Empirically, there is a fairly strong correlation between the term length of British Prime Ministers and their historical rankings as assessed by journalists for The Times: for the people on that page, the correlation coefficients between number of days as PM and ranking are -0.45 for Matthew Parris, -0.60 for Peter Riddell and -0.53 for Ben MacIntyre. (The rankings are such that 1 is best and 53 is worst; hence these numbers show that longer-lasting Prime Ministers were more positively assessed. I don't have p-values because after two hours of inputting data I really didn't feel like doing a whole new load of number crunching, and I can't see any easy way to find the p-value in Excel).
Now there are a couple of obvious ways to go about justifying this. First, if we are to assess leaders by their achievements then a longer tenure gives you more time to rack up achievements. I'm a bit sceptical of this, since (a) it also gives you more time to rack up failures, and (b) it doesn't take all that long to establish an impressive list of achievements. Churchill (unanimously ranked in first place by the Times journalists) is remembered almost entirely for his first term, during which he led Britain to victory in the Second World War; one could ignore his second term of four years without denting his position. David Lloyd George was Prime Minister for less than six years, yet is credited with winning the Great War and founding the welfare state. Peel lasted barely more than five years, but abolished the Corn Laws and established the first official Police. To be sure, longevity helps in this regard, but I suspect that diminishing marginal returns set in pretty quickly after the first five years.
A better argument would be that historians are likely to assess leaders on similar criteria as voters (or other people with the ability to topple a Premier). The better you perform, the more likely you are to continue surviving. Obviously there's considerable variation in this across parliaments - a US president can usually be unseated only every four years, the exceptions of course being Nixon and the four presidents who were assassinated, while Australian PMs can go from being unchallenged to being out of office in a matter of mere hours.
The question, then, becomes one of whether the criteria upon which leaders tend to be assessed are appropriate. This question is one I hope to return to.