Birks – Model building in economics
Many economics courses and textbooks focus very heavily on models as representations of the economic phenomena under investigation. These could be considered as analogies for the real world, alternative representations which may be similar, in some respects, to the real world.
An understanding of the process of model development can be helpful for understanding the material, especially when the texts or courses do not make the process explicit. Consider the following brief outline:
- A model in economics is constructed by specifying selected variables, some of which are exogenous and others which are endogenous.
- The values of exogenous variables are set outside the model. They are not explained by the model, so their values are treated a ‘given’.
- The values of endogenous variables are set within the model according to specified relationships with the other variables, both endogenous and exogenous.
- Note that the exogeneity of some variables is a feature of the model, not of the variables themselves. Policy variables can be included in this category, with their values set by policy makers.
- Some exogenous variables are not even explicitly mentioned. Instead, you may see the term, ‘ceteris paribus’, a Latin expression meaning, ‘everything else staying the same’. In other words, all omitted variables are assumed to be constant. Of course, they may not be constant in reality, but we assume them to be constant in the model so as to separate out the effects of the variables of interest (this might be unrealistic, in which case the model could give misleading results).
- Note also that the term ‘exogenous preferences’ in the context of model construction does not mean that preferences are fixed. It just means that they are not determined within the model. Anything that might cause preferences to change (such as advertising) is beyond the scope of the model.
- Similarly short- and long-run in relation to models are not issues of time, they relate to assumptions as to what can vary. This can be clearly seen in the theory of the firm. It is for reasons such as this that it is important to understand the process of model building and the nature of models.
- In the initial stages, a model could be quite basic and unrealistic. A basic model ‘explains’ very little. It commonly has few endogenous variables.
- An economist might use this as a starting point, attempting to understand its characteristics.
For a microeconomics example, consider a supply and demand model where quantities supplied and demanded depend on price alone. All other possible variables are assumed to be constant (the ceteris paribus assumption) or irrelevant to the relationships, but a general structure is established, including the concept of equilibrium, along with questions of its existence, uniqueness and stability. The ceteris paribus assumption can then be relaxed. Additional variables (consumers’ incomes, prices of other goods, a sales tax) can be introduced and their effects considered. Their determinants can then be included also, or the functional relationships made more complex.
For a macroeconomics example, consider a simple circular flow diagram with no leakages or injections. Then add savings as a withdrawal and investment as an injection. Also add a government sector with taxation as a withdrawal and government expenditure on goods and services as an injection. Add government transfer payments as an additional injection (or as negative taxes). Bring in international trade with expenditure on imports as a withdrawal and income from exports as an injection. So far the interest rate and price level are assumed constant (as exogenous variables, perhaps not even mentioned). Allow the interest rate to vary, considering its effects on investment, perhaps other aspects of expenditure, maybe international capital movements, balance of payments and exchange rates. Add a financial sector to give some explanation of interest rate determination. These are the sorts of steps you will see in a standard course.
- As the models develop, they get more involved, giving some possible explanations for more and more phenomena. The explanations are all limited by the dimensions and specification of the model, although it may be hoped that they gradually come to more closely resemble the real world.
- Where models are intended to aid in policy making, the design will depend in part on the policy variables being considered. A Reserve Bank will focus on monetary variables, interest rates and reserve ratios, for example, whereas Treasury analysts would be more concerned about the effects of fiscal variables.
- One book which goes into some detail on the nature of models in economics is Morgan (2012).
- As a word of caution, some economists (Lawson, 1997, 2003) argue that these models represent ‘closed’ systems, whereas the real world is a ‘open system’, so the models can never be realistic.
It is important to consider the constraints of a model, including its simplifying assumptions and additional aspects to be considered in the real world. These can be thought of in terms of the nature of framing and the processes that may be followed when analysing issues. This perspective implicitly conveys the limited, partial nature of any approach and hence the possibility of numerous, similarly limited, alternatives.
There is a benefit from seeing the material this way, even for those who wish only to learn the basic textbook material. It is clear that the findings relate to abstract models. They do not provide definitive representations of real world phenomena. The approach has the added advantage that it helps to build up critical skills and to see processes whereby issues can be analysed, along with the limitations of those processes.
Lawson, T. (1997). Economics and reality. London: Routledge.
Lawson, T. (2003). Reorienting economics. London: Routledge.
Morgan, M. S. (2012). The World in the Model: How Economists Work and Think. Cambridge: Cambridge University Press.
Commentary added 20th November 2014, last updated 5 January 2015