Using the Models as a Teaching Device

We have found this type of model pedagogically useful in two ways, as a method of demonstration in the classroom and as the basis for guided assignment work. 

Classroom Demonstration

Most modern classrooms are equipped with computer projection equipment, and Excel is almost universal, so using the models for demonstration is usually quite easy. If you use Powerpoint to organize your classes, it is straightforward to insert a link in your presentation to open the simulation at the appropriate juncture. 

In terms of mechanics, a problem that may arise is that, because the model interfaces are sometimes quite large, it may be necessary to take measures to ensure it will fit legibly on the screen. Screen real estate can be maximized by minimizing the task bar and viewing in full screen mode. If the image still doesn't fit, a custom zoom can be used. Many of the models also contain multiple graphics, only some of which are likely to be used at any one time, others can be hidden or shifted off screen. We also tend to use spinners to control inputs where possible, to avoid accidental typos that might necessitate a restart.

We usually use these models for demonstrations in class after teaching the underlying theory, in the same way as a worked numerical example might be used as a reinforcement device in other contexts. To engage the students in the process, we solicit student input on the direction and magnitude of the changes to be considered, and get their predictions (along with reasoning) on the outcome before making the change in the model. For example, if we have learned about Stolper-Samuelson, we would ask the students for a recommended change in one of the prices. Then, given the data in the sheet, we query the students on the pattern of factor intensity. We then decide on the pattern of changes in factor prices. Finally, we simulate the change using the appropriate model, and verify (or in some cases refute, if the student's logic is incorrect) the prediction, and discuss again why the results look the way they do. Because the Excel models can quickly handle multiple scenarios, we can rapidly consider many changes, something that is difficult using other approaches. Again using Stolper-Samuelson as an example, while formal notes might contain only a single price change scenario, with a simulation model we can quickly demonstrate price changes for both goods in both directions. This is useful in reinforcing the general pattern.

Guided Assignments

After we have introduced models during class and the students are comfortable with the mechanics of working with them, we often use them as the basis of class assignments. The models are made available on the class website, sometimes in a specially parameterized version, along with assignment sheets that request the students undertake a series of experiments and record and analyze the results. In general we try and construct the assignments so that they are quasi-experimental. The experiments are designed to generate a particular outcome if the instructions are correctly followed, but the students are not told what to expect. We feel that this approach heightens the feeling of discovery by the students, and encourages them to think carefully about and take ownership of the results. Many students have commented that they have found undertaking the assignments and discovering the patterns for themselves made the results more compelling and real to them. A few examples illustrate the approach:

Example 1: This experiment is based on the large country tariff and retaliation ('trade war') sheet. In it students simulate the process of a trade war that leads to a Nash equilibrium. The learning objectives are that the optimal tariff argument breaks down in the face of retaliation, that a trade war can result in an equilibrium in which there are no real winners, and that the WTO can be seen as a mechanism for maintaining a non-Nash outcome (free trade) that Pareto dominates the Nash outcome (i.e., a type of solution to the prisoners' dilemma). <Download>

Example 2: This is based on the customs union sheet. Students simulate a customs union under various conditions. The objective is to help students get a feel for factors that are likely to minimize trade diversion and maximize trade creation, i.e., high initial tariffs, low post-union tariffs, relative efficiency of the partners, and so on. <Download>

Example 3: This assignment is based on the two-country version of the HOS model. The students simulate a trading equilibrium and then consider what happens to one country when the other economy grows in various ways (the assignment is used to motivate discussion of US-China trade relations). The students find that the welfare level of the country that is not growing is affected by the changes in the economy that is growing. They should find that growth that makes the partner more apt to trade will raise welfare, while growth that makes the partner less apt to trade will lower welfare. Some students conclude that we should not trade when a partner grows the 'wrong' way. However, by comparing the welfare outcomes to autarky, many students realize that even though growth in a trading partner might lower welfare, it remains superior to autarky. <Download>