Instrumental Variable
Definition
An instrumental variable helps, under certain assumptions, to identify causal effects despite unobserved confounders.
A valid instrument influences the explanatory variable, affects the outcome only through this variable, and must not be connected to relevant unobserved confounders. These assumptions are strong and must be justified.
Simply explained
Like using an external lever that changes the treatment but does not influence the outcome by any other path.
Why it matters in practice
Relevant for causal analyses when randomized experiments are not possible and suitable natural variation exists.
Example from business
A research team uses a plausibly exogenous policy change as an instrument to study the effect of a measure.
Limits and pitfalls
- Weak instruments; unjustified assumption of exclusion restriction; ignoring correlation of the instrument with confounders; generalizing local effects.
Origin & history
Instrumental variables are used under suitable assumptions to address endogenous regressors. The Stata documentation explains this for linear IV models.
Use in mid-sized companies
- Possible application: sophisticated impact analyses; in SMEs usually rather with statistical expertise or external specialists.
Frequently asked questions
Is correlation of the instrument with the treatment sufficient?
No. What is a weak instrument? One that explains the treatment only slightly and can make estimates unstable.
Sources
- Imbens & Rubin (1996), Identification of Causal Effects Using Instrumental Variables · Angrist, Imbens & Rubin (1996), Identification of Causal Effects Using Instrumental Variables
- ivregress, lineare IV-Modelle · Stata, ivregress, lineare IV-Modelle
Source check: 10./11.10.2026.
Editorially reviewed on: