Deviations from the reference implementations

Where this package's defaults differ from a reference implementation, the table gives the reason and, where a published number depends on it, the switch that reproduces the reference. The audit pages document each reference in detail.

Mei, Sheng & Shi (LP_panel_all.r, xtlp)

topicreferencepanellp defaultswitch
two-way demeaningone unit-then-time sweep (R code): exact only on balanced samplesalternating projections to convergenceabsorb = OneSweep()
split-panel jackknife halvesfirst $\lceil n_i/2 \rceil$ rows (R code, xtlp); the paper writes $\lfloor n_i/2 \rfloor$split = :ceil, as the codeSplitPanelJackknife(split = :floor)
small-sample factors (R code, robust = T)$N/(N-1) \cdot (n-1)/(n-K)$; two-way $G/(G-1) \cdot (n-1)/(n-N-K)$, $G = \min(N, T)$reghdfe's factors (ssc = :stata)Clustered(:unit or :unit, :time; ssc = :mss)
repair of a two-way variance that is not positive semi-definiteR code: clip the eigenvalues of V (eigens = T); xtlp: clip in the metric of the regressors' standard deviationspsd = :stata, as xtlp and reghdfepsd = :raw
Driscoll–Kraay lags (xtlp, xtscc)by position among the sample's periodsby calendar periodtimelags = :position
xtlp's e(irf) bands±1.96 standard errors whatever the variancethe specification's critical valuessummarize(…; dist = :normal)
transformed outcomes (xtlp)stored as floats unless set type doubledouble—
a horizon whose variance is undefined (xtlp)coefficients posted as missingcoefficients kept, variance NaN—

Ugarte-Ruiz (2026)

topicpaperpanellpswitch
specification rulesstated in the textchecked (CDOLS())CDOLS(validate = false)
variance"robust", formula not stated (heteroskedasticity-robust, presumably)any; unit clusters are valid, robust ones are notRobust()

Almuzara & Sancibrián (panel_LP.m/.R)

topicreference codepanellp defaultswitch
lags$\min(h, p_{\max})$: no lag at $h = 0$; no information criterionthe paper's $\max\{1, \min(h, \hat p)\}$ with AIC (ICLags())LagAugmented(CappedLags(p_max))
the HC2/Imbens–Kolesár refinementoff by defaulton (the paper's recommendation)TLAHR(refine = false)
synthetic regressionraw exposure $s_{it}$; periods unweightedthe exposure residualised per period (the paper's $\hat s_i$, and on the exposures of the other regressors of the same aggregate series); periods weighted by $\omega_t$ (fn 24: CR2 by period)TLAHR(weights = :raw)
degrees-of-freedom contrast$e_1$ for every coefficient (a bug)each coefficient's own—
singletonskeptdropped, as reghdfe and xtlpsingletons = :keep
no fixed effectsno interceptan intercept0 +
Date time index (R)broken (day counts)calendar periods with frequency—

Dube, Girardi, Jordà & Taylor (lpdid 1.1.0)

topiclpdidpanellp (LPDiD)switch
reweighting factor of the weighted regression (rw)the treated share of the period among clean rows with the outcome, including rows whose $\Delta D$ is undefined and that the regression dropsthe treated share among the rows of the regression (the equally weighted average of the cohorts' comparisons)reweight = :lpdid
a treatment that switches off, without nonabsorbing()kept, with a regressor of −1an errornonabsorbing = L
nocomp with a one-sided windowalso requires the pre-treatment horizon 2 (or horizon 0)the requested horizons (sample = :common)horizon = -2:H
pooled regressions under nocompthe flags of both pooled windowsthe horizons of the window—
$h = -1$ under pmd()reported as zeroestimated—
standard errors of regression adjustmentno small-sample factor, normal critical valuesthe specification'sClustered(:unit; ssc = :none)
wild bootstrapbootstrap() through boottestnot implemented—

Ilzetzki (2024, ivreg2 with dummies)

topicreplication packagepanellp defaultswitch
varianceclassical, $RSS/N$ (no option to ivreg2)none: the variance is chosen at summarizeHomoskedastic(ssc = :none)
critical valuesStudent t with $N - K$, on a standard error without correctionnormal for ssc = :nonethe degrees of freedom of Homoskedastic()
singletons of the line and month dummieskeptdroppedsingletons = :keep
first-stage FCragg–DonaldKleibergen–Paap with the chosen variancefirststage(r).cdF

Variance conventions shared with other software

topicreferencepanellpswitch
Driscoll–Kraay bandwidthxtscc, xtlp: $\lfloor 4(T/100)^{2/9} \rfloor$ lags; Almuzara & Sancibrián: kernel bandwidth $\lceil 4((T-h)/100)^{2/9} \rceil$DriscollKraaySE(:nw94):as2026 ($M - 1$ lags), :mss2026, an integer
Driscoll–Kraay factorxtscc and xtlp …, nodfadj: $(n-1)/(n-K-1) \cdot T/(T-1)$xtlp's default (ssc = :stata)ssc = :xtscc; ssc = :none (ase)
fixest's two-way varianceeigenvalues floored at 1e-16 (vcov_fix)psd = :statapsd = :raw (clips at 0)
equal-weighted cosine varianceharreg: $B = \lfloor 0.41\,T^{2/3} \rfloor$, t($B$)the same (CosineHAR())an integer $B$
absorption stopping rulextlp: absolute step below 1e-9; the Almuzara–Sancibrián code: 1e-8 (MATLAB: absolute norm of the change; R: fixest::demean)relative to the column after its first sweep, 1e-12AlternatingProjections(tol = …)