HOW TO BUILD AN ECONOMIC MODEL
An economic model is a simplified representation of reality that is used to generate testable hypotheses regarding the economic activity. Because there are no objective indicators of economic results, one of the most fundamental characteristics of an economic model is that it must be subjective in design. Different economists will arrive at different conclusions on what is required to explain their views on reality.
Theoretical and empirical economic models are the two types of models. Theoretical models attempt to extract provable conclusions about economic behavior based on the assumption that agents maximize certain objectives within well-defined model restrictions (for example, an agent’s budget). They provide qualitative answers to specific concerns, such as the consequences of asymmetric knowledge (where one party to a transaction knows more than the other) or how to best deal with market failures.
Empirical models, on the other hand, strive to validate the qualitative predictions of theoretical models and translate them into precise, numerical outcomes. A theoretical model of an agent’s consuming behavior, for example, would normally suggest that expenditure and income have a positive connection. The theoretical model’s empirical adaption would seek to assign a numerical number to the average increase in expenditure as income rises.
A set of mathematical equations that describe a theory of economic activity is called an economic model. Model designers strive to include enough equations to provide useful insights into how rational agents behave or how an economy works. The structure of the equations reflects the model builder’s attempt to simplify reality by assuming an endless number of competitors and market participants with perfect foresight, for example. In actuality, economic models can be relatively simple: for example, if all other factors remain constant, demand for apples is inversely related to price. The demand for apple increases as the price of apples decreases. Models can also be quite complicated: some models that attempt to anticipate an economy’s true level of output employ hundreds of complex formulations known as “nonlinear, interconnected differential equations.”
Economic models can also be divided into categories based on the regularities they are supposed to describe or the problems they are supposed to answer. Some models, for example, describe the economy’s ups and downs in terms of a changing long-run path, emphasizing the demand for products and services rather than being too precise about long-run growth sources. Other models are built to focus on structural concerns like the influence of trade changes on long-term output levels while neglecting short-term fluctuations. Economists also use models to investigate “what-if” scenarios, such as the influence of a value-added tax on the broader economy.
How economists build empirical models
Empirical economic models, despite their differences, share some characteristics. Each will allow for inputs, or exogenous variables, that the model does not need to explain. Policy variables, such as government expenditure and tax rates, as well as nonpolicy variables, such as the weather, are examples of these. Then there are the outputs, which are referred to as dependent variables (for example, the inflation rate), which the model will attempt to explain when one or more exogenous factors are present.
Every empirical model will include coefficients that determine how a dependent variable responds to changes in input (for example, the responsiveness of household consumption to a $100 reduction in income tax). Typically, such coefficients are estimated (given numbers) using historical data. Finally, empirical model builders include a catchall variable in each behavioral equation to account for individual differences in economic behavior. (A $100 tax rebate will not elicit the same response from agents as in the previous scenario.)
However, economists have fundamental disagreements on how an empirical model’s equations should be derived. Some economists argue that the equations must include maximizing behavior (for example, an agent picks future spending to maximize its level of happiness within a given budget), efficient markets, and forward-looking conduct. The expectations of agents and how they react to policy changes are crucial in the equations that result. As a result, users of the model should be able to track the impact of specific policy changes without having to worry about whether the change affects the behavior of agents.
Others advocate a more nuanced approach. Their favorite equations reflect, in part, what they’ve learned about observable data from their personal experiences. Economists who design models in this manner are essentially casting doubt on the reality of the behavioral constructs in more technically derived models. Because the underlying equations do not explicitly account for changes in agent behavior, incorporating experience frequently makes it hard to untangle the effect of specific shocks or estimate the effects of a policy change. The benefit, according to these same economists, is that they are better at forecasting (especially for the near term)
What makes a good economic model?
The scientific process (many fields, such as physics and meteorology, develop models) necessitates that every model gives exact and provable conclusions about the economic phenomena it is attempting to describe, regardless of the approach. The model’s principal implications are tested, and its capacity to generate stylized facts is assessed in a formal evaluation. Case studies, lab-based experimental research, and statistics are all used by economists to evaluate their models.
Still, because economic data is often unpredictable, economists must be precise when claiming that a model “effectively explains” something. In terms of forecasting, this indicates that errors are unpredictable and, on average, inconsequential (zero). When two or more models meet this requirement, economists typically use the forecast error volatility to break the tie — lower volatility is normally favored.
If an empirical model creates systematic forecasting errors, this is an objective indicator that needs to be changed. Systematic errors indicate that one or more of the model’s equations are erroneous. Understanding why such errors occur is an important aspect of how economists evaluate models on a regular basis.
Why models fail
No matter how complex, all economic models are subjective approximations of reality intended to explain observed occurrences. As a result, the model’s predictions must be tempered by the unpredictability of the underlying data it is attempting to explain, as well as the validity of the theories used to create the model’s equations.
The ongoing dispute about existing models’ failure to forecast or unravel the causes of the recent global financial crisis is a notable example. It has been blamed on a lack of focus on the connections between overall demand, wealth, and, in particular, excessive financial risk-taking. There will be a lot of research done in the coming years to identify and grasp the lessons learned from the catastrophe. Current economic models will be updated with new behavioral equations as a result of this research. It will also necessitate the modification of current equations (for example, those relating to family saving behavior) in order to relate them to the new financial sector equations. The upgraded model’s ultimate test will be its ability to regularly identify levels of financial danger that necessitate a proactive policy response.
There is no such thing as a flawless economic model that accurately describes reality. The process of building, testing, and refining models, on the other hand, compels economists and policymakers to tighten their beliefs on how an economy operates. As a result, scientific discussion about what drives economic behavior and what should (or should not) be done to address market failures is gaining traction. Adam Smith would most likely agree.
How To Build An Economic Model In Your Spare Time
Here are the points to take away:
1. Look for ideas in the world, not in the journals.
2. First, make your model as simple as possible, then generalize it.
3. Look at the literature later, not sooner.
4. Model your paper after your seminar.
5. Stop when you’ve made your point.