Responses to two interesting messages from Prof. Hiwatari:

1. On electoral systems, my picture is that there is a logic to a given electoral system (or other institutions of that sort) that means that people who don't behave according to that logic will not do as well as people who do. As Reed argues, it is a process of eliminating losers (like evolution) plus some learning by mistakes, more than people making conscious calculations about the future. Other forces certainly affect behavior as well, including intertia, the tendency of people to do what they have always done.
The implication of that is that when there is a lot of change, things are getting shaken up for some reason (such as urbanization), the processes of elimination and of learning happen quicker, and the logic of the instituiton like the electoral system will take hold more rapidly. That point can explain the anomoly that Japan moved to a five-party system in the cities in the 1960s, but retained a two-party system in rural areas despite the electoral system.
Two footnotes to that: (1) it helps that many rural areas had three-seat constituencies, which presumably would have less of an effect than bigger ones;
and (2) having said all that, I think it is still an interesting puzzle that the LDP continued to dominate rural areas so much despite the substantial amount of social change that went on there.
As to factions, maybe Steve Reed will say something.

2. As to bureaucrats vs. politicians, often I think I can't stand that question anymore, but yet I keep coming back to it again and again. I think the usual question is wrongly posed because we don't pay enough attention to what we mean by power. The P-A line of thought often comes down to a matter of ultimate or constitutional or veto power. On those grounds it surely lies with the politicians in the postwar system, but for most of the things I am interested in, mostly about how public policy changes, I don't think that notion of power helps us answer many questions. (As NH observes, it is the problem of slack that does us in).
I prefer to look at who is pushing issues on to the agenda that lead to policy change. That can be bureaucrats (or one might better say agencies), politicians, interest groups, public opinion, environmental changes, gaiatsu, etc, with a lot of variation by policy area, era, and just randomness. It is nonetheless quite valid to look at the balance between bureaucrats and politicians in that regard, and in fact it is not all that hard to do empirically through historical case studies of policy change.
For what it is worth, on those grounds I see the politicians becoming much more active in the 1960s, many important bureaucratic agencies as being pretty ineffectual in the 1970s, and an upsurge of the bureaucrats in the 1980s.
As to the comparative question, I am really not too sure whether Japan is especially distinctive here, but I am certainly convinced that the US is unique in the very weak role played by its bureaucratic agencies in most policy areas (with the usual exceptions of defense, monetary policy, etc, and for a period up to the 1970s probably social policy).
I could go on and on . . .