Thursday, May 12, 2011

The geographic effect

May 11, 2011
GE 2011
By Ivan Png

THE General Election of May 7 was a watershed in Singapore politics. The Workers' Party (WP) overcame the dominance of the People's Action Party (PAP) to win a Group Representation Constituency (GRC).

Supporters of the Workers' Party have hailed its success as vindication of its call for a First World Parliament. Various commentators have speculated that the dynamics of Singapore politics has changed. The new model is one of competition between two mainstream parties - PAP versus WP.

However, looking at the electoral map in The Sunday Times on May 8, my son and I were struck by a geographical pattern. Anti-PAP sentiment seemed to be strongest in the east, and then tapered off towards the west. The exceptions to the geographical pattern were the PAP's relatively strong showing in Ang Mo Kio and Pasir Ris-Punggol GRCs - helmed by Prime Minister Lee Hsien Loong and Deputy Prime Minister Teo Chee Hean respectively.

So what drove opposition support last Saturday - the WP or geography? This is an ideal setting for the statistical technique of multiple regression. This method is now widely taught - undergraduates routinely use multiple regression in their projects and even secondary school students are learning it.

Using data from the General Election, I regressed the percentage support for the opposition along three groups of factors that possibly influenced the election. They were: constituency size in terms of both the number of candidates (ranging from one to six) and number of electors; the identity of the opposition party; and geography.

Overall, the most significant factor explaining opposition support was the identity of the opposition party - the Worker's Party being worth between 11 and 13 percentage points more than the Singapore Democratic Alliance (SDA). But the average WP support was 46.6 per cent compared with the SDA's average of 30.1 per cent. So the identity of the party was only part of the story.

The second most significant factor explaining opposition support was geography. Constituencies in the east - which I defined as being Nee Soon, Ang Mo Kio, Bishan-Toa Payoh, Moulmein-Kallang, and Tanjong Pagar and further east - gave the opposition between 4.5 and 5.8 percentage points more support than those in the west.

The effect of geography was even stronger when I excluded Ang Mo Kio (the Prime Minister effect) from the analysis. Excluding Ang Mo Kio, the impact of being in the east on opposition support was 7.6 percentage points.

So did the Workers' Party do especially well by concentrating its attention in the east? Among the GRCs that the WP contested, their weakest results were in Nee Soon and Moulmein-Kallang - right on the boundary between east and west.

A tale of two Singapores - but of two parties or two geographies? And to the extent that geography is a major factor, the parties must be asking why. The reasons could be differences in income, housing, new immigrants. With the appropriate data, the multiple regression analysis could be expanded to include all of these and we (as well as the political parties) would know more and be better prepared for the next election.

The writer is the Lim Kim San Professor at the NUS Business School, and Professor of Economics and Information Systems at the National University of Singapore. He was a Nominated MP in 2005-06.

No comments: