[bashtage]

cov_cluster computes the clustered score covariance of the system, IV and panel models. It accepted two kinds of cluster ids that give a meaningless result:

• Missing values. np.unique counts all the missing values as one group, but the group boundaries found from the sorted ids treat each missing value as a group of its own. The loop over the groups then added the scores of the first observation with a missing cluster and left out the scores of all the others, without an error. In an example with 60 observations and 5 missing ids, the covariance was exactly that of the clusters of the 55 observations with an id plus the first of the other five as a cluster of its own, which is neither a covariance that drops the five observations nor one that treats each of them as a cluster.
• A single group. The scores sum to zero at the estimates, so the covariance is zero up to rounding. SUR with a single cluster reported standard errors of 2e-16 and PooledOLS 9e-17, while SUR with group_debias=True and the IV models raised a bare ZeroDivisionError.

cov_cluster now raises a ValueError for both. The ids cannot be dropped with the rows of the data that have missing values, since the model was estimated with them, so an error is the only result that is not silently wrong. As the check is in the shared function, every model that clusters inherits it. No test relied on either input.

The tests call cov_cluster directly, with float and object ids, and the system covariance estimators with one-way and two-way clusters. They fail without the change.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 99.59%. Comparing base (6cbd30b) to head (ec8d1ce).
⚠️ Report is 67 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #789      +/-   ##
==========================================
+ Coverage   99.58%   99.59%   +0.01%     
==========================================
  Files         128      128              
  Lines       24188    24473     +285     
  Branches     1809     1823      +14     
==========================================
+ Hits        24087    24374     +287     
+ Misses         40       39       -1     
+ Partials       61       60       -1     

Flag	Coverage Δ
adder	99.57% <100.00%> (+0.01%)	⬆️
subtractor	99.57% <100.00%> (+0.01%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.