Will Indonesia’s presidential election be stolen?

Indonesia presidential candidate Joko “Jokowi” Widodo gestures as chats with members of the media in Jakarta on July 10. The elections commission is expected to announce the official result around July 22. (Darren Whiteside/Reuters)
As part of our continuing series of Monkey Cage election reports, the following is a post-election report on the recent Indonesian presidential elections from Cornell University political scientist Tom Pepinsky.
*****
On July 9, Indonesians went to the polls to elect their next president. The official results won’t be known until early next week, as Indonesia’s Electoral Commission works to count over 150 million votes from across this sprawling archipelago. Early results from “Quick Counts” have Joko Widodo and running-mate Jusuf Kalla with a 4 to 5 point lead over Prabowo Subianto and Hatta Rajasa. But both camps are claiming victory.
The two presidential contenders represent two starkly different options for Indonesia. Joko Widodo — universally known as “Jokowi” among Indonesians — is billed as a populist reformer, who hails from an ordinary middle-class background and rose to national prominence as the extraordinarily popular mayor of Solo, a regional city in central Java. His reputation is for low-key, honest, pragmatic leadership, and he promises to confront the corruption that plagues Indonesia and reform the country’s institutions. Prabowo, by contrast, is the quintessential insider, the ex-son-in-law of deceased former dictator Suharto, a former general and amember of a prominent family. His message is one of decisive leadership that will address Indonesia’s lackluster economic performance, while restoring a sense of order and greatness to Indonesia.
Prabowo’s personal history makes him something of an unlikely candidate. He is widely known to have a short temper, and has been linked to both human rights violations during the Indonesian occupation of East Timor and to the disappearance of pro-democracy activists during his time in the military. Taking advantage of a poorly run campaign by Jokowi, however, Prabowo cobbled together a broad coalition of political parties and wealthy business backers to mount a quite competitive race.
By the eve of the election, Jokowi held only a slight lead. Public opinion surveys conducted by the well-regarded Saiful Mujani polling company during the first week of July showed 44.9 percent of likely voters choosing Prabowo-Hatta, 47.6 percent choosing Jokowi-JK, and 7.5 percent undecided.
The demographic breakdown of Mujani’s results provides a window into the sources of popular support for each candidate.

Jokowi’s supporters are on the whole more likely to be rural and less educated than Prabowo’s supporters, and also were more likely to be employed in lower skill occupations (farmers, fishermen, workers) or to be women working at home. This probably reflects urban middle class frustrations with Indonesia’s rampant corruption, and the belief that only Prabowo’s decisive leadership can bring it under control. Muslims are roughly divided between Jokowi and Prabowo, with a slight edge to Prabowo, but non-Muslims massively favor Jokowi, by a roughly 4-to-1 margin.
In addition to conducting pre-election surveys, Mujani’s firm and a number of others conduct “Quick Counts,” and these are the key piece of quantitative evidence about the results of the July 9 election given that the official results have not been released. Unlike exit polls, Quick Counts use the actual results from entire polling stations. They rely on well-understood statistical principles to produce highly accurate estimates of the final vote share by combining a large number of polling stations that represent (in a statistical sense) the Indonesian voting population in its entirety. I plot the results from eight firms below alongside Mujani’s most recent pre-election survey. (I also show the confidence intervals around that survey’s estimates of the vote share going to each candidate to reflect our uncertainty.)

All eight Quick Count results above fall within the 95 percent confidence intervals of Mujani’s pre-election survey. Each shows Jokowi with a substantial lead; six of the eight have his margin at greater than 4 percent. It is almost certain that if Indonesia’s electoral commission faithfully reported the results of the July 9 election, Jokowi would win handily.
Once elections are over, vote buying is no longer an option. So given these numbers, how could Prabowo steal the election? Two political scientists at the Australian National University in Canberra, Edward Aspinall and Marcus Mietzner, have outlined one plausible scenario. The basic strategy is to combine deliberate misinformation from Quick Count results by “non-independent” firms with “manipulation of the formal counting and vote tabulation process” — that is, vote stealing. Four firms (not shown above) also released Quick Count results, but these showed narrow victories for Prabowo. Notably, as Aspinall and Mietzner point out, none of these firms have a track record of conducting accurate polling, and all have links to Prabowo. The Prabowo camp is accordingly using these four questionable Quick Count results to create sense that the electoral results are too close to call.
A great deal depends on the ability of thousands of officials from the electoral commission to resist strong pressure, both financial and otherwise, to shift the results in Prabowo’s favor. These two weeks accordingly represent a critical period for Indonesian democracy. Both candidates claim to have received a popular mandate, but Prabowo — the presumptive loser — is almost certainly willing to press his case using non-democratic tactics. The weakness of Indonesia’s rule of law is well known, but Indonesia has seen nothing approaching this level of pressure on election officials since democratization in 1999.
One facet of Indonesia’s vote counting procedures may prove important in raising the costs of vote stealing. The Indonesian electoral commission has committed to placing scanned images of the official reports from every single polling station around the country online. Nearly all have been submitted already (see here for progress, and you can click here to see an example of one report from central Jakarta). Enterprising Indonesians have created Web sites that allow visitors to process these images into numerical data on vote totals — in effect, crowdsourcing election monitoring. There is even a Tumblr where observers can flag problematic reports. As of the morning of July 18, the results of these efforts have Jokowi and Jusuf Kalla with 52.8 percent of the vote, broadly in line with the Quick Count results.
Whether or not such crowdsourced public audits will affect the electoral commission’s official outcome remains to be seen. What is clear even now, though, is that Indonesia’s most competitive election in recent history is not yet over.