In the summer of 1982, Tasmania was building a dam on the Franklin, one of the last wild rivers in Australia. The year before, the state had held a referendum that asked voters to choose between two dams. There was no box for neither, so one third of people wrote "No Dams" across the ballot instead.
On 14 December 1982, the day the river was listed as World Heritage, protesters began a blockade. Over the next three months more than 1,200 were arrested, but construction went on.
Three days before the federal election of March 1983, the Tasmanian Wilderness Society ran full-page ads in the Sydney Morning Herald and The Age. Each carried a photograph of the river and one line: "Could you vote for a party that will destroy this?" The Wilderness Society also put volunteers into more than a dozen mainland marginal seats, and on election day 13,000 of its supporters staffed the polling booths.
Labor, with a new leader and the economy in recession, won in a landslide. It had promised to stop the dam, and it kept the promise. The new government passed a law to protect the river, and in July the High Court upheld it, four judges to three. There is still no dam on the Franklin.
Who changed their mind?
Hardly anyone. Labor held no seats in Tasmania, where the dam meant jobs, so it had little to lose. It needed marginal seats on the mainland, where most people had never seen the Franklin until it appeared in their newspaper and volunteers met them at the polling booth. The photograph worked because it spoke to something mainland voters already felt about wild places, even ones they would never visit. For Labor, that made stopping the dam a promise worth making.
Without a movement, the people who profit from a harm usually win. There are few of them and they each have a lot at stake, so they organise. The many who pay for the harm each pay a little, so they don't organise. The economist Mancur Olson called this the logic of collective action.
A movement is what takes millions of people who each care a little and turns them into votes a politician has to count. That changes what it takes to win, whatever the politician believes.
Economists found a natural experiment that backs this up in the United States. In 2009 the weather helped decide how big the Tea Party rallies were, because rain kept people home in some towns and not others. Where the sun was shining and the rallies were bigger, local members of Congress voted more conservatively afterwards, and Republicans won more votes the next year.
Why some movements win
Movements win when they ask for something a politician can actually deliver, like stopping a dam or closing a stretch of ocean to fishing. They also win when the solution is ready before the window opens. The World Heritage listing gave Canberra the legal power to act, so a new government could move on day one.
Big numbers aren't enough if they don't turn into votes. In February 2003, hundreds of thousands of Australians marched against the coming war in Iraq, and it went ahead anyway. The next year's election was fought on interest rates, and John Howard was re-elected with a bigger majority.
How a policy survives
Winning once isn't enough, because the next government can undo it. What keeps a policy alive is the people it pays.
Australia tested this. The carbon price and the Renewable Energy Target came out of the same climate movement. Labor expanded the target after winning the 2007 election, in which climate change was a big issue, and a Labor government brought in the carbon price in 2012.
The carbon price paid no one. Its cost showed up on every power bill, while its benefit, a safer climate, was spread across the world and decades away. Coal and mining groups spent about $10 million on advertising against it, and in 2013 the Coalition won government on a promise to "axe the tax". It was repealed the next year.
The target was different. In The Problems Nobody Pays to Fix I argued that a government's most powerful move is to give a problem a customer, and the target did exactly that. It made retailers buy renewable power and helped pay for rooftop solar, and by 2013 more than a million Australian homes had panels on the roof, with an industry of installers and developers grown up around them.
The same Coalition government's own review wanted the target closed to new projects, but by then too many people were benefiting from it. The government left household solar alone and managed only to cut the large-scale target by about a fifth.
The pattern holds well beyond Australia. A political scientist who studied American reforms, from tax to airline deregulation, found that the ones that last give people a stake in keeping them.
One movement had won both policies, but only the target paid anyone and they fought to keep it. A movement wins a policy by making it a way to win seats, but the people the policy pays keep it alive after the marchers go home.