diff --git a/doc/source/whatsnew/v0.22.0.txt b/doc/source/whatsnew/v0.22.0.txt
index d43d5bec7175f..150e86f564ad0 100644
--- a/doc/source/whatsnew/v0.22.0.txt
+++ b/doc/source/whatsnew/v0.22.0.txt
@@ -81,6 +81,7 @@ Other API Changes
 - Inserting missing values into indexes will work for all types of indexes and automatically insert the correct type of missing value (``NaN``, ``NaT``, etc.) regardless of the type passed in (:issue:`18295`)
 - Restricted ``DateOffset`` keyword arguments. Previously, ``DateOffset`` subclasses allowed arbitrary keyword arguments which could lead to unexpected behavior. Now, only valid arguments will be accepted. (:issue:`17176`, :issue:`18226`).
 - :func:`DataFrame.from_items` provides a more informative error message when passed scalar values (:issue:`17312`)
+- When created with duplicate labels, ``MultiIndex`` now raises a ``ValueError``. (:issue:`17464`)
 
 .. _whatsnew_0220.deprecations:
 
diff --git a/pandas/core/indexes/multi.py b/pandas/core/indexes/multi.py
index 456999b94c523..761dbc6086b53 100644
--- a/pandas/core/indexes/multi.py
+++ b/pandas/core/indexes/multi.py
@@ -177,8 +177,8 @@ def _verify_integrity(self, labels=None, levels=None):
         Raises
         ------
         ValueError
-            * if length of levels and labels don't match or any label would
-            exceed level bounds
+            If length of levels and labels don't match, if any label would
+            exceed level bounds, or there are any duplicate levels.
         """
         # NOTE: Currently does not check, among other things, that cached
         # nlevels matches nor that sortorder matches actually sortorder.
@@ -198,6 +198,11 @@ def _verify_integrity(self, labels=None, levels=None):
                                  " level  (%d). NOTE: this index is in an"
                                  " inconsistent state" % (i, label.max(),
                                                           len(level)))
+            if not level.is_unique:
+                raise ValueError("Level values must be unique: {values} on "
+                                 "level {level}".format(
+                                     values=[value for value in level],
+                                     level=i))
 
     @property
     def levels(self):
diff --git a/pandas/tests/groupby/test_functional.py b/pandas/tests/groupby/test_functional.py
index bc13d51c4f4f6..b9718663570bd 100644
--- a/pandas/tests/groupby/test_functional.py
+++ b/pandas/tests/groupby/test_functional.py
@@ -52,10 +52,10 @@ def test_frame_describe_multikey(self):
         desc_groups = []
         for col in self.tsframe:
             group = grouped[col].describe()
-            group_col = pd.MultiIndex([[col] * len(group.columns),
-                                       group.columns],
-                                      [[0] * len(group.columns),
-                                       range(len(group.columns))])
+            # GH 17464 - Remove duplicate MultiIndex levels
+            group_col = pd.MultiIndex(
+                levels=[[col], group.columns],
+                labels=[[0] * len(group.columns), range(len(group.columns))])
             group = pd.DataFrame(group.values,
                                  columns=group_col,
                                  index=group.index)
@@ -67,8 +67,9 @@ def test_frame_describe_multikey(self):
                                          'C': 1, 'D': 1}, axis=1)
         result = groupedT.describe()
         expected = self.tsframe.describe().T
-        expected.index = pd.MultiIndex([[0, 0, 1, 1], expected.index],
-                                       [range(4), range(len(expected.index))])
+        expected.index = pd.MultiIndex(
+            levels=[[0, 1], expected.index],
+            labels=[[0, 0, 1, 1], range(len(expected.index))])
         tm.assert_frame_equal(result, expected)
 
     def test_frame_describe_tupleindex(self):
diff --git a/pandas/tests/indexes/test_multi.py b/pandas/tests/indexes/test_multi.py
index 5c2a0254b072b..a2c0a75e21f43 100644
--- a/pandas/tests/indexes/test_multi.py
+++ b/pandas/tests/indexes/test_multi.py
@@ -1618,7 +1618,9 @@ def test_is_(self):
         # shouldn't change
         assert mi2.is_(mi)
         mi4 = mi3.view()
-        mi4.set_levels([[1 for _ in range(10)], lrange(10)], inplace=True)
+
+        # GH 17464 - Remove duplicate MultiIndex levels
+        mi4.set_levels([lrange(10), lrange(10)], inplace=True)
         assert not mi4.is_(mi3)
         mi5 = mi.view()
         mi5.set_levels(mi5.levels, inplace=True)
@@ -2450,13 +2452,11 @@ def test_isna_behavior(self):
             pd.isna(self.index)
 
     def test_level_setting_resets_attributes(self):
-        ind = MultiIndex.from_arrays([
+        ind = pd.MultiIndex.from_arrays([
             ['A', 'A', 'B', 'B', 'B'], [1, 2, 1, 2, 3]
         ])
         assert ind.is_monotonic
-        ind.set_levels([['A', 'B', 'A', 'A', 'B'], [2, 1, 3, -2, 5]],
-                       inplace=True)
-
+        ind.set_levels([['A', 'B'], [1, 3, 2]], inplace=True)
         # if this fails, probably didn't reset the cache correctly.
         assert not ind.is_monotonic
 
@@ -3083,3 +3083,16 @@ def test_million_record_attribute_error(self):
         with tm.assert_raises_regex(AttributeError,
                                     "'Series' object has no attribute 'foo'"):
             df['a'].foo()
+
+    def test_duplicate_multiindex_labels(self):
+        # GH 17464
+        # Make sure that a MultiIndex with duplicate levels throws a ValueError
+        with pytest.raises(ValueError):
+            ind = pd.MultiIndex([['A'] * 10, range(10)], [[0] * 10, range(10)])
+
+        # And that using set_levels with duplicate levels fails
+        ind = MultiIndex.from_arrays([['A', 'A', 'B', 'B', 'B'],
+                                      [1, 2, 1, 2, 3]])
+        with pytest.raises(ValueError):
+            ind.set_levels([['A', 'B', 'A', 'A', 'B'], [2, 1, 3, -2, 5]],
+                           inplace=True)