Part of the problem is that so many different things were happening in the Late Permian—massive volcanism, changes in ocean and atmospheric chemistry, global warming, and most animal life dying—that it is often difficult to separate possible causes of the extinction from consequences. We have some evidence for an extraterrestrial impact (but evidence that is still, in my view, less than overwhelming); the evident simultaneous dates of the so-called Siberian flood basalts [the geological remnants of lava flows] and the mass extinction; strong evidence for some degree of low oxygen and other changes in ocean chemistry; and a sudden spike in global temperatures.
Since I am suspicious of coincidences, right now the eruption of the Siberian flood basalts seems a more plausible cause of the extinction than an extraterrestrial impact. Although quite a few scientists are sure that an impact caused this extinction, the fact remains that repeatability is a critical aspect of scientific research, and at this point, none of the various claims for impact debris have been independently confirmed by other scientists. This does not mean they are wrong, but it does suggest that impact is not the best bet.
Some pieces of the puzzle still don't quite fit.
While the simultaneity of eruption and extinction is impressive, the problem is how one caused the other. Impact enthusiasts claim that the simplest explanation is that an impact triggered the Siberian flood basalts. That would certainly be an interesting result, and may be the only way the Permian will ever succeed in Hollywood, but computer simulations suggest that this would require a meteor about 125 miles in diameter, versus the six-to-nine-mile-wide meteor thought to have caused the end of the dinosaurs. If something this big had hit the Earth, we would not have to search for the evidence, as it would be quite apparent.
We have so little understanding of the climatic effects of massive eruptions like the Siberian flood basalts that the repercussions may have been far greater than we imagine. I suspect that the progressive effects of acid rain produced by sulfuric aerosols, massive releases of carbon dioxide from volcanic eruptions, and the burning of vast coal beds in the volcanic region caused global cooling then warming, and it is within this nexus that I believe the cause of the extinction lies.
But some pieces of the puzzle still don't quite fit. Did terrestrial life really die out at the same time as marine life? Evidence from South Africa and new work from China suggest that they did, but studies in Greenland indicate the marine extinction may have occurred before the one on land. What really happened in the seas, and how was this related to extinctions on land? The complexity of the geochemical changes in the ocean make me worry that we have missed something about the forces that triggered the extinction.
Our difficulties in understanding this event may lie more within how we tend to define a satisfactory answer than in the event itself. In the wake of the apparent success of the extraterrestrial impact hypothesis for the end of the Cretaceous, many of us seem to prefer a single dramatic cause as an explanation. History is messy, and just as there is no simple answer to what caused the First World War (despite what we were taught in high school), I see little reasona priori to expect such a neat and tidy resolution to this riddle.
Read more about the evidence and controversies in the quest to solve the Permian riddle in Doug Erwin's book Extinction: How Life on Earth Nearly Ended 250 Million Years Ago (Princeton University Press, 2006).