Reference

This page documents the public scpcR API.

Overview

Function	Description
scpc()	Run spatial correlation-robust inference for a fitted model
print()	Print the main SCPC coefficient table
summary()	Print the coefficient table plus confidence intervals
coef()	Extract the coefficient estimates
confint()	Extract confidence intervals for selected coefficients

Conventions

Supported Models

scpc() accepts fitted:

• lm models
• fixest::feols() models, including IV specifications

Coordinates

Provide exactly one coordinate specification:

• lon and lat for geographic coordinates
• coords_euclidean for planar coordinates

Do not pass both specifications in the same call.

Clustering and Method Selection

• cluster names an optional clustering variable in data
• coordinates must be constant within clusters
• when a model uses absorbed fixed effects and external clustering, the conditional adjustment is not implemented; use uncond = TRUE
• method = "auto" selects the exact branch for smaller problems and the approximation branch once the large-n threshold is reached

Core Inference

scpc()

Run spatial correlation-robust inference.

Signature

scpc(
  model,
  data,
  lon = NULL,
  lat = NULL,
  coords_euclidean = NULL,
  cluster = NULL,
  ncoef = NULL,
  avc = 0.03,
  method = "auto",
  large_n_seed = 1,
  uncond = FALSE,
  cvs = FALSE
)

Parameters

• model: fitted model object. Supported classes are lm and fixest::feols(), including IV models.
• data (data.frame): data used to fit model. It must contain the coordinate columns referenced by lon / lat or coords_euclidean.
• lon, lat (character | NULL): longitude and latitude column names for geographic distance calculations.
• coords_euclidean (character | NULL): Euclidean coordinate column names. Use instead of lon and lat.
• cluster (character | NULL): optional clustering variable in data.
• ncoef (integer | NULL): number of coefficients to report. NULL reports all coefficients.
• avc (numeric): upper bound on average pairwise correlation. Must lie in (0.001, 0.99). Default: 0.03.
• method (character): spatial algorithm.
• "auto": choose between exact and approximate branches
• "exact": always use the full distance matrix
• "approx": always use the large-n approximation branch
• large_n_seed (numeric): integer-valued seed used by the large-n approximation branch. Ignored when method = "exact".
• uncond (logical): if TRUE, skip the conditional adjustment and report unconditional critical values only.
• cvs (logical): if TRUE, store per-coefficient critical values at the 32%, 10%, 5%, and 1% levels.

Returns

• object of class "scpc"
• scpcstats: matrix with coefficient estimates, SCPC standard errors, t-statistics, p-values, and 95% interval bounds
• scpccvs: matrix of stored critical values, or NULL when cvs = FALSE
• W: spatial projection matrix
• avc
• c0
• cv
• q
• method
• large_n_seed
• call

Methods

print.scpc()

Print the main SCPC coefficient table.

Signature

print(x, ...)

Parameters

• x: object of class "scpc"

Returns

• the input object, invisibly

The printed output shows the coefficient table with estimates, SCPC standard errors, t-statistics, and p-values. If cvs = TRUE, stored critical values are printed as well.

summary.scpc()

Print the fuller SCPC summary.

Signature

summary(object, ...)

Parameters

• object: object of class "scpc"

Returns

• the input object, invisibly

The summary adds the 95% confidence interval table and reports avc alongside the main coefficient output.

coef.scpc()

Extract the coefficient estimates.

Signature

coef(object, ...)

Parameters

• object: object of class "scpc"

Returns

• named numeric vector of coefficient estimates

confint.scpc()

Extract confidence intervals for selected coefficients.

Signature

confint(object, parm = NULL, level = 0.95, ...)

Parameters

• object: object of class "scpc"
• parm (character | numeric | NULL): coefficient names or indices to include. NULL uses all coefficients.
• level (numeric): confidence level to request.
• 0.95 is always available
• 0.68, 0.90, and 0.99 are also available when cvs = TRUE

Returns

• numeric matrix of lower and upper interval bounds

Return Object

"scpc"

Returned by scpc().

Fields

• scpcstats
• scpccvs
• W
• avc
• c0
• cv
• q
• method
• large_n_seed
• call

Methods

• print()
• summary()
• coef()
• confint()

The primary inferential output is stored in scpcstats. Its rows correspond to coefficients, and its columns contain the estimate, standard error, t-statistic, p-value, and 95% interval endpoints.