diff --git a/doc/source/whatsnew/v0.20.0.txt b/doc/source/whatsnew/v0.20.0.txt
index 74fe7916523c5..493f50617aceb 100644
--- a/doc/source/whatsnew/v0.20.0.txt
+++ b/doc/source/whatsnew/v0.20.0.txt
@@ -1072,6 +1072,7 @@ Groupby/Resample/Rolling
 - Bug in ``.rolling()`` where ``pd.Timedelta`` or ``datetime.timedelta`` was not accepted as a ``window`` argument (:issue:`15440`)
 - Bug in ``Rolling.quantile`` function that caused a segmentation fault when called with a quantile value outside of the range [0, 1] (:issue:`15463`)
 - Bug in ``DataFrame.resample().median()`` if duplicate column names are present (:issue:`14233`)
+- Bug in ``.groupby()`` when calling ``first()`` or ``last()`` on TZ-aware timestamps (:issue:`15884`)
 
 Sparse
 ^^^^^^
diff --git a/pandas/core/groupby.py b/pandas/core/groupby.py
index fe764a099bb63..8091a976c07ee 100644
--- a/pandas/core/groupby.py
+++ b/pandas/core/groupby.py
@@ -1162,11 +1162,10 @@ def first_compat(x, axis=0):
 
             def first(x):
 
-                x = np.asarray(x)
                 x = x[notnull(x)]
                 if len(x) == 0:
                     return np.nan
-                return x[0]
+                return x.iloc[0]
 
             if isinstance(x, DataFrame):
                 return x.apply(first, axis=axis)
@@ -1177,11 +1176,10 @@ def last_compat(x, axis=0):
 
             def last(x):
 
-                x = np.asarray(x)
                 x = x[notnull(x)]
                 if len(x) == 0:
                     return np.nan
-                return x[-1]
+                return x.iloc[-1]
 
             if isinstance(x, DataFrame):
                 return x.apply(last, axis=axis)
diff --git a/pandas/tests/groupby/test_groupby.py b/pandas/tests/groupby/test_groupby.py
index c17c98c5448be..0582cb0b79903 100644
--- a/pandas/tests/groupby/test_groupby.py
+++ b/pandas/tests/groupby/test_groupby.py
@@ -81,6 +81,15 @@ def test_select_bad_cols(self):
             # will have to rethink regex if you change message!
             g[['A', 'C']]
 
+    def test_first_last_timestamp(self):
+        # GH15884
+        df = pd.DataFrame({'time': [pd.Timestamp('2012-01-01 13:00:00+00:00')],
+                           'A': [3]})
+        result = df.groupby('A', as_index=False).first()
+        assert_frame_equal(df, result)
+        result = df.groupby('A', as_index=False).last()
+        assert_frame_equal(df, result)
+
     def test_first_last_nth(self):
         # tests for first / last / nth
         grouped = self.df.groupby('A')