In recent years many applications of Physics methods have been applied to understand financial systems. Main focusing is on financial time series of stock market. Here we focus on another type of market: the interbank money market, i.e. the market of liquidity management in the banking system. This kind of market can be represented as a network, where banks are nodes and transactions among banks are links (1). In particular we focus on the Italian segment of this market. Banks involved are highly heterogeneous by their capital sizes and their kinds of transactions. We can study the topology of the network and use methods of graph theory to understand the auto-organization of this set of banks. In fact the money market is totally screen based in Italy and in principle each bank can interact with any other bank. From our studies emerges that the system self-organises in such a way that reveals a division of roles among different banks (2,3). Moreover a changing in the topology of the network can be observed during the time, revealing typical time scales and a dynamical evolution of the system (1,2). 1) "A network analysis of the Italian overnight money market",

1) G. Iori, G. De Masi, G. Caldarelli et al., accepted in "Journal of Economic Dynamic and Control"

2) "Trading strategies in the Italian inter-bank money market", R. Reno', G. Iori, G. De Masi, G. Caldarelli, submitted

3)''Emerging communities in the interbank complex network'', G. De Masi, G. Iori, G.Caldarelli, submitted