A week ago, protesters were swimming in Gotabaya Rajapaksa’s presidential pool.
This week, Rajapaksa left the country.
He flew to the Maldives on Wednesday as thousands of protesters returned to the streets, then travelled onwards to Singapore. On Thursday, he submitted his resignation. Parliament formally announced it on Friday.
Sri Lanka no longer has President Gotabaya Rajapaksa.
For the protest movement that spent months demanding “Gota Go Home”, that is an extraordinary achievement. A president who won a large electoral mandate in 2019 and belonged to the family that dominated Sri Lankan politics for much of the last two decades has been forced from office by a collapse in public legitimacy.
But the slogan has now accomplished what it was designed to accomplish.
Gota went.
Sri Lanka still needs a government.
That transition—from rejecting authority to constructing authority—may be considerably harder than removing Rajapaksa.
The immediate problem is Ranil Wickremesinghe. Under Sri Lanka’s constitutional arrangements, the prime minister became acting president after Rajapaksa’s departure, and parliament is preparing to elect someone to complete the remainder of the presidential term.
Constitutionally, this provides continuity.
Politically, it is deeply controversial.
Wickremesinghe is himself one of the people many protesters want removed.
His private residence was set on fire last weekend. Protesters subsequently occupied the prime minister’s office. Many view him as part of the same political establishment they believe failed Sri Lanka, and some accuse him of protecting or enabling the Rajapaksa system.
This creates a difficult collision between constitutional legitimacy and revolutionary legitimacy.
The constitution says one thing.
The streets may say another.
Which should govern?
My instinct is that the constitutional process has to remain the foundation, even when the political anger against the people operating through that process is legitimate.
That does not mean protesters should stop demanding change.
It means popular mobilisation cannot become an alternative constitution.
This distinction matters because protest has a powerful democratic function. Elections occur periodically. Governments can become deeply unpopular between them. Institutions can become unresponsive. Peaceful protest allows citizens to communicate political preferences outside the electoral calendar.
Sri Lanka’s protesters have done that with extraordinary effectiveness.
They demonstrated that Rajapaksa’s formal constitutional authority no longer translated into sufficient political legitimacy to govern.
But protest is much better at measuring intensity than distribution.
A million people in the streets tells us that a very large number of citizens care intensely about an issue.
It does not tell us automatically what fifty per cent plus one of the electorate wants.
People who stay home do not appear in photographs.
Different protesters may want completely different replacements.
Some want immediate elections.
Some want constitutional reform.
Some want a temporary all-party government.
Some want Wickremesinghe removed.
Some may favour particular opposition parties.
The protest movement can be united around rejection while remaining divided over construction.
This is why elections remain important even when street mobilisation appears overwhelmingly legitimate.
Voting gives every citizen an equal formal unit of political power.
One person.
One ballot.
Protest does not.
A person capable of spending three days demonstrating in Colombo has more visible political influence than an elderly citizen in a rural village who cannot travel.
That does not make protest undemocratic.
It makes protest a different democratic instrument.
Healthy political systems need both mobilisation and formal aggregation.
The danger begins when either claims to be the only authentic expression of “the people”.
Governments make this mistake when they say: we won the last election, therefore protests are illegitimate until the next one.
Protest movements can make the reverse mistake: we filled the streets, therefore parliament no longer possesses legitimate authority.
Both interpretations are too absolute.
Democracy is an ecosystem of institutions through which different kinds of political information become visible.
Elections measure broad preference.
Protests measure intensity.
Parliaments negotiate.
Courts interpret rules.
Civil society organises interests.
Journalists reveal information.
No single institution perfectly represents society.
The system becomes resilient when they constrain and inform one another.
Sri Lanka’s immediate challenge is therefore converting the political information produced by the protests into constitutional change without destroying the mechanism through which authority can be transferred peacefully.
That may require uncomfortable compromises.
The people capable of forming a government today are mostly politicians already inside parliament.
That is frustrating for protesters whose movement is partly a rejection of the political class itself.
But states cannot replace every experienced administrator and political institution overnight.
This is the same institutional continuity problem that appears after revolutions.
If everyone associated with the previous order is removed simultaneously, the new political leadership may discover that it has also removed much of the state’s administrative capacity.
If almost nobody changes, the revolution becomes cosmetic.
The challenge is determining which continuity preserves the state and which continuity preserves the failures citizens were trying to remove.
There is no simple formula.
Sri Lanka’s economic crisis makes experimentation particularly dangerous because the state has almost no time.
Fuel remains scarce.
Inflation remains severe.
Foreign-exchange reserves remain desperately limited.
The country remains in sovereign default.
Negotiations with the International Monetary Fund are continuing.
Every week of political paralysis makes economic stabilisation harder.
This gives experienced political actors an argument for continuity: the country cannot afford months of institutional reconstruction while essential imports are unavailable.
Protesters have an equally strong response: the political establishment claiming it alone can solve the crisis contains many of the people responsible for creating it.
Both arguments contain truth.
That is what makes transitional politics difficult.
Competence and legitimacy do not always exist in the same people.
A technocrat may understand debt restructuring but possess almost no popular support.
A protest leader may possess enormous legitimacy but little experience negotiating with the IMF.
A veteran politician may understand parliament but represent everything voters want changed.
Successful transitions somehow have to combine these resources rather than pretending one substitutes for the others.
This is where a temporary national or all-party government may have value.
Its purpose would not necessarily be implementing an ambitious ideological programme. It could focus on stabilisation, negotiating external financing, restoring basic supplies and establishing a credible timetable for elections and institutional reform.
Temporary governments, however, create another problem.
Temporary according to whom?
Political actors frequently enter emergency positions promising to leave once stability returns.
Stability is conveniently difficult to define.
This is why transitional arrangements need clear endpoints.
A date.
An election.
A constitutional mechanism.
Something that does not depend entirely on the temporary leader deciding when the emergency has ended.
Sudan demonstrated the danger of transitions without sufficiently credible transfers of power.
Libya demonstrated the danger of temporary governments whose electoral endpoints become contested.
Sri Lanka’s institutions are much stronger than either country’s, but the principle still applies.
Emergency legitimacy should contain an expiration mechanism.
Wickremesinghe’s position is particularly interesting because he represents almost the purest version of the legitimacy problem.
He is one of Sri Lanka’s most experienced politicians and has served as prime minister multiple times.
That experience may be useful during an economic emergency.
Yet his party performed disastrously in the 2020 parliamentary election, winning almost no constituency representation. Wickremesinghe entered parliament through his party’s single national-list seat.
He later became prime minister because Rajapaksa appointed him during the crisis.
Now constitutional succession has made him acting president.
Every step can be legally explained.
The cumulative result still produces an obvious democratic question: how much contemporary electoral support does the acting president actually possess?
Legality cannot answer that completely.
But neither can opinion on the streets automatically determine the replacement.
This is why early elections may eventually become necessary to restore political legitimacy.
They are not necessarily the immediate solution.
Holding elections during severe economic collapse carries costs. Governments need administrative capacity. Campaigns can delay urgent policy. Politicians facing an election may avoid necessary but unpopular economic reforms.
An IMF programme negotiated immediately before an election could become politically toxic.
Again, sequencing matters.
Stabilise first and risk allowing an unpopular political class to remain too long.
Vote first and risk delaying stabilisation while the economy deteriorates further.
There is no costless sequence.
The best option may involve a short, clearly bounded transition followed by elections once essential economic functions are sufficiently stable.
But that arrangement requires trust.
And trust is precisely what Sri Lanka lacks.
This is why transparency becomes particularly important during transitions. If citizens are being asked to tolerate politicians they distrust because those politicians possess necessary administrative experience, every major economic decision needs unusually strong public scrutiny.
What are the IMF negotiations producing?
Which creditors are being repaid?
Who receives scarce fuel?
Which taxes will increase?
Which subsidies will be reduced?
Which state enterprises will be reformed?
Who bears losses from debt restructuring?
These decisions cannot be treated as purely technical because they determine the distribution of the crisis.
Sri Lanka has already experienced the consequences of policymaking concentrated among too few people.
The solution should not be replacing one opaque decision-making structure with another simply because the new one calls itself technocratic.
Expertise needs accountability.
The debt problem itself reinforces this.
Public discussion sometimes searches for one villain responsible for Sri Lanka’s foreign debt: China, international bondholders, corrupt politicians, expensive infrastructure projects or the pandemic.
Reality is more complicated.
Sri Lanka borrowed from multiple sources. International sovereign bonds became a major part of its external debt. Chinese lending financed significant infrastructure and remains important, but describing the entire crisis simply as a Chinese “debt trap” would obscure domestic fiscal weakness, tax cuts, tourism collapse, declining reserves and policy mistakes.
Complex crises rarely have one cause.
Political narratives prefer one because one cause gives us one actor to blame.
Good analysis should resist that temptation.
China’s lending should be scrutinised.
So should commercial borrowing.
So should Rajapaksa-era fiscal policy.
So should the central bank’s decisions.
So should structural weaknesses preceding Gotabaya’s presidency.
Responsibility can be distributed without becoming meaningless.
Several factors can each be necessary parts of an explanation.
This matters for reform because diagnosing the wrong cause produces the wrong solution.
If the crisis were simply caused by Chinese debt, renegotiating Chinese loans would solve it.
It won’t.
If the crisis were simply caused by corruption, replacing corrupt politicians would solve it.
It won’t.
If the crisis were simply caused by COVID, tourism recovery would solve it.
It won’t.
Sri Lanka needs fiscal reform, debt restructuring, foreign financing, restored export and tourism earnings, institutional credibility and political legitimacy simultaneously.
That is much harder.
The protest movement achieved something concrete by removing the president most closely associated with the collapse.
That may improve political legitimacy.
It does not produce dollars.
This is the distinction between political accountability and economic causality.
Removing a leader can be justified even when removing them does not solve the underlying economic problem.
Accountability has independent value.
A political system in which catastrophic mismanagement produces no consequences creates terrible future incentives.
Politicians need to know failure can cost them office.
But citizens also need realistic expectations about what changing officeholders accomplishes.
Rajapaksa leaving does not refill petrol stations.
A new government still has to do that.
There is also a longer constitutional question about the Sri Lankan presidency itself.
Sri Lanka’s executive presidency concentrates substantial power in one individual. Critics have argued for years that the office contributes to excessive centralisation and weakens parliamentary accountability.
The Rajapaksa period intensified those concerns because constitutional changes strengthened presidential authority while members of the same family occupied several major institutions.
The current crisis therefore creates an opportunity to reconsider not merely who should be president but how powerful the presidency should be.
This is the difference between personnel reform and institutional reform.
Personnel reform says the wrong person held too much power.
Institutional reform asks why anyone should possess that much power.
The second question is usually harder because opposition politicians who criticise executive authority can become considerably more enthusiastic about it once they expect to control the executive themselves.
This is why constitutional reform is most credible when political actors design rules without knowing who benefits next.
Would today’s opposition still support reducing presidential power if it knew it would win the presidency tomorrow?
Would today’s government support parliamentary oversight if its opponents controlled parliament?
Again, the veil-of-ignorance test becomes useful.
Rules are strongest when people accept them even under unfavourable distributions of power.
Sri Lanka may now have a rare political opportunity for that conversation because confidence in the existing arrangement has collapsed so dramatically.
Crises create reform windows.
They also create opportunities for opportunism.
Emergency conditions allow institutions to change quickly because old vetoes weaken.
That can produce necessary reform.
It can also produce new concentrations of power justified as temporary solutions.
The distinction will depend on process.
This is why the behaviour of the protest movement matters after victory.
Its greatest achievement may be forcing Rajapaksa from office.
Its greatest contribution now may be ensuring that the transition remains accountable without attempting to become an unelected government itself.
Civil society is strongest when it can pressure institutions while remaining independent enough to criticise whoever occupies them next.
If protest leaders become absorbed entirely into government, the movement may lose that monitoring function.
If they reject every institutional compromise, they may contribute to paralysis.
Again, mature political action often exists between purity and co-option.
The events of this week therefore feel less triumphant than the images might suggest.
Gotabaya Rajapaksa is gone.
That matters.
A political leader who lost the ability to govern has finally lost the office too.
But states cannot exist permanently in the moment of overthrow.
Eventually somebody has to sign the budget.
Somebody has to negotiate with creditors.
Somebody has to decide fuel allocations.
Somebody has to explain which taxes rise.
Political authority has to become ordinary again.
That is the paradox of successful protest.
Its objective is often to make extraordinary politics unavoidable.
Its success eventually depends on making ordinary politics possible again.
Sri Lanka has demonstrated that citizens can withdraw legitimacy from a government with extraordinary force.
The next test is harder.
Can the political system construct a government people dislike less without destroying the institutions needed to govern at all?
Removing power is dramatic.
Rebuilding legitimate authority rarely is.
It happens through parliamentary votes, negotiations, budgets, elections and compromises.
Those processes will never produce images as memorable as protesters swimming in a presidential pool.
They will determine whether the protest actually changed Sri Lanka.
