Aurelia hits the nail right on the head - Japan has superimposed the institutions of a Westminster-style parliamentary system on a Prussian-inspired bureaucratic system. And, as Ehud points out, deeply entrenched informal bureaucratic practice were one of the factors that undermined the DPJ's efforts to implement political control of the ministries. Indeed, informal practices often trump formal enactments in bureaucrat-politician power relations (e.g., Tanaka Makiko vs. Nogami Yoshimi/Suzuki Muneo).

Even though today's central state bureaucracy is a shadow of its former self (although it remains a formidable force in executive affairs), a great deal of institutional inertia must be overcome if Abe and his cabinet are going to assume the driver's seat. Younger bureaucrats may have a different mindset, but one still detects an "Emperor's servant" (tenno no kanri) attitude - i.e., servant of the state as opposed to an agent of the politically-appointed minister - among senior bureaucrats. Also, despite a good deal of mid-career "loaning out" (shukko), bureaucrats still regard themselves as officials of a particular ministry. As long as this continues, bureaucratic "sectionalism" and inter-ministerial rivalries will present obstacles to the realization of centralized political control (e.g., cabinet government).
Historically, major institutional change in Japanese executive affairs has followed "critical junctures"
(e.g., Meiji Restoration or MacArthur's reforms).
Unless the triple disasters of 3.11 can be construed to represent a similarly dramatic tipping point (which, to date, does not appear to be the case), it is difficult to imagine what will drive the sweeping power shift
alluded to in the Reynolds article.

Best,

Brian