An asymptotic distribution is the probability distribution that a statistic approaches as the sample size becomes very large.
$$$$
Why economists use it
Exact finite-sample distributions are often complicated or unknown. Large-sample approximations let economists build confidence intervals, hypothesis tests, and standard errors even when the exact distribution is hard to derive.
A common example
For many estimators:
$$
\sqrt{n}(\hat{\theta}-\theta_0) \xrightarrow{d} N(0,V)
$$
This means the scaled estimation error converges in distribution to a normal random variable with variance (V).
What “asymptotic” does and does not mean
It does not mean the approximation is perfect in small samples. It means the approximation becomes more accurate as sample size grows. Economists still need to ask whether the available sample is large enough for the asymptotic result to be useful in practice.
Knowledge Check
### An asymptotic distribution describes:
- [x] the limiting behavior of a statistic as sample size grows
- [ ] the exact finite-sample distribution in every case
- [ ] only macroeconomic trends
- [ ] the path of a price index over time
> **Explanation:** It is a large-sample approximation, not necessarily an exact small-sample result.
### Why is asymptotic distribution theory useful in econometrics?
- [x] It allows inference even when exact finite-sample distributions are hard to derive
- [ ] It removes the need for assumptions
- [ ] It applies only to descriptive statistics
- [ ] It guarantees perfect inference in tiny samples
> **Explanation:** Large-sample approximations are a practical foundation for statistical inference.
### If a sample is small, an asymptotic approximation may:
- [x] be less accurate than in large samples
- [ ] become automatically exact
- [ ] eliminate sampling error
- [ ] make the estimator unbiased by definition
> **Explanation:** The usefulness of asymptotics depends on how quickly the large-sample approximation becomes reliable.