diff --git a/doc/source/whatsnew/v1.5.0.rst b/doc/source/whatsnew/v1.5.0.rst
index 55bfb044fb31d..47b673ab4a663 100644
--- a/doc/source/whatsnew/v1.5.0.rst
+++ b/doc/source/whatsnew/v1.5.0.rst
@@ -850,6 +850,7 @@ I/O
 - Bug in :func:`read_sas` returned ``None`` rather than an empty DataFrame for SAS7BDAT files with zero rows (:issue:`18198`)
 - Bug in :class:`StataWriter` where value labels were always written with default encoding (:issue:`46750`)
 - Bug in :class:`StataWriterUTF8` where some valid characters were removed from variable names (:issue:`47276`)
+- Bug affecting :func:`read_csv`, :func:`read_excel`, :func:`read_html`, and :func:`read_table`, which accepted decreasing multi-index header arguments even though they don't work (:issue:`47011`)
 
 Period
 ^^^^^^
diff --git a/pandas/io/common.py b/pandas/io/common.py
index 5aecc55bb363a..56bc344d47c40 100644
--- a/pandas/io/common.py
+++ b/pandas/io/common.py
@@ -198,6 +198,8 @@ def validate_header_arg(header: object) -> None:
             raise ValueError("header must be integer or list of integers")
         if any(i < 0 for i in header):
             raise ValueError("cannot specify multi-index header with negative integers")
+        if list(header) != sorted(set(header)):
+            raise ValueError("multi-index header elements must be increasing")
         return
     if is_bool(header):
         raise TypeError(
diff --git a/pandas/tests/io/parser/test_header.py b/pandas/tests/io/parser/test_header.py
index 3fc23525df89e..46e53f11ff722 100644
--- a/pandas/tests/io/parser/test_header.py
+++ b/pandas/tests/io/parser/test_header.py
@@ -61,6 +61,20 @@ def test_negative_multi_index_header(all_parsers, header):
         parser.read_csv(StringIO(data), header=header)
 
 
+@pytest.mark.parametrize("header", [([0, 0]), ([1, 0])])
+def test_nonincreasing_multi_index_header(all_parsers, header):
+    # see gh-47011
+    parser = all_parsers
+    data = """1,2,3,4,5
+        6,7,8,9,10
+        11,12,13,14,15
+        """
+    with pytest.raises(
+        ValueError, match="multi-index header elements must be increasing"
+    ):
+        parser.read_csv(StringIO(data), header=header)
+
+
 @pytest.mark.parametrize("header", [True, False])
 def test_bool_header_arg(all_parsers, header):
     # see gh-6114