The general theme of this research is to introduce the notion of large-scale economic networks into the mainstream of economics, in particular in macroeconomics and international trade. Economic agents often do not have access to all the relevant information they may need: whom they know, whom they interact with represents a small fraction of all possible interactions. I model this limited set of interactions as a network: agents are nodes, and they only interact with other agents they have formed a link with. What is the shape of this network of linkages between agents, and how does it evolve? More importantly, what are the aggregate implications of the shape of this network? These are the broad questions I will address in this research. I will consider six specific applications of this unifying idea in various fields: international trade, IO, macroeconomics and growth. In international trade, we have only a very crude understanding of the frictions that prevent most firms from exporting. I propose to model trade frictions as a dynamic network: at a point in time, a given exporter only has information about a limited set of potential customers in a few foreign countries; over time, this exporter discovers new export opportunities, and its network of customers evolves dynamically. I offer theoretical and empirical tools to understand and analyze the properties of this network, and show how it shapes aggregate trade patterns. In IO and macroeconomics, most plants only have few suppliers. I will model the input-output linkages between plants as a dynamic network; I offer theoretical and empirical tools to analyze this network, and show how it shapes the propagation of plant level shocks to generate aggregate fluctuations. Human capital accumulation is key to economic growth and development, with workers learning from each other. I will model these growth-enhancing interactions as a dynamic network; I will show how the properties of this network shape long run growth.