Overcoming Challenges in African Beryllium Mining

Beryllium mines and tribulations
Introduction
In the case at hand we seek to help Fine Metals Inc. in, among others, improving efficiency in their two African subsidiaries.
We are first introduced to their subsidiary in Mozambique, which is suffering from an uncooperative local manager and generally low productivity. Second, we face a challenge in the Tanzanian subsidiary, which has a problem with employees stealing beryllium in the mines.
Both of these mines deal with the mining and processing of the fairly expensive metal beryllium. Therefore, efficiency is very important, as low prices are key for staying competitive in the raw materials industry.
In order to deal with this problem, we take the principal-agency theory as a starting point, and assume that the workforce in both mines are rational and opportunistic actors. In addition, given their similar utility functions (U=w^½-e), we assume that they act equally when facing the same opportunities – i.e. breaking a core assumption of

The reason for the workers stealing is obviously that they realise they can improve their expected income, and thereby utility, by stealing beryllium.
Assuming the problem will evolve to a substantial problem, it only makes sense that the prison sentence is the correct way to punish the workers. If they work at a $1 wage currently, then there is no punishment in firing and sending them to the farm, where they can work for an equal pay. The prison is disincentivising for the workers, since opportunity costs are high, and they are risk-averse.
One way to disincentivise stealing would be to introduce seniority incentives in the mine, and dispose of the workers caught stealing. While this might not be the most effective plan in the short run, due to the time value of money (Westerfield, 2011), it can be good for keeping loyal employees and building