diff --git a/pandas/core/arrays/interval.py b/pandas/core/arrays/interval.py
index ebabc7edcbf43..1011381f235ca 100644
--- a/pandas/core/arrays/interval.py
+++ b/pandas/core/arrays/interval.py
@@ -812,9 +812,7 @@ def take(self, indices, allow_fill=False, fill_value=None, axis=None, **kwargs):
 
         fill_left = fill_right = fill_value
         if allow_fill:
-            if (np.asarray(indices) == -1).any():
-                # We have excel tests that pass fill_value=True, xref GH#36466
-                fill_left, fill_right = self._validate_fill_value(fill_value)
+            fill_left, fill_right = self._validate_fill_value(fill_value)
 
         left_take = take(
             self.left, indices, allow_fill=allow_fill, fill_value=fill_left
diff --git a/pandas/io/formats/excel.py b/pandas/io/formats/excel.py
index cc7b6b0bfea97..0140804e8c7b5 100644
--- a/pandas/io/formats/excel.py
+++ b/pandas/io/formats/excel.py
@@ -629,7 +629,9 @@ def _format_hierarchical_rows(self):
                 ):
 
                     values = levels.take(
-                        level_codes, allow_fill=levels._can_hold_na, fill_value=True
+                        level_codes,
+                        allow_fill=levels._can_hold_na,
+                        fill_value=levels._na_value,
                     )
 
                     for i in spans: