining' Maps" (American Economic Journal: Economic Policy, 2021), found that the maps are linked to lasting differences in homeownership, house values and segregation in the affected neighbourhoods.
None of this is a footnote. If owned land helps families build and pass on wealth, then being shut out of ownership, or having it taken, does the opposite, and the effect also passes down.
Renting and owning, in software
We do not think software is land, and we are wary of any analogy that makes a database sound like a moral question. The comparison we want to make is narrower, and it is about the structure of the relationship.
A tenant has a home but not the terms. The landlord can raise the rent, change the rules, or decline to renew. The tenant can leave, but only by paying to move, and some of what they built stays behind. An owner carries real costs the tenant does not: maintenance, taxes, risk, and the work of running the place. Owning is not better in every case. Plenty of people rent on purpose, and plenty of teams should rent cloud services for the same reasons.
Renting a cloud works the same way. You build on someone else's platform, in formats and APIs that suit that platform. The price list is theirs to change. Your data may be easy to put in and costly to take out, and your account can be suspended by a policy you did not write. A good provider behaves well on all of this, and many do. The point is that the terms are theirs.
The alternative we are building is to keep the ground. With Loams, your compute runs in your account, and your data sits in your bucket, in open formats, whether that bucket is S3, GCS, Azure Blob, RustFS or a folder on your disk. Loams is free and open source under Apache-2.0, and it runs on ground you own. If you stop using Loams, your data is still there in formats other tools can read. On that ground you can run coding agents side by side in The Field, keep their toolchain beside them in The Greenhouse, and open it all from the console that ships with your deployment.
Be honest about the limits. A bucket on a public cloud is still on someone else's hardware, so you are still a tenant at the hardware level. What changes is that the data is yours to move, in formats that do not need our software to read. You also take on operations: you run it, you upgrade it, and when something breaks you are the one on call. We say this in our own docs too, and our post on what Loams is bad at today lists the gaps.
Why we use the word
Land ownership helped many families build and pass on wealth, and it was withheld from many others by laws we can read and maps we can look at. We would rather keep both of those facts in view than tell only the cheerful half. The lesson we take for software is small and practical: terms you set yourself are better than terms set for you, and it is easier to keep ownership than to win it back.
So we say it plainly. Don't rent a cloud. Loam it.