diff --git a/tools/import_validation/runner.py b/tools/import_validation/runner.py
index bcf83fb06e..c2f5ba271a 100644
--- a/tools/import_validation/runner.py
+++ b/tools/import_validation/runner.py
@@ -191,7 +191,8 @@ def _load_differ_df_from_mcf(self, input_dir: str) -> pd.DataFrame:
                     stats[current_var][diff_type] += 1
 
         if not stats:
-            return pd.DataFrame()
+            return pd.DataFrame(
+                columns=['StatVar', 'ADDED', 'DELETED', 'MODIFIED'])
 
         rows = []
         for var, counts in stats.items():
diff --git a/tools/statvar_importer/mcf_file_util.py b/tools/statvar_importer/mcf_file_util.py
index 7f6177c52c..8120b2c055 100644
--- a/tools/statvar_importer/mcf_file_util.py
+++ b/tools/statvar_importer/mcf_file_util.py
@@ -826,7 +826,7 @@ def normalize_list(value: str, sort: bool = True) -> str:
             value_list = get_value_list(value)
             has_quotes = True
         else:
-            value_list = value.split(',')
+            value_list = [v.strip() for v in value.split(',')]
         values = []
         if sort:
             value_list = sorted(value_list)
@@ -931,6 +931,10 @@ def normalize_value(
             if value[0] == '"' and value[-1] == '"' and len(value) > 100:
                 # Retain very long strings, such as geoJsonCoordinates, as is.
                 return value
+            if value.startswith('[') and value.endswith(']') and ',' in value:
+                inner_list = value[1:-1].strip()
+                normalized_list = normalize_list(inner_list)
+                return f'[{normalized_list}]'
             if ',' in value and maybe_list:
                 return normalize_list(value)
             if value[0] == '[':