There are several reasons why.
a) To explain is to give a reason, which is to give a cause. Hence, any explainable phenomenon is causal. This does not tell us that all phenomena are causal unless all phenomena are explainable which is, of course, not obviously the case. (In fact, I believe it is not the case that all phenomena are explainable by us, but more on that later). But we can at least say that acausality may never feature in any coherent explanation. Thus, to explicitly invoke acausality in an explanation is to engage in contradiction. This makes almost any video featuring Michio Kaku virtually unwatchable.
b) Causality implies conservation (if the world is causal, you cannot have entities and forces popping into/out-of existence). Conservation of all physical variables (except entropy) is perhaps the most well-established fact of physics. While this does not prove the antecedent, it is strongly consistent with it.
c) The most subtle aspect of the issue has to do with the complexity of explanations. We can rank explanations in order of their complexity and when explanations are so ranked, it turns out that there is a concrete reason why Ockham’s razor is the case - simpler mechanisms are more probable than complex mechanisms, ceteris paribus. In other words, if you have a set of phenomena A, and two or more explanations (mechanisms) that are consistent with A, the simplest mechanism is to be preferred because if we imagined mechanisms being assembled at random, the simplest mechanism is more likely to occur. I’m stating this in very imprecise language for the sake of communicating the essence of the argument. For more information, search “kolmogorov complexity” “chaitin” “solomonoff induction”, etc.
When applied to physical systems, this idea suggests that random phenomena (acausal phenomena) are the most improbable of all. This reinforces point a) above and gives it a more specific meaning… there’s a reason why we should treat random phenomena as incoherent… the implied “explanation” is maximally complex.
Finally, on the point of our inability to explain everything, this goes into the fact that a computation system of complexity X cannot fully explain phenomena of complexity greater than X. Again, I’m using extremely imprecise language (caveat). The brain is not infinitely complex and, thus, it can only explain phenomena of a certain complexity. There is no reason to suppose that the Universe itself (its fundamental operation) is less complex than the human brain. Hence, there is no reason to expect that we can explain all real phenomena. This has some implications to the nature of noise/randomness in physical theories.
Clayton -