The Importance of Sharing in Knowledge Growth

Main Menu           Previous Topic                                                           Next Topic

Sharing Rate Strongly Influences Knowledge Endowment over Time

Note: in order to run the simulation referred to in this slide, go here, to the Java Applet version. You will be directed to download the latest version of the Java plug-in.

Let's start with a new society of 11 agents with 2 channels each, which is shown to the left. What we want to show is how much better off a network can be by improving its sharing rate.

As with the learning functions, we set everyone to be equal in terms of their baseline rates. Here agents only share about 10% of their private information per time step. What we now wish to do is to increase the sharing rate of the third tier network which has average resources much lower than those of the top tier network. The following shows what happens in more open environments when agents are willing to share more of their private information.

First we create a sharing function which causes the agents to share 50% of their private resources rather than just 10%. Then we assign it to the members of the third-tier network and run the simulation forward by 25 steps. Press "Go" to run the simulation with the changed functions.

Below, you can see a graph of what happens to the agent knowledge endowments. The third tier network has outstripped the other two by a substantial margin over the 25 periods. Increasing the sharing rate has substantially increased the growth in overall resources and provides an indication of the advantages that are possible in more open environments or ones with groupware and other sharing systems in place.

A situation of growing resources suggested by this observation derives some empirical support from a case study of high technology firms in Massachusetts and in California [Saxenian 1995]. In 1975, firms in the two regions employed roughly the same number of people but 15 years later the firms in California had created three times as many net new technology jobs. Additionally, in the last several years of the study, firms in Massachusetts had increased their market capitalization by roughly 1 billion dollars as compared to 25 billion in California. The study suggests that (1) firms in New England were more vertically integrated so their business units traded less information with buyers and suppliers (2) these firms also had more defense contracts so they couldn't share their research and (3) scientists and engineers did more job hopping and traded more information in California leading to a greater distribution of knowledge in California.

                   Previous Slide                                                           Next Slide