householdbench/householdbench-full

HouseholdBench: full data

Code and rebuild instructions · Website and leaderboard · Test prompts scored in the paper · Paper

This dataset holds all 1,670,349 rows of the 17 tasks whose source data we may redistribute, out of the benchmark's 4,663,568 rows over 32 tasks. The other 15 tasks use IPUMS, PSID, Michigan Surveys of Consumers or ICPSR data, which we may not redistribute. To build them, download those files under your own access and run the pipeline in the code repository together with the raw files in this dataset.

Contents

Folder	What it holds	Who reads it	Size
data/	The prompts of each task as Parquet, split into train, validation, pre_cutoff_test and post_cutoff_test	load_dataset and the dataset viewer	0.24 GB
householdbench/	The release files: prompts (JSONL), task tables, split assignments and XGBoost reference predictions (CSV), with checksums in three manifest.csv files	The code repository's scripts	6.5 GB
raw/	The 94 raw source files the pipeline starts from (CEX, CPI, LAUS, BEA, FHFA, SCE, two replication archives and six release-date tables)	Only a rebuild from scratch	2.9 GB

data/ and householdbench/prompts/ hold the same records; householdbench/ adds the task tables, split files and reference predictions. raw/ is the input to the pipeline and householdbench/ is its output for these 17 tasks.

Tasks

Task	Topic	Survey	Mode	Target type	Rows	Where to get it
cons_cex_categories	Consumption and saving	CEX	Baseline	Numeric	500,000	This dataset
cons_cex_rebate01	Consumption and saving	CEX	Policy response	Numeric	9,168	This dataset
cons_cex_sspay	Consumption and saving	CEX	Policy response	Numeric	20,769	Rebuild with the code
cons_cex_stimulus08	Consumption and saving	CEX	Policy response	Numeric	17,940	This dataset
cons_cex_total	Consumption and saving	CEX	Baseline	Numeric	500,000	This dataset
cons_psid_jobloss	Consumption and saving	PSID	Policy response	Numeric	3,676	Rebuild with the code
cons_psid_wealth	Consumption and saving	PSID	Baseline	Numeric	74,578	Rebuild with the code
cons_sce_growth	Consumption and saving	SCE	Baseline	Numeric	28,969	This dataset
cons_sce_shock	Consumption and saving	SCE	Policy response	Numeric	20,169	This dataset
house_census_move	Housing and location	Census	Baseline	Categorical	500,000	Rebuild with the code
house_psid_owner	Housing and location	PSID	Baseline	Categorical	70,170	Rebuild with the code
house_sce_financing	Housing and location	SCE	Policy response	Numeric	995	This dataset
house_sce_lockin	Housing and location	SCE	Policy response	Probabilistic	734	This dataset
house_sce_move	Housing and location	SCE	Baseline	Probabilistic	148,995	This dataset
income_cps_displace	Income dynamics and household resources	CPS	Policy response	Numeric	22,553	Rebuild with the code
income_mich_finance	Income dynamics and household resources	Michigan	Baseline	Numeric	252,870	Rebuild with the code
income_psid_earnings	Income dynamics and household resources	PSID	Baseline	Numeric	76,938	Rebuild with the code
income_sce_growth	Income dynamics and household resources	SCE	Baseline	Numeric	179,318	This dataset
income_sce_policy	Income dynamics and household resources	SCE	Policy response	Categorical	19,254	This dataset
labor_cps_displace	Labor supply, job search, and retirement	CPS	Policy response	Categorical	53,787	Rebuild with the code
labor_cps_jobfind	Labor supply, job search, and retirement	CPS	Baseline	Categorical	505,774	Rebuild with the code
labor_cps_retire	Labor supply, job search, and retirement	CPS	Baseline	Categorical	528,680	Rebuild with the code
labor_cps_separation	Labor supply, job search, and retirement	CPS	Baseline	Categorical	656,473	Rebuild with the code
labor_cps_ui	Labor supply, job search, and retirement	CPS	Policy response	Categorical	158,989	Rebuild with the code
labor_psid_addedworker	Labor supply, job search, and retirement	PSID	Policy response	Numeric	1,117	Rebuild with the code
labor_sce_offer	Labor supply, job search, and retirement	SCE	Baseline	Categorical	3,623	This dataset
labor_sce_reswage	Labor supply, job search, and retirement	SCE	Policy response	Numeric	5,994	This dataset
labor_sce_risk	Labor supply, job search, and retirement	SCE	Baseline	Probabilistic	85,720	This dataset
labor_sce_search	Labor supply, job search, and retirement	SCE	Baseline	Probabilistic	2,636	This dataset
macro_mich_outlook	Macroeconomic expectations	Michigan	Baseline	Numeric	66,845	Rebuild with the code
macro_sce_revision	Macroeconomic expectations	SCE	Policy response	Numeric	7,100	This dataset
macro_sce_uncertainty	Macroeconomic expectations	SCE	Baseline	Probabilistic	139,734	This dataset
In this dataset (17 tasks)					1,670,349
Rebuild with the code (15 tasks)					2,993,219
Benchmark total (32 tasks)					4,663,568

Use

from datasets import load_dataset

ds = load_dataset("householdbench/householdbench-full", "cons_sce_growth")  # splits: train, validation, pre_cutoff_test, post_cutoff_test

The code repository explains the record fields, the splits, and how to download and use these files.

Known issues

• Two prompt templates contain wording errors that every evaluated model saw, so they are kept to preserve comparability: "about the same than" (SCE household-finance questions) and a doubled "Your household your household" (cons_sce_growth).
• macro_sce_revision record IDs in this release drop the __m01_YYYYMM__m12_YYYYMM suffix used during the paper's evaluation runs; the records are otherwise identical.

License

The source surveys keep their own terms; no single license covers the data. LICENSE.md lists each source's terms and the attribution notices you must keep when you use or share these files. The Survey of Consumer Expectations (SCE) license requires these notices:

Some survey questions were taken or adapted from the Survey of Consumer Expectations, © 2013-2019 Federal Reserve Bank of New York (FRBNY). The SCE questions are available without charge at https://www.newyorkfed.org/microeconomics/sce and may be used subject to license terms posted there. FRBNY did not participate in or endorse HouseholdBench, and FRBNY disclaims any responsibility or legal liability for the administration of the survey and the analysis and interpretation of data collected.

Source: Survey of Consumer Expectations, © 2013-2019 Federal Reserve Bank of New York (FRBNY). The SCE data are available without charge at https://www.newyorkfed.org/microeconomics/sce and may be used subject to license terms posted there. FRBNY disclaims any responsibility or legal liability for this analysis and interpretation of Survey of Consumer Expectations data.

Citation

@article{huang2026householdbench,
  title   = {{HouseholdBench}: Evaluating Large Language Models as Predictors of Household Economic Behavior},
  author  = {Huang, Jin and Ferreras Garrucho, Diego and Xie, Yutong and Yuan, Walter M. and Mei, Qiaozhu and Lian, Chen and Hazell, Jonathon},
  journal = {arXiv preprint arXiv:2610.07563},
  year    = {2026}
}