# What changed in statsmodels

> 19 changes in statsmodels 0.14 to 0.15 from its official release notes, upgrade guides and advisories: 4 breaking, 12 new APIs and options. Each one cited. Sources checked 2026-10-28. HTML: https://unstale.sh/pypi/statsmodels/

To see which of these apply to a repository, run `npx @unstale/cli init` in it.

## statsmodels 0.15

### Breaking changes

- The minimum supported versions of NumPy, SciPy, and pandas are raised. (0.15.0) — [release notes](https://github.com/statsmodels/statsmodels/releases/tag/v0.15.0)
- Functions with variable-length tuple returns now return documented NamedTuple results. (0.15.0) — [release notes](https://github.com/statsmodels/statsmodels/releases/tag/v0.15.0)
- Randomness is controlled with rng across the package, replacing seed and random_state. (0.15.0) — [release notes](https://github.com/statsmodels/statsmodels/releases/tag/v0.15.0)
- Input validation is tighter across many string-valued options; invalid values now raise a clear ValueError instead of failing silently or with a confusing error. (0.15.0) — [release notes](https://github.com/statsmodels/statsmodels/releases/tag/v0.15.0)

### New APIs and options

- A new robust estimator is available. (0.15.0) — [release notes](https://github.com/statsmodels/statsmodels/releases/tag/v0.15.0)
- A new delete-k block jackknife estimator is available. (0.15.0) — [release notes](https://github.com/statsmodels/statsmodels/releases/tag/v0.15.0)
- A new Diebold-Mariano test is available. (0.15.0) — [release notes](https://github.com/statsmodels/statsmodels/releases/tag/v0.15.0)
- HurdleCountModel gains fit_regularized. (0.15.0) — [release notes](https://github.com/statsmodels/statsmodels/releases/tag/v0.15.0)
- A new abstracted formula engine supports patsy and formulaic as backends. (0.15.0) — [release notes](https://github.com/statsmodels/statsmodels/releases/tag/v0.15.0)
- The Hamilton filter is newly available. (0.15.0) — [release notes](https://github.com/statsmodels/statsmodels/releases/tag/v0.15.0)
- A new Jonckheere-Terpstra test is available. (0.15.0) — [release notes](https://github.com/statsmodels/statsmodels/releases/tag/v0.15.0)
- An L1-penalized GLM solver is newly available. (0.15.0) — [release notes](https://github.com/statsmodels/statsmodels/releases/tag/v0.15.0)
- A new Leybourne-McCabe test is available. (0.15.0) — [release notes](https://github.com/statsmodels/statsmodels/releases/tag/v0.15.0)
- A local false discovery rate correction is newly available. (0.15.0) — [release notes](https://github.com/statsmodels/statsmodels/releases/tag/v0.15.0)
- A new Pesaran-Timmermann test is available. (0.15.0) — [release notes](https://github.com/statsmodels/statsmodels/releases/tag/v0.15.0)
- Polars DataFrames and Series can be used as model input. (0.15.0) — [release notes](https://github.com/statsmodels/statsmodels/releases/tag/v0.15.0)

### Behavior and default changes

- MICEData is now iterable. (0.15.0) — [release notes](https://github.com/statsmodels/statsmodels/releases/tag/v0.15.0)

## statsmodels 0.14

### Fixes and other changes

- Includes minor changes to improve future compatibility with NumPy. (0.14.6) — [release notes](https://github.com/statsmodels/statsmodels/releases/tag/v0.14.6)
- Fixes an issue with pandas 3.0.0 that prevented statsmodels from importing. (0.14.6) — [release notes](https://github.com/statsmodels/statsmodels/releases/tag/v0.14.6)