His broad point is not to look at central bank history from an institutional perspective. In this way we look at certain central banks like Fed, BoE etc and look at how they have evolved over the years. This way we miss some other organisations which worked as central banks in the past but do not exist today. The research focus instead should be from a functional perspective where we list the functions of central banks and see which entities did them in the past.
He begins saying we ask wrong qs from economic history:
A fundamental precondition to this exercise is, of course, not making an anachronistic use of the past. Unfortunately, this is what a good deal of economic historians have often being doing, just asking historical sources: “When did we learn to do it right?” Based on the assumption that today we are better than any of our ancestors at understanding the world, this teleological approach is what drove us directly into the crisis. The events of recent years suggest that a more neutral stance should be adopted. The kind of question we should rather ask sounds like: “What did we do, and did it work?”
In a recent survey of the long-term evolution of central banking (Ugolini 2011), I ask this very question. To this aim, the paper takes a functional approach (Merton 1995). It does not look at what today’s central banks used to do in the past. Rather, it looks at what kinds of organisations used to perform, in the past, the very functions central banks perform today. This approach allows us to broaden considerably the scope of analysis, as it puts into comparative perspective a number of institutional arrangements – going back to as far as the Middle Ages.
He introduces microfoundations of central banking:
My survey starts from the microfoundations of central banking. In any sufficiently advanced financial system, the need for the centralisation of interbank payments naturally arises. Proponents of free banking maintain that such demand can be adequately met by a private clearinghouse. A number of objections have been advanced against this idea. In particular, it has been pointed out that there exist scope economies in the clearing process, which lead to its concentration to a single intermediary (Goodhart 1988). Therefore, the payments system can be described as a natural monopoly. Because of their great liquidity, claims on the organisation sitting at the centre of the payments system tend to acquire the status of money – even when they are not granted legal-tender status. Thus, the intermediary dominating the clearing process becomes, de facto, a money-issuing organisation.
He looks at this whole debate of central banks buying bonds and says hostory shows central banks have often retorted to these measures. This does not lead to eroding independence of central banks:
Should we conclude that a “subjugation” of monetary policy to fiscal policy must be expected in the near future? Not necessarily. What emerges from the past is that while the relationship between central banking and politics may have been dynamic, it has always been very close. It also comes out that deficit monetisation has very often occurred almost everywhere. Sometimes it ended in catastrophe, but on many other occasions, it did not. This suggests that the debate we should be confronted with today is not whether monetisation is admissible or not, but whether we can appropriately assess its long-term costs and benefits. In some circumstances, it may not be optimal for monetary authorities to guarantee government debts; in others, it may well be (Kocherlakota 2011).
Similarly, we should not take an ideological approach to the question of central bank independence. Central banking is the outcome of collective bargaining. What history shows is that central banks did not derive their actual strength from formal independence, but from the credibility of the institutional arrangement in force. In order to prevent the equilibrium from being fragile, monetary and fiscal authorities must not be perceived as free-riding the one on the other. This would patently be inconsistent – after all, both are but the two sides of the same coin.