Draw SPI incomes within earnings groups set by FRS employment status

changelog.d/spi-income-earnings-groups.fixed.md

@@ -0,0 +1 @@
+Draw SPI incomes for the enhanced FRS's SPI-synthetic rows within earnings groups (employee, self-employed, both, neither) set by each row's FRS employment status, so employees draw pay, the self-employed draw a trade, people out of work draw no earnings, and children keep their own (zero) incomes (#504). Calibrate to the ONS LFS counts of employees and the self-employed, and report rather than train on HMRC's local counts of taxpayers with employment income, which are annual and include part-year earners the FRS records as out of work.

policyengine_uk_data/datasets/create_datasets.py

@@ -7,6 +7,15 @@
 
 logging.basicConfig(level=logging.INFO)
 
+# Local targets the calibration logs but does not train on. HMRC's counts of
+# income-tax payers with employment income by area (SPI table 3.15) are annual:
+# they include people with pay for part of the year whose FRS status at
+# interview is out of work, and who therefore have no pay in the FRS. Training
+# on them moves weight from people out of work to employees, away from the LFS
+# employee count (targets/sources/ons_labour_market.py). The area amounts of
+# employment income, and the national counts by income band, still train.
+VALIDATION_ONLY_LOCAL_TARGETS = ["hmrc/employment_income/count"]
+
 
 def _get_positive_int_env(name: str, default: int) -> int:
     raw_value = os.environ.get(name)
@@ -252,7 +261,7 @@ def main():
                 area_count=650,
                 weight_file="parliamentary_constituency_weights.h5",
                 dataset_key=str(frs_release.calibration_year),
-                excluded_training_targets=[],
+                excluded_training_targets=VALIDATION_ONLY_LOCAL_TARGETS,
                 log_csv="constituency_calibration_log.csv",
                 verbose=True,  # Enable nested progress display
                 area_name="Constituency",
@@ -279,7 +288,7 @@ def main():
                 area_count=360,
                 weight_file="local_authority_weights.h5",
                 dataset_key=str(frs_release.calibration_year),
-                excluded_training_targets=[],
+                excluded_training_targets=VALIDATION_ONLY_LOCAL_TARGETS,
                 log_csv="la_calibration_log.csv",
                 verbose=True,  # Enable nested progress display
                 area_name="Local Authority",