Why does any species live as long as it does? The high-level answer is that present length of life is an evolved consequence of adaptations that allow a species to succeed in occupying its niche, via competitive success for individuals in propagating their genes. You don't see the losers in this process, as they have vanished. There is fierce debate over the nature of the relationship between desirable adaptions that provide evolutionary success and consequent length of life, and the debate is very different for different species and different circumstances. In general, however, researchers who hold that aging is a process caused by accumulated cellular and molecular damage view aging as a sort of shadow cast by natural selection operating on youthful individuals. There is great selection pressure upon young biology, the fight for success in early life, and a species can boast successful adaptations that enhance reproductive success but which nonetheless exist at the expense of individual health and survival in later life.
Some models suggest that this shadow of aging is in and of itself necessary for the long-term survival and success of species, as aging species tend to outcompete ageless species during periods in which the environment changes. Since there have been innumerable such episodes in our evolutionary past, it is perhaps not too surprising that aging species are far more numerous than those few species that might be ageless.
A process that is advantageous in youth but becomes harmful in aging is known as a form of antagonistic pleiotropy: the human immune system is perhaps a good example of a system that has this property. It is structured so as to be effective in youth, but some of the very same aspects that make it so effective at the outset of life - such as the inflammatory immune response and the ability to remember specific threats indefinitely - contribute to immune system failure in later life.
In humans a big question lingers over our comparative longevity. Why did we evolve to live for so much longer than our primate cousins? One proposal is the grandmother hypothesis, which essential points to a combination of culture and intelligence as the root of human longevity. Once a post-reproductive individual can materially contribute to the success of his or her descendants, then a selection effect for greater longevity comes into play, one that doesn't exist in primate species that are not intelligent enough to have this ability.
To pick another species in which the situation is radically different, we could look at salmon. Salmon undergo a very sudden aging process after spawning, and the details of this process are strongly driven by the habits of bears who feed on salmon. Some bear populations prefer to feed on older salmon, and in those rivers salmon age more slowly.
The influence of predation on the evolution of aging is an established body of theory in the evolutionary science community. It is generally accepted that greater predation will tend to favor the evolution of shorter life spans, even if only because it allows for adaptations that ensure early reproductive success but which place a great physiological burden on later life. Mechanisms that instead ensure a longer reproductive life span will not be selected if individuals are largely eaten young.
All species are different, however, and there is always room for argument in this field. Which is why you will see such things as researchers crunching the numbers in support of the dominant viewpoint on predation and longevity:
In birds, variation in life-span extends from parrots such as the Sulphur-crested cockatoo that can become more than 100 years old, to the small Allen's hummingbird with a maximum life-span of only 4 years, a 25 fold difference. How can this variation be explained?
The classical evolutionary theory of ageing, first proposed by the famous evolutionary biologist George C. Williams over 50 years ago, gives an answer. The theory predicts that high mortality rates in adult animals due to predation, exposure to parasites and other randomly occurring events will be associated with shorter maximum life-spans. This is because under high external mortality most individuals will already be dead (eaten or succumbed to disease) before natural selection can act on rare mutations that cause healthier ageing. The theory has since been further developed and tested in a number of experimental and comparative studies. Yet contradictory results have caused scientists to cast doubt on its validity.
[Researchers] have now tested this theory using a comprehensive database on estimates of maximum life-span of 1396 bird species, 1128 from free-living species and 268 from birds kept in captivity. The researchers used a global distribution map of these species, included data on their morphology and reproductive rate, and estimated predation rate.
By means of complex statistical analysis methods they found that in the investigated bird species maximum longevity is negatively related to the number of predator species occurring within the same geographical area. This means that the more predator species are present in the same habitat and the more evenly they are distributed, the lower is the life span of the respective species. This relationship supports the classical theory of ageing, and remains valid when other life history traits known to influence longevity such as body mass and clutch size are included into the statistical model. Indeed, larger species live longer, and those that reproduce fast (lay more eggs) live shorter lives. Remarkably, the observed pattern showing longer life-spans when fewer predators are present emerges no matter how the analysis was done: at the species level, at a finer regional scale (groups of species within a certain area) or even when comparing entire bioregions.