That rule, however, may be changing. A study in the 'Economic and Political Weekly' this month shows that in the last five years, for the first time, the percentage of governments losing elections has slipped below the 50% mark — to just 46%.

The story of anti-incumbency being a powerful force is not factually off the mark — the attrition rate for incumbent governments was inordinately high in the post-Congress polity after the collapse of the one-party dominance system.

A pattern was established through the 1990s: after the election, the short honeymoon, followed by the seemingly inevitable disillusion, anger and the eventual defeat of the incumbent in the next election.

But some recent electoral verdicts indicate that the story of state politics in India may be turning again. There have been signs of a shift — from change to continuity.

Several state governments have won another mandate in the last five years. Incumbents retained power in Madhya Pradesh, Chhattisgarh, and Delhi in the latest round of state elections in November-December 2008. Before that, incumbents won not just in West Bengal, but also in Maharashtra, Assam, Orissa, and most famously in Gujarat.

In the EPW article, political scientists Yogendra Yadav and Suhas Palshikar assess the performance of ruling parties in Assembly elections from 1989 to 2008 and describe a startling new pattern. There has been an eye-catching decline in governmental anti-incumbency at the state level: If 77 per cent of the governments lost elections between 1989 and 1998, the rate fell sharply to 62 per cent between 1999 and 2003. And in the last five years, it has come down to 46 per cent.

In 1999-2003, 10 out of 29 state-level incumbents won another mandate to rule. The number increased to 13 out of 28 in 2004-2008. The reduction in the swing against the ruling party over the same period looks only deceptively small: from -3.03 per cent to -1.73 per cent.

What could be the reason? Or reasons? Can there be a uniform explanation like the emergence of the era of “better governance” that transcends other factors that vary across states — like differences in the format of party competition, the make-up of winning social coalitions, popularity of state-level leaderships, or failure of the opposition to seize the moment?

Just as anti-incumbency was not an air-tight formula — for instance, there was “irregular oscillation” in states like Madhya Pradesh, Andhra Pradesh, Assam, Maharashtra, and Delhi where the ruling party sometimes bucked the trend, and “rotation” of power among more than two parties in states like UP, Karnataka and Haryana — the new tilt towards continuity cannot be viewed as an evenly infectious or undifferentiated trend.

Yadav and Palshikar point to differences in the victories of governments even in states where incumbents returned to power in 2004-2009. Incumbents in states like West Bengal 2006, Nagaland 2008, and Chhattisgarh 2008, won with handsome margins and consolidated their base in the state. In others states like Madhya Pradesh 2008, Gujarat 2007 and Orissa 2004, incumbents won only a “qualified extension” with a reduced majority. And in Maharashtra 2004 and Delhi 2008, the ruling party or coalition actually lost votes heavily while retaining power.