Economics of learning

This note is primarily directed to students. It is the fourth and final in a series on strategic cost analysis. This note will focus on the experience curve (also called the learning curve).

The learning curve was first described in psychology in the early 20th century. If we complete a task several times we will gradually become better at the task. In the beginning, we learn and improve quickly, but gradually it becomes more difficult to improve and eventually we reach a point where we do not learn anymore. Management scholars have applied the same basic model to companies. In certain industries (e.g. complicated assembly), there is a strong learning curve. In other industries there is no major room for learning (e.g. power plants). Learning can occur at different levels of an organisation; individuals learn, work teams learn, managers learn, business-units learn. If all learning happened at the individual level, it would be possible for a small company to catch up just by hiring personnel from the leading company. However, if a lot of the learning happens on the organisational level, it does not sufficient to just hire personnel.

Economics of learning is different from economics of scale. In economics of scale, we have the same shape of the cost curve in each year. In economies of learning, we are able to produce at a lower cost in each subsequent year.

Figure. Comparison between two companies that produce identical and same volume every year. Economies of scale are determined by the yearly production. Since the yearly production does not change, the blue company will have the same cost every year. In contrast, economies of learning are determined by the cumulative production. The red company will experience lower average cost with time.

Implications

The implications on industry analysis are mainly related to new entry.

• Threat of new entry. Large economies of learning can make it very difficult for a new entrant. With economies of scale, the new entrant will have a good cost position as long as it can produce at the MES. With economics of learning, the new entrant will have to go through a similar learning process as the incumbents so it is not sufficient to produce at the MES. In addition, the new entrant must gain a cumulative production volume to be able to learn. Eventually, the new entrant will catch up with the incumbents accumulated learning, but it could take years. (It is possible that there will be spillover learning across companies, giving the new entrant a head start.)
• Rivalry. It is hard to make a general statement regarding economics of learning and rivalry. In the 1970s, Boston Consulting Group argued that companies should strive to increase their market share so that they more quickly could move down the learning/experience curve. This lead to more rivalry in some industries. This view has since been discredited because it presupposed that learning is always a consequence of cumulative production volume (i.e. innovation does not exist). Tentatively, I would argue that there is no strong relationship between economies of learning and rivalry.

In industries where economies of learning are important, a company with a low-cost strategy needs to be quite concerned with cumulative production volume.

Summary

I have covered four different kinds of "economies". It is hard to say which is more important in general, but here is my attempt to a ranking: (1) economies of scale, (2) economies of capacity, (3) economies of learning, and (4) economies of scope.