What it means
A model describes relationships among variables, with some inputs treated as given while others are solved through those relationships. Calling a variable endogenous identifies its role in that particular model, not a permanent characteristic of the quantity itself.
In a basic supply-and-demand model, if supply and demand equations jointly determine market price and quantity, both are endogenous outcomes, while an externally specified tax rate or input cost can be treated as exogenous even though a broader model might explain it too. Changing the model's scope can change the classification: a small firm's forecast may treat the market interest rate as an outside assumption, whereas a model of monetary policy and financial conditions may explain that rate through interactions inside the system.
Model determination is not the same as a proven causal relationship, since equations express relationships that need evidence, and a variable can be solved inside a poorly specified model without the endogenous label making the model correct. The Iowa State University economic-modelling reference explicitly separates endogenous, model-determined variables from exogenous variables given outside the model, which helps a reviewer see which results come from the model and which depend on chosen inputs, and both the relationships and the external assumptions deserve scrutiny.
Econometric usage requires particular care, because an explanatory variable is endogenous in a regression when it is correlated with the unobserved error term, which can arise from omitted factors, simultaneous determination or measurement problems. For example, advertising expenditure and sales may respond to the same unobserved demand outlook, so a simple regression can attribute that shared influence to advertising, and a positive coefficient does not by itself establish the additional sales caused by spending another dollar.
Reverse causality is another possibility, since a business may spend more on advertising because it expects stronger sales while advertising may also affect sales, and observing that both variables move together cannot settle the direction of influence. Do not call every correlated pair endogenous, because ordinary correlation among measured variables differs from correlation between an explanatory variable and the regression error, and an exogenous input need not be unrelated to every other variable in a dataset.
Methods such as instrumental variables can address certain endogeneity problems, but they require additional assumptions, and a proposed instrument must satisfy relevant validity conditions rather than simply have a different name or come from another spreadsheet. A complicated estimation method does not remove the need for evidence.
Scenario analysis also depends on boundaries, since changing an outside input while allowing endogenous variables to adjust differs from fixing every outcome independently, which can create combinations that violate the model's own supply, demand or financial constraints. For a non-finance manager, ask which variables are inputs, which are outcomes and what evidence supports causal claims, and when reviewing a regression, ask whether omitted influences or feedback affect the explanation.
These questions help separate a useful forecast from an unsupported prescription.
In practice
Real-world examples.
Example
In a market model, demand is 100 minus price and supply equals price. Solving the two relationships determines price and quantity together. Neither should be inserted independently without checking whether the equations remain consistent.
Example
A company treats the policy rate as a fixed forecasting assumption. A central-bank model explains that same rate through policy responses to inflation and output. The variable's classification changes because the models have different boundaries.
Example
A retailer finds that stores with larger advertising budgets sell more. Management also assigns larger budgets to stores in stronger local markets. The analyst considers that omitted demand condition before claiming advertising caused the entire difference.
Formula
Calculation
Illustrative simultaneous model: demand Q = 100 - P and supply Q = P. Equating them gives 100 - P = P, so P = 50 and Q = 50. This simplified calculation demonstrates internally determined variables, not a reliable forecast for an actual market or proof that its relationships are causal.Case study
Seen in the real world.
Fictional case: A manager recommends higher discounts after a regression links discounts with greater sales. Finance notes that discounts were concentrated in stores facing weak demand and excess stock. The team revises the analysis and tests the decision more carefully rather than treating the initial coefficient as a causal answer.
Watch out
Common mistakes.
- Treating endogenous status as a permanent property independent of the model.
- Confusing correlation between observed variables with correlation involving a regression error.
- Assuming an internally solved equation or estimated coefficient automatically proves causation.
Questions
People also ask.
Can the same variable be endogenous in one model and exogenous in another?
Yes. Its role depends on the model's scope and assumptions.
Does endogenous mean every variable is correlated with every other variable?
No. That is not the definition in either economic modelling or econometric estimation.
Does a positive regression coefficient establish causation?
No. Feedback, omitted factors or other endogeneity problems can affect the estimate.
From the founder's library
Take it further with the book.
Build your financial confidence beyond this definition. Shihan's full-length guide, Accounting Fundamentals, takes the same plain-English approach and turns it into a complete, practical playbook for non-finance managers, business owners and students - with chapter-end quiz answers and presentation slides included.
25% off with code MMHQ25, applied at checkout. Priced in USD - checkout may show the equivalent in your local currency.
View the book and save 25%