I eventually decided that we needed a theory of leadership that could
account for the fact that, for every success story in public administration,
there seems to be an equal and opposite failure story, and that different
evaluations of the same approach (be it public housing or school vouchers or
whatever) are irreconcilable. I therefore came to the conclusion that
governance (the capacity "to guide or steer" in its original Latin meaning)
should be analyzed on the basis of two dimensions: "political hardware"
(referring to rules, procedures, technology, organizational arrangements,
methods, etc.) and "political software" (referring to the quality of relationships
between leaders and followers essential for the effectiveness of political
hardware).