Who gets to own
the future?

A conversation here in Sri Lanka began with taxes. We were discussing the difference between an employee whose income is taxed before it reaches their bank account and a business owner who can deduct eligible expenses and carry forward permitted losses. There are sound reasons for that distinction: revenue is not profit.¹ But the conversation led to a harder question. After years of keeping an economy going, what should people reasonably expect to be able to build?

An economy can recover while a household is still replacing the savings it used to survive. Wages can rise while the deposit needed for a home moves further out of reach. Technology can make work cheaper while leaving access to customers in the hands of the same few firms. Growth changes what people can earn and what they must pay to get started. Their freedom depends on the balance between the two.

That is the question behind ownership. Some people hold assets whose value rises as the future becomes more promising. Others must earn their way into that future at its new price. The rules governing entry, competition and protection against failure help determine how far apart those experiences become.

Sri Lanka makes the distinction between survival and advancement difficult to ignore. The Easter attacks of 2019 damaged tourism. The pandemic followed, then the economic crisis of 2022, default, shortages and inflation. External shocks exposed weaknesses that domestic decisions had helped create. The large tax cuts of 2019 left already fragile public finances less able to absorb what followed.²

Businesses kept operating and families found ways to manage. That resilience deserves recognition. It also has a cost. A business can survive by postponing investment or a hire. A household can draw on savings, cut consumption or borrow. Sri Lanka's official crisis survey records reduced spending and withdrawals from savings among the ways households coped.³ Some of what financed survival could otherwise have financed education, a better home or a business. The country keeps going, but part of its future has been spent getting through the present.

The World Bank's October 2026 assessment reports real GDP growth of 4.7% in the first half of the year. Output over four quarters had recovered to around its level in early 2018. Yet poverty was still estimated at 16.9% in 2025, compared with 11.5% before the crisis.⁴ A recovery in national output and a recovery in household prospects run on different clocks.

Nor does returning to an earlier level restore the opportunities lost along the way. A firm that has rebuilt its balance sheet may be healthier than it was three years ago and smaller than it might otherwise have become. The lost investment is absent from both the business and the comparison with its starting point.

Stability gives people a chance to invest again. Continued inflation, shortages and uncertainty would consume more of that chance. The difficult question is how to restore public finances while preserving the capacity to build.

Between 2021 and 2024, Sri Lanka's primary balance, which excludes interest payments, improved by 7.9 percentage points of GDP. The adjustment involved higher taxes, reduced real public-sector wages and compressed public investment.⁵ These contributors to the fiscal adjustment leave different households under pressure and different assets available for future use.

Spending more is not sufficient either. The World Bank's review identifies weaknesses in selecting and delivering public investment. An unused allocation and an insufficient allocation can leave the same road unfinished, but they call for different remedies. The return on a tax payment depends partly on the state's ability to turn money into something useful.

That return includes shared services, protection against risks and payment of past obligations. It cannot be calculated as an equivalent service owed to each taxpayer. But people can reasonably ask whether comparable means face comparable enforcement, whether useful projects are delivered and whether today's adjustment leaves a stronger foundation for tomorrow.

A household's foundation is easier to see in its home. Sri Lanka's 2024 census found that 84.6% of occupied housing units were owned by a member of the household.⁶ That is a tenure measure, not a count of debt-free individual owners. It nevertheless directs attention to a form of security that income alone misses.

A modestly paid household with secure accommodation and little debt may have more room to withstand a lost salary than a higher earner with large monthly housing payments. The second household may enjoy more comfort. The first has fewer cash commitments to meet when something goes wrong.

Shelter still cannot pay an electricity bill. Nor does living in a family-owned home give a young adult an asset they can sell or take to another city. Security can belong to the household while independence remains out of reach for one of its members.

Ownership is one way to make essential needs less vulnerable to a setback. Secure renting, liquid savings, insurance and dependable public services can do some of that work too. Their value lies in the choices they keep open when income falls. A buffer buys time. It does not determine the price of the next step.

Imagine saving an additional €5,000 during the year while the deposit required for the home you want rises by €10,000. Your wealth has increased. Holding everything else constant, your deposit shortfall has widened by €5,000. You have made financial progress and moved further from your goal.

Choosing a smaller home or moving elsewhere might close the gap. But that means changing the goal, with consequences for space, travel or family life. It is different from reaching the original goal more easily. Economic progress has to be understood in relation to what the additional money will buy access to.

Someone already living in the neighbourhood faces another calculation. Their housing payments may remain manageable while the deposit needed to enter the market rises. Maintaining a position and acquiring it are different financial tasks. An improvement for settled households can coexist with a higher barrier for those trying to join them.

The owner's apparent gain needs care too. A higher house price increases equity, but borrowing against it adds a liability. Selling releases money, but the next home may have risen in price as well. Appreciation does more for someone able to sell without buying equivalent accommodation than for someone hoping to move to a larger home nearby.

These distinctions matter because ownership performs several economic jobs. A home supplies shelter and can serve as collateral. Shares provide a claim on future company profits. A business owner can create an asset by adding productive capacity. A landowner may gain because a desirable location has become harder to obtain. All can appear as growing wealth, although their consequences for the next entrant differ.

The roles also overlap. Workers hold pensions and own homes. Entrepreneurs can begin with little capital. The useful dividing line is between the claims people already hold and the terms on which they can acquire another one.

Supply helps determine those terms. If higher incomes encourage more homes to be built where people want to live, additional purchasing power can finance additional accommodation. If sites, infrastructure or permission cannot respond, more money competes for much the same stock. Part of the income gain can then become a higher price for access.

This is why making buyers richer and making housing more attainable need not be the same policy. Deposit support can help a constrained buyer. When it expands what many buyers can bid without expanding supply, sellers may receive part of the benefit.⁷ The result depends on which obstacle is removed and what remains scarce.

Restrictions can serve legitimate purposes. Infrastructure has limits; construction affects neighbours and the environment. The economic question is whether a restriction addresses those costs proportionately or preserves a valuable position by preventing others from entering. An existing owner may gain from both a better neighbourhood and fewer competing homes. A prospective resident bears the cost of exclusion without yet belonging to the neighbourhood making the decision.

Finance can open a route and raise its price at the same time. Easier borrowing helps a buyer with sufficient income but too little upfront cash. It does less for someone whose earnings cannot support the repayments. Expanding credit without distinguishing those problems risks increasing bids without making the resulting debt manageable.

Tighter credit offers no painless reversal. A lower property price is little help if the mortgage payment rises or the buyer loses their job. ECB researchers modelling quantitative easing in four large euro-area economies found that employment gains among poorer households reduced income inequality, while relative wealth inequality changed little.⁸ An account based only on rising asset prices misses that benefit. The relevant test is whether someone can obtain and sustain a position, not whether its advertised price has fallen.

There is another advantage an existing claim can confer: access to gains before they are realised. If investors expect a company to become more profitable, its shares can rise before the additional profits arrive. Interest rates and the return demanded for bearing risk also affect the price.⁹ Expectations can be wrong. But an existing holder may be able to sell the claim now, converting part of a hoped-for improvement into purchasing power.

The new buyer pays for that expectation. They still bear the risk that future profits will disappoint. Owning a successful company and buying its shares at any price are very different propositions.

A worker receives gains through another route. Greater demand for their skills, stronger bargaining power or competition among employers can lift pay. Some responses are quick; patent research has found nearly immediate gains for incumbent employees in its setting.¹⁰ Others take time or never reach a particular worker. Someone saving from wages must continue meeting today's expenses while waiting. For an existing asset owner, the future can arrive in today's price. For a worker, it may take years to arrive in their pay.

This timing difference can alter who is able to buy the next opportunity. It also explains why higher income, greater wealth and an easier path forward should not be treated as interchangeable measures. A year of growth can improve one while leaving another unchanged. The price of entry is only part of what people see. The position reached by someone else is much more visible than the route that paid for it.

A beautiful home reveals little about the finances behind it. Its owner may have saved for years, received help from family or borrowed heavily. Someone who knows only the owner's salary can mistake an inherited starting point for a reward available from similar work.

The same problem appears in a successful business. The result is visible. Failed attempts, outside support and years without income are harder to see. Copying the visible choices may do little for someone who lacks the conditions that made them possible.

Visibility itself can affect well-being. When Norway made already-public tax records much easier to search online, the relationship between income rank and reported happiness became stronger. Ricardo Perez-Truglia estimated an increase of 29% in that relationship.¹¹ The result concerns the importance of relative income, not a 29% fall in happiness. It helps explain why learning more about other people's circumstances can change the experience of one's own.

Information can also open a route. Knowing what comparable workers earn may improve a pay negotiation. Understanding how a business found its customers can make success more reproducible. The useful discovery is how someone got there, including which resources were theirs to begin with.

But better perspective cannot close a real deposit shortfall. Comparison may reveal a material obstacle rather than create discontent from nothing. A person who earns more and still cannot establish an independent household has a problem that advice about expectations cannot solve.

Whether that frustration becomes a demand for government action depends partly on whom people hold responsible and whether they expect government to help.¹² Promising a better destination is therefore only the beginning. People must believe the route is open to them and that they can bear the cost of taking it.

Imagine a government proposing to restructure an industry. New investment should create better opportunities, it argues, but some existing jobs will disappear before alternatives develop. For the country, this is a claim about future gains. For a worker in one of those jobs, it is a decision about next month's income.

Rent, food and school expenses continue during a job search. Savings might cover a short interruption but not several years. Borrowing postpones the pressure while adding another commitment to uncertain earnings. A promise of greater prosperity in five years has less practical value to someone who cannot get through the next five months.

Even the ability to wait does not settle the matter. New jobs may require different skills, appear elsewhere or pay less. The industry could become more productive while the worker loses permanently. Calling the loss a transition cost does not make it temporary.

The uncertainty is personal. Someone can believe a reform will benefit the country and reasonably doubt that it will benefit their household. Fernandez and Rodrik showed how uncertainty about individual gains and losses can obstruct a reform that produces gains overall.¹³ Repeating the aggregate benefit does not answer the household's calculation.

Resources affect what people can withstand, not simply what they want. A secure incumbent may resist losing protection from competition. A poorer entrant may welcome change because the existing arrangement offers so little. Neither consent nor opposition can be read directly from a bank balance.

Protection can still make a risk worth taking. A French unemployment-insurance reform insured unemployed people who started businesses and increased firm creation. Employment in those firms largely displaced jobs at incumbents, but the entrants were more productive.¹⁴ Insurance helped people enter and changed where work was done, with little change in total employment. Security need not arrive through property ownership to support enterprise.

Protecting a person and protecting an incumbent position are different choices. Income support can give someone time to find other work. Keeping an unproductive firm permanently insulated from competition can prevent others from finding an opening. A transition policy has to decide which of those things it is financing.

It must also make its promises credible beyond the next election. The government may lose office before the gains arrive; a successor may abandon the policy. Meanwhile, the people facing losses know who they are. Some eventual beneficiaries may not yet know that they have a stake in the change.

Continuity is valuable when it allows useful work to accumulate. The World Bank's study of China's high-speed rail development identifies standardised designs, coordinated supply chains and sustained construction as factors that helped control costs.¹⁵ Equipment and expertise could be used repeatedly. That is an organisational achievement, more specific than simply having the power to keep building.

Building efficiently and choosing something worth building remain separate achievements. In its February 2026 assessment of China, the IMF identified property stress, local-government financing problems and declining returns from investment. It recommended shifting resources towards household consumption and addressing unsustainable financing obligations, alongside near-term fiscal support and later consolidation.¹⁶ The issue was the composition and timing of spending, not a blanket instruction to spend less.

An institution can become good at finishing what it starts and bad at admitting that the original case has changed. Contractors, lenders and officials may all have reasons to continue. The ability to sustain a good decision matters. So does the ability to correct a bad one.

Correction requires someone with both evidence and authority to act. In an experiment involving Indonesian village roads, increasing government audits reduced the gap between reported expenditure and independent engineers' cost estimates. Measures encouraging community participation had little average effect on that gap.¹⁷ An invitation to watch was less effective, on this measure, than a functioning check.

Elections provide one means of removing decision-makers. Courts, legislatures and public scrutiny can challenge conduct between elections.¹⁸ Their economic value depends on whether they expose failures and change decisions. A published target without a consequence for missing it can preserve the appearance of accountability while leaving control intact.

For the worker facing restructuring, a credible promise would answer practical questions. Are replacement opportunities appearing? Is support reaching the people who need it? Who must respond if the plan fails? Asking for patience is easier than accepting responsibility for what the waiting costs.

The same test applies to the opportunities offered in return. Training is useful when it leads to skills someone needs. A business needs customers as well as startup finance. Protection gives people room to try; productive possibilities determine what there is to try for.

AI now promises to make some of those possibilities cheaper. The question is whether it lowers the cost of establishing a viable position, or mainly improves the returns to positions already held.

The price of achieving a given level of performance on AI tests has fallen sharply.¹⁹ The economic opening is larger than a cheaper way to complete existing work. Some problems may become worth solving for the first time. A task that once required more specialist effort than a small business could justify might finally pay for itself.

Whether that happens depends on the full cost of a usable result. Someone must connect the tool to reliable records, check its work and take responsibility when it fails. A subscription is only one part of that calculation.²⁰

A software company's rollout in 2020–21 makes the opportunity concrete. Support agents, mostly in the Philippines, received suggestions drawn from previous interactions, which they could edit or ignore. A study of 5,172 agents estimated that issues resolved per hour rose 15% on average, and about 30% for less experienced and lower-skilled staff. The most experienced and highest-skilled agents gained little speed and lost some quality.²¹

Expertise became available where it was needed, but access still had a price and a process. Three-hour training sessions preceded access; limited trainer capacity slowed rollout. A contractual seat limit eventually stopped new onboarding. The study also found suggestive evidence of learning that persisted without the tool.

Making knowledge easier to reproduce changes who can perform a task well. It leaves a separate question about who controls the conditions of its use.

This is where a productivity result becomes a distributional question. Faster resolution might improve service, reduce staffing needs or allow the company to serve more customers. Pay depends on the alternatives employees can command and how the gains are shared. A tool can make a worker better at the job without making the worker more valuable in the labour market if it also enables many others to do the same work.

There is a related problem at the first rung of a career. IMF research identifies employment risks for younger workers in occupations where AI exposure is high and opportunities to work alongside it are lower.²² One possible mechanism is the removal of routine work through which beginners learn. If firms automate that work and recruit fewer juniors, learning may become faster for the people admitted while fewer people get in. The support study gives reason for optimism about the former; it does not resolve the latter. Better tools for learning do not, on their own, keep the door open.

The opportunity for smaller businesses is nevertheless substantial. Consider a firm that cannot justify commissioning a system to reconcile orders with stock records. If AI assistance reduces the cost of building and maintaining one enough, the firm may obtain a capability previously beyond its scale. Checking quantities, approving changes and correcting bad records remain real work. The gain exists when the time and errors saved exceed those costs.

The same threshold applies to a product or service. A lower fixed cost means fewer sales may be needed to recover the initial investment. An experiment that required a large commitment can become affordable to someone with less capital. That is a route from cheaper intelligence to wider entry, provided customers value the result enough to cover the remaining costs.

Judgment remains part of those costs. In a Kenyan trial, access to an AI business adviser produced no detectable average improvement in revenues and profits; initially weaker businesses did worse.²³ Choosing which advice to implement is harder than obtaining an answer. Cheap expertise can make an experiment affordable without making it a good experiment.

Employment depends on what happens next. A company that needs fewer staff hours per customer needs fewer hours for the same output. Lower prices or better service might bring enough additional demand to offset the saving. Growth could also come at a rival's expense. Customers may spend their savings elsewhere, creating demand in other activities, but those jobs need not reach the people displaced.²⁴ The conditions that make a business more productive and those that expand employment are different.

Nor do digital gains remove physical requirements. A robot needs components, power, maintenance and a place in a working process. A data centre needs a grid connection as well as chips. The International Energy Agency expects data-centre electricity demand to rise substantially, with grid and equipment constraints shaping how quickly capacity can be delivered.²⁵ More efficient models can ease those constraints; investment can expand supply. Their importance depends on how quickly supply responds relative to demand.

Housing makes the distributional consequence especially clear. Suppose AI sharply reduces the cost of architectural design. More households can explore layouts and obtain plans. That is a real gain. Its effect on the price of a completed home may still be small because land, finance, permission, materials and construction remain necessary.

If cheaper design encourages more people to build while suitable sites and permissions remain limited, some of the saving can be bid into the price of those scarce rights. The technology improves, but the holder of a different asset receives part of the benefit. Cheap plans alone do not raise land values; the effect requires additional demand meeting a constrained supply. AI could make designing the home cheaper long before it makes the home itself cheaper.

There is also a route through to the buyer. When sites and infrastructure are available and builders can expand, lower costs can support more construction. Competition can carry savings into prices, quality or choice. A household can then benefit without owning the technology, the building firm or the land in advance.

Who keeps a saving depends partly on who can prevent the next person from offering an alternative. That power may sit in distribution, essential data or access to infrastructure. If a business saves on production but must still pass through a gatekeeper to reach its customers, cheaper production does not remove the gatekeeper's position. If customers and suppliers can switch, and competitors can enter, keeping the entire saving becomes harder.

Temporary scarcity is less powerful than durable control. Capacity can expand, substitutes can improve and yesterday's bottleneck can become tomorrow's excess investment. Identifying a scarce input is therefore not enough to establish a lasting source of exceptional returns.²⁶ The important question is what prevents a response.

For a Sri Lankan business, access to AI can lower the cost of serving a foreign market without securing a place in it. Translation may make an enquiry easier to understand. Software may make a small order economical to process. Trust, product specifications, working capital and delivery still determine whether the sale can be completed.²⁷

Countries need not build the most advanced models to benefit from adopting and adapting them.²⁸ But the local complements matter. Dependable electricity, usable records and people able to judge an output can determine whether an inexpensive tool becomes a productive capability. The practical task is to identify the obstacle that remains after the technology becomes available.

Foreign competitors can use similar tools. A supplier may need to improve simply to retain its customers, while buyers receive better service and margins narrow. Payments to overseas technology providers may also send part of the revenue elsewhere. Local users can still gain. The benefits from using a technology and the profits from supplying it are different claims.

The hopeful possibility is that more people can test an idea before committing resources they cannot afford to lose. Someone who understands a product or a customer may no longer need to assemble every capability in-house before starting. When the business can reach demand, deliver reliably and retain enough value to continue, cheaper intelligence has widened the opportunity to build.

The conversation in Sri Lanka began with what people pay. It led to what their effort makes possible. Stability is part of the return, as are services and protection that households cannot readily provide for themselves. But recovering the means to survive is an incomplete answer to someone hoping to build a life.

Who gets to own the future? Existing owners can gain when expectations are capitalised into today's prices. Those who control scarce access can capture part of a saving made elsewhere. Entrants gain when lower costs are accompanied by a workable route to customers, housing or employment. Workers gain when demand for their contribution or their bargaining position improves. Consumers can gain through competition without owning any of the assets involved.

These outcomes depend on decisions. Permission to build affects whether additional purchasing power finances homes or bids up sites. Protection against failure affects who can afford to try. Accountability affects whether a failed investment is corrected or its costs are passed on to people with little power to object.

One distinction runs through these choices: rewarding the creation of useful capacity serves a different purpose from preserving the value of a position by keeping others out. The boundary is not always simple. Investment needs a prospect of return; restrictions can protect legitimate interests. But an existing owner's gain cannot, by itself, establish that the arrangement serves those trying to enter.

The test should be close to ordinary life. Does another year of work bring a home within reach? Can a business use a new capability to earn enough to continue? Can a worker change direction without putting the household's essential needs at risk? Greater output helps make those things possible. Its value depends on the routes through which people can claim a share.

That gives public policy a demanding task. It must protect people without freezing every existing position, enable investment without treating every increase in asset prices as a productive achievement, and make commitments durable enough to rely on without placing them beyond correction.

We should expect recovery and technological progress to leave more people able to make consequential choices of their own. The answer to the tax conversation cannot end with a country that has survived, or a market that has risen. People should not have to own the right assets in advance to benefit from what comes next.