The political ferment triggered by Ariel Sharon's disengagement policy and by his formation of a new party has sparked renewed discussion of another sweeping reform: changing Israel's electoral system. One prominent proposal is to reintroduce a system of direct election of the prime minister, this time in a more presidential version. A fresh look at the precursor of this idea - Israel directly elected its prime minister from 1996 to 2001 based on legislation introduced in 1992 - offers a cautionary tale about grandiose projects meant to fix all the system's shortcomings. The idea of direct elections had many fathers in Israel's 1980s political and academic landscape. It eventually landed in the Knesset, tabled by four MKs from different parties a day after the collapse of the National Unity Government led by Yitzhak Shamir as a result of what came to be remembered as "the dirty trick." It may now be embraced by the Kadima party if it wins the March elections. There is nothing intrinsically wrong in the idea of directly electing a prime minister or a president; the question is how to craft the reform to give the elected leader not just a popular mandate but also a workable parliament that will endorse his policies. It is not enough, in other words, for a leader to be chosen by the people. The 1992 reform's failure to do this was no small oversight, given that the direct election proposal was ushered in to address the previous system's failure to produce workable majorities, strong leadership and a clear sense of direction in the face of existential political challenges. But the problem of direct election was not in its effort to take the process of leadership selection away from the parties to restore it to the people. It was in the na ve assumption that a directly elected prime minister could, on the strength of a popular mandate alone, be able to govern more effectively. The lack of government backing exposed the legislation to last-minute amendments that emasculated its aims. Still, even had the law passed as originally envisaged, it is doubtful that things would have been so different. History played its tricks and even a carefully crafted and extensively debated reform ended up producing results and incurring challenges that neither its opponents nor its promoters had anticipated. TAKE TWO main arguments voiced against directly electing the prime minister. Right- wing tribunes warned that the Arab vote would automatically go to the Labor candidate, thus ensuring that the Left would return to power and bring about the partition of the Land of Israel and the creation of a PLO state. Israel indeed negotiated with the PLO and began partitioning the Land of Israel. But while the former was a result of a clear victory of the Left under the old system, partition's most tangible step - evacuating settlements - happened under Ariel Sharon, not under Labor. Liberal voices warned that a strongman would seize power and destroy what remained of Israeli democracy. In the heated exchanges that accompanied the legislation, one man's possible rise to power was evoked to warn against the reform: Ariel Sharon. History ridiculed these arguments: the Arab vote might have been influential, but that did not prevent two Likud prime ministers - out of three directly elected before the reform was scrapped in 2001 - from winning. The direct election system did not prevent Barak from soundly defeating Netanyahu in 1999. As for the strongman, Sharon eventually made it to the prime minister's seat, but by the time he won his second term in 2003 he was a popular figure and a centrist leader, not the Napoleon many had warned against. Sharon, by then, had become the panacea against the effects of direct election, a man many wanted not because he endangered democracy - but because his predecessors' inexperience and ineffective leadership needed to be rectified. Besides, no directly elected prime minister could aspire to be a strongman thanks to a popular mandate: they all had to contend with a deeply fragmented Knesset - something that direct election supporters had not factored into their forecasts. The ragtag coalitions that Netanyahu, Barak and Sharon had to put together were one of direct election's unintended side effects: though forecast by a few clairvoyant Cassandras, the breakdown of Israel's party system made the elected prime minister look like a general without an army. This breakdown was the result of the least debated and ill-conceived aspect of the reform. While the prime minister was directly elected he still had to contend with the Knesset's confidence to install his government. This mechanism - which reformers had originally removed but the Knesset reintroduced as an afterthought - led to disastrous consequences. Direct election was meant to weaken parties' influence in the process of government formation. Thrown out of the window, this influence came back from the backdoor through the pernicious mechanism of the double ballot - one for prime minister, one for the Knesset - which itself became a central element of the reform without much debate or thought about its consequences. What was feared would strengthen the prime minister too much achieved nearly the opposite of what it was expected. By contrast, it was under the old system, shunned for its inability to foster strong leadership and effective governance, that the late Yitzhak Rabin led a homogeneous coalition to sign the Oslo accords - a clear cut political choice and an example of strong leadership. And it was after 2003, under an improved version of the old system, that Sharon led the Likud out of the political wilderness inherited from Netanyahu, cemented Israel's alliance with the US in unprecedented ways, defeated the Intifada, built the fence and proceeded to disengage from Gaza. Israel's next elections will be no doubt dominated by events that nobody could anticipate and might yield yet again an uncertain political landscape, where the next prime minister will have to build a broad coalition with narrow space for political maneuver on the crucial issues of the day. Ultimately, the push to reform the system to improve its performance at the decision making level is both commendable and desirable. But the recent history of failed reform teaches a humbling lesson to promoters of grandiose and carefully crafted changes. Even direct election, which was debated during eight Knesset debates, 44 committee meetings, in countless public and party forums and through thousands of articles in the Israeli and international press, ended up yielding outcomes that few anticipated and never produced the awesome consequences so many hoped for and others feared. Next time reform is considered seriously it is perhaps inevitable that some results will, again, not be fully anticipated. Yet if we learn from past failures we may be able to keep such unpleasant surprises to a minimum. The writer teaches Israel Studies at Oxford University. His book Israel's Electoral Reform will appear later this year.