diff --git a/README.md b/README.md index 12e6179..4b50b94 100644 --- a/README.md +++ b/README.md @@ -1,6 +1,8 @@ # PanelLocalProjections.jl [![CI](https://github.com/andrerecio/PanelLocalProjections.jl/actions/workflows/CI.yml/badge.svg?branch=main)](https://github.com/andrerecio/PanelLocalProjections.jl/actions/workflows/CI.yml) +[![docs: stable](https://img.shields.io/badge/docs-stable-blue.svg)](https://andrearecine.me/PanelLocalProjections.jl/stable/) +[![docs: dev](https://img.shields.io/badge/docs-dev-blue.svg)](https://andrearecine.me/PanelLocalProjections.jl/dev/) Panel local projections in Julia. For each horizon `h`, the package estimates a panel regression of `y_{i,t+h}` on a shock, and gathers the estimators and inference @@ -157,7 +159,7 @@ Plots.plot(s) # the same with Plots.jl (Makie and Plots both export `plo Small-sample factors, critical values, the Driscoll–Kraay bandwidth and the repair of a two-way variance are options of each specification (`ssc`, `psd`, `timelags`); see -[Inference](docs/src/guide/inference.md). +[Inference](https://andrearecine.me/PanelLocalProjections.jl/stable/guide/inference/). ## Usage by method @@ -226,8 +228,8 @@ panellp(@formula(ldiff(y2) ~ firstdiff(D2) + fe(t)), data; id = :id, time = :t, Other options: `control = :nevertreated` or `:notyet`, `firsttreat`, `oneoff` (a treatment that lasts one period), `untreated_before`, `pmd = k` or `:max` (pre-mean differencing), `weights`. It is written from the paper and reproduces the Stata command `lpdid` 1.1.0 -option by option; the [LP-DiD page](docs/src/methods/lpdid.md) has the map between the -two, and the [audit](docs/src/audit/lpdid.md) the places where they differ (the +option by option; the [LP-DiD page](https://andrearecine.me/PanelLocalProjections.jl/stable/methods/lpdid/) has the map between the +two, and the [audit](https://andrearecine.me/PanelLocalProjections.jl/stable/audit/lpdid/) the places where they differ (the reweighted estimate on unbalanced panels, among them). ### Pre-trends and the path of the treatment @@ -243,7 +245,7 @@ summarize(r, Clustered(:unit)) firststage(r; vcov = Clustered(:unit)) # Kleibergen–Paap, effective and Cragg–Donald F ``` -The [replication page](docs/src/replication/ilzetzki2024.md) reproduces his Figure 9 on +The [replication page](https://andrearecine.me/PanelLocalProjections.jl/stable/replication/ilzetzki2024/) reproduces his Figure 9 on the replication data (`singletons = :keep` and `Homoskedastic(ssc = :none)` for the published numbers). @@ -281,7 +283,7 @@ standard errors are 1.5 to 2.7 times those clustered by state, and none of the t bands excludes zero. The first stage shows the same thing: its Kleibergen–Paap F is about 2 clustered by period, but 35 robust and 200,000 clustered by state. The shock is aggregate, so the effective sample size is the 84 quarters, not the 4,284 state-quarters. -See the [example page](docs/src/examples/us_states.md) for the details. +See the [example page](https://andrearecine.me/PanelLocalProjections.jl/stable/examples/us_states/) for the details. Data: BEA Regional Economic Accounts; Board of Governors of the Federal Reserve System (H.15), via FRED; Jarociński & Karadi (2020, *AEJ: Macroeconomics*), updated surprises @@ -299,32 +301,32 @@ split-panel rule and critical values. The references are the Stata commands `xtl `xtivreg2`, `weakivtest`, `harreg` and `lpdid`, R `fixest`, and the authors' own R code (Mei–Sheng–Shi's `LP_panel`, Almuzara–Sancibrián's `panel_LP.R`). [`crosscheck/README.md`](crosscheck/README.md) maps each reference command to the `panellp` options that reproduce it, and -[Deviations](docs/src/deviations.md) lists the defaults that differ from a reference. +[Deviations](https://andrearecine.me/PanelLocalProjections.jl/stable/deviations/) lists the defaults that differ from a reference. ## Documentation -The pages are in [`docs/src`](docs/src): - -- Guide: [getting started](docs/src/guide/example.md), [panel data](docs/src/guide/data.md), - [formula](docs/src/guide/formula.md), [inference](docs/src/guide/inference.md). -- Methods: [fixed-effect absorption](docs/src/methods/absorption.md), - [split-panel jackknife](docs/src/methods/spj.md), [CD-OLS](docs/src/methods/cdols.md), - [t-LAHR](docs/src/methods/tlahr.md), [panel LP-IV](docs/src/methods/lpiv.md), - [lead paths and negative horizons](docs/src/methods/leads.md), - [LP-DiD](docs/src/methods/lpdid.md). -- Example and replications: [US states](docs/src/examples/us_states.md), - [Mei, Sheng & Shi](docs/src/replication/mss2026.md), - [Ugarte-Ruiz](docs/src/replication/ugarte2026.md), - [Ilzetzki](docs/src/replication/ilzetzki2024.md) (on his replication data, not +The documentation is at for the latest release and for `main`: + +- Guide: [getting started](https://andrearecine.me/PanelLocalProjections.jl/stable/guide/example/), [panel data](https://andrearecine.me/PanelLocalProjections.jl/stable/guide/data/), + [formula](https://andrearecine.me/PanelLocalProjections.jl/stable/guide/formula/), [inference](https://andrearecine.me/PanelLocalProjections.jl/stable/guide/inference/). +- Methods: [fixed-effect absorption](https://andrearecine.me/PanelLocalProjections.jl/stable/methods/absorption/), + [split-panel jackknife](https://andrearecine.me/PanelLocalProjections.jl/stable/methods/spj/), [CD-OLS](https://andrearecine.me/PanelLocalProjections.jl/stable/methods/cdols/), + [t-LAHR](https://andrearecine.me/PanelLocalProjections.jl/stable/methods/tlahr/), [panel LP-IV](https://andrearecine.me/PanelLocalProjections.jl/stable/methods/lpiv/), + [lead paths and negative horizons](https://andrearecine.me/PanelLocalProjections.jl/stable/methods/leads/), + [LP-DiD](https://andrearecine.me/PanelLocalProjections.jl/stable/methods/lpdid/). +- Example and replications: [US states](https://andrearecine.me/PanelLocalProjections.jl/stable/examples/us_states/), + [Mei, Sheng & Shi](https://andrearecine.me/PanelLocalProjections.jl/stable/replication/mss2026/), + [Ugarte-Ruiz](https://andrearecine.me/PanelLocalProjections.jl/stable/replication/ugarte2026/), + [Ilzetzki](https://andrearecine.me/PanelLocalProjections.jl/stable/replication/ilzetzki2024/) (on his replication data, not distributed), - [firm-level studies](docs/src/replication/firm_level_wip.md) (call mappings, not run on + [firm-level studies](https://andrearecine.me/PanelLocalProjections.jl/stable/replication/firm_level_wip/) (call mappings, not run on the data). -- [Cross-checks](crosscheck/README.md), [deviations](docs/src/deviations.md) and - [audits of the reference code](docs/src/audit). +- [Cross-checks](https://andrearecine.me/PanelLocalProjections.jl/stable/crosscheck/), [deviations](https://andrearecine.me/PanelLocalProjections.jl/stable/deviations/) and + audits of the reference code, starting with [`xtlp`](https://andrearecine.me/PanelLocalProjections.jl/stable/audit/xtlp/) and [`lpdid`](https://andrearecine.me/PanelLocalProjections.jl/stable/audit/lpdid/). - API: every exported name has a docstring (`?panellp`); the list is in - [`docs/src/api.md`](docs/src/api.md). + the [API page](https://andrearecine.me/PanelLocalProjections.jl/stable/api/). -To build the HTML pages locally (Julia 1.11 or later), into `docs/build`: +The sources are in [`docs/src`](docs/src). To build the HTML pages locally (Julia 1.11 or later), into `docs/build`: ```bash julia --project=docs -e 'using Pkg; Pkg.instantiate()'