On the eve of the 2012 presidential election, the Central Election Commission announced that election forecasts made by the Exchange of Future Events (Xfuture.org), an electronic exchange operated by National Chengchi University’s Center for Prediction Markets (CPM) and xPredict, fall within the scope of the Presidential and Vice Presidential Election and Recall Act (總統副總統選舉罷免法). As such, it instructed Xfuture to stop the forecasts the following day.
This was a questionable decision, from both an academic and a legal perspective. The commission must recognize the academic value of the predictions, it should not distort or misinterpret the law and neither should it confuse Xfuture election forecasts with opinion polls.
Xfuture is a futures exchange built upon a prediction markets mechanism. It is not an opinion poll: It is a predictive method used in the social sciences. The University of Iowa first created the Iowa Electronic Markets (IEM) online futures market 26 years ago, and the concept has attracted the attention of the international academic world ever since. The Brookings Institute and the American Enterprise Institute in the US produce a joint special paper on prediction markets, and the University of Buckingham in the UK publishes the periodical Journal of Prediction Markets.
There are many exchanges around the world operated by major universities conducting prediction markets. In the US, for example, there are the Massachusetts Institute of Technology, the University of Pennsylvania and the University of Virginia, to name a few. There are also privately operated prediction markets, such as the Hollywood Stock Exchange, the Microsoft Prediction Lab and those operated by Google and Hewlett-Packard, among others.
There is already a lot of academic research out there comparing the results obtained from the various prediction markets and those from opinion polls. Three authors from Iowa University, for example, compared their university’s prediction market with a total of 964 polls carried out by Gallup and the Harris Poll, as well as by major media outlets such as the New York Times, and how they fared in predicting the results of US presidential elections between 1988 and 2004. They found that the prediction markets had a 74 percent higher degree of accuracy than opinion polls.
In fact, when IEM ran a prediction of Taiwan’s 2000 presidential election, the first set of prediction market contract prices showed that Chen Shui-bian (陳水扁), who later won the election, had a 55 percent chance of being elected compared with 38 percent for James Soong (宋楚瑜) and 28 percent for former vice president Lien Chan (連戰). The second set placed the respective values at 39 percent, 37.5 percent and 26 percent. In the end, the vote counts were 39.3 percent, 36.8 percent and 23.1 percent, respectively. The second set of predictions was virtually spot-on. That was people in the US predicting the outcome of a Taiwanese election: It was not an opinion poll.
In view of the strong capabilities of prediction markets, in 2008, 22 economists led by three Nobel prize winners published a paper in the renowned periodical Science entitled “The Promise of Prediction Markets.” The paper reiterated the predictive potential of the “wisdom of crowds,” and encouraged the use of this strong predictive capability to enhance the decision making ability of public and private institutions, thereby improving social welfare provision. It also recommended that these should be free of unnecessary government regulations or limitations.
Xfuture was established by the CPM. At present, it is the only prediction market exchange in the Chinese-speaking world. It deals with a broad range of topics, including politics, sports, finance, entertainment, fashion, science, social and international issues, with more than 180,000 people participating from more than 130 countries around the world, providing data to predict future events and trends to help with risk avoidance, including the risks of cross-strait political and economic exchanges.
Returning to the law, the Presidential and Vice Presidential Election and Recall Act is very clear on what constitutes an opinion poll. An opinion poll must have an organizer or presider who oversees the surveying, a sampling method, a sample size, financial sources and an error value before it can be considered a poll. The people participating in prediction markets come from all over the world, so there is no organizer or presiding institution to speak of, nor sample size, and there is certainly no sampling method tying it all together. It is not, therefore, an opinion poll. The commission itself decided in 2004 that estimates of vote counts or vote percentages are not opinion polls.
Academics from overseas and in the CPM have already published a great deal of academic papers on prediction markets, comparing the differences between these and opinion polls. Also, Exchange of Future Events, a book published by Yuan-Liou Publishing, describes and evaluates in great depth how Xfuture works.
We hope that the commission will consider the academic and practical issues here, and get a better understanding of prediction markets. It should not retard progress in the country.
Tung Chen-yuan is a professor at National Chengchi University’s Graduate Institute of Development Studies.
Translated by Paul Cooper
China has not been a top-tier issue for much of the second Trump administration. Instead, Trump has focused considerable energy on Ukraine, Israel, Iran, and defending America’s borders. At home, Trump has been busy passing an overhaul to America’s tax system, deporting unlawful immigrants, and targeting his political enemies. More recently, he has been consumed by the fallout of a political scandal involving his past relationship with a disgraced sex offender. When the administration has focused on China, there has not been a consistent throughline in its approach or its public statements. This lack of overarching narrative likely reflects a combination
Father’s Day, as celebrated around the world, has its roots in the early 20th century US. In 1910, the state of Washington marked the world’s first official Father’s Day. Later, in 1972, then-US president Richard Nixon signed a proclamation establishing the third Sunday of June as a national holiday honoring fathers. Many countries have since followed suit, adopting the same date. In Taiwan, the celebration takes a different form — both in timing and meaning. Taiwan’s Father’s Day falls on Aug. 8, a date chosen not for historical events, but for the beauty of language. In Mandarin, “eight eight” is pronounced
US President Donald Trump’s alleged request that Taiwanese President William Lai (賴清德) not stop in New York while traveling to three of Taiwan’s diplomatic allies, after his administration also rescheduled a visit to Washington by the minister of national defense, sets an unwise precedent and risks locking the US into a trajectory of either direct conflict with the People’s Republic of China (PRC) or capitulation to it over Taiwan. Taiwanese authorities have said that no plans to request a stopover in the US had been submitted to Washington, but Trump shared a direct call with Chinese President Xi Jinping (習近平)
It is difficult to think of an issue that has monopolized political commentary as intensely as the recall movement and the autopsy of the July 26 failures. These commentaries have come from diverse sources within Taiwan and abroad, from local Taiwanese members of the public and academics, foreign academics resident in Taiwan, and overseas Taiwanese working in US universities. There is a lack of consensus that Taiwan’s democracy is either dying in ashes or has become a phoenix rising from the ashes, nurtured into existence by civic groups and rational voters. There are narratives of extreme polarization and an alarming
RSS