Spatial Unit Root Diagnostic Tests

Usage

spurtest(
  formula,
  data,
  coords = NULL,
  type = c("i1", "i0", "i1resid", "i0resid"),
  latlong = FALSE,
  q = 15,
  nrep = 1e+05,
  seed = NULL,
  ...
)

spurtest_i1(
  formula,
  data,
  coords = NULL,
  latlong = FALSE,
  q = 15,
  nrep = 1e+05,
  seed = NULL,
  ...
)

spurtest_i0(
  formula,
  data,
  coords = NULL,
  latlong = FALSE,
  q = 15,
  nrep = 1e+05,
  seed = NULL,
  ...
)

spurtest_i1resid(
  formula,
  data,
  coords = NULL,
  latlong = FALSE,
  q = 15,
  nrep = 1e+05,
  seed = NULL,
  ...
)

spurtest_i0resid(
  formula,
  data,
  coords = NULL,
  latlong = FALSE,
  q = 15,
  nrep = 1e+05,
  seed = NULL,
  ...
)

Arguments

formula

A formula specifying the variable(s) to test:

• For simple tests: y ~ 1 or just the variable name.

• For residual tests: y ~ x1 + x2.

Alternatively, a feols model object.

data

A data frame containing the variables.

coords

Coordinate specification: a formula (e.g., ~ s_1 + s_2), a matrix, or NULL to auto-detect s_* variables.

type

Type of test: "i1", "i0", "i1resid", or "i0resid".

latlong

Logical; are coordinates latitude (s_1) and longitude (s_2) in decimal degrees?

q

Number of low-frequency weighted averages. Default: 15.

nrep

Number of Monte Carlo replications. Default: 100000.

seed

Optional seed for reproducibility.

...

Additional arguments passed to methods.

Value

An object of class "spur_test" (and subclass depending on type) with components:

statistic

LR test statistic.

p_value

Monte Carlo p-value.

ha_parm

Parameter for the alternative hypothesis.

critical_values

Named vector of critical values (1\ typeType of test performed. qNumber of low-frequency components used. nrepNumber of Monte Carlo replications. callThe matched call.

Implements four tests from Mueller and Watson (2024) for diagnosing spatial unit roots: I(1) test, I(0) test, and their residual variants. Mueller, U. K. and Watson, M. W. (2024). "Spatial Unit Roots and Spurious Regression." Econometrica, 92, 1661–1695.