enc_new – Enc-New forecast encompassing test (Clark-McCracken 2001).#

Back to nested_test axis | Back to L6 | Browse all options

Operational op under axis nested_test, sub-layer L6_B_nested, layer l6. Standalone callable: mf.functions.enc_new_test.

Function signature#

mf.functions.enc_new_test(
    loss_small: np.ndarray,
    loss_large: np.ndarray,
) -> EncNewTestResult

Parameters#

name

type

default

constraint

description

loss_small

np.ndarray

—

—

Squared losses for the small model.

loss_large

np.ndarray

—

—

Squared losses for the large model.

Returns#

EncNewTestResult — frozen dataclass with fit results.

Attribute

Type

Description

stat

float or None

Enc-New statistic

pvalue

float or None

One-sided p-value

decision

bool

Reject H0 at 5%

n_obs

int

Observations used

Behavior#

Tests whether the large model’s forecast contains information beyond the small (nested) model. Uses raw loss improvement f_t = loss_small - loss_large without CW adjustment, then applies one-sided DM inference. Complementary to the Clark-West test when the user does not want the CW penalty.

When to use

Testing forecast encompassing in nested model settings without the CW adjustment term.

When NOT to use

When the CW adjustment for bias is desired – use clark_west instead.

In recipe context#

Set params.nested_test = "enc_new" in the relevant layer to activate this op within a recipe:

# Layer L6 recipe fragment
params:
  nested_test: enc_new

References#

  • macroforecast design Part 3, L6: ‘tests must report (statistic, p-value, kernel, lag) and respect HAC dependence-correction.’

  • Clark & McCracken (2001) ‘Tests of Equal Forecast Accuracy and Encompassing for Nested Models’, JoE 105(2): 1-28.