diff --git a/pandas/core/common.py b/pandas/core/common.py
index dc2ee31bbaf3d..a8d61d7ec0297 100644
--- a/pandas/core/common.py
+++ b/pandas/core/common.py
@@ -473,7 +473,7 @@ def _infer_dtype_from_scalar(val):
         dtype = np.dtype('M8[ns]')
 
     elif isinstance(val, (np.timedelta64, timedelta)):
-        val = tslib.convert_to_timedelta(val, 'ns')
+        val = lib.Timedelta(val).value
         dtype = np.dtype('m8[ns]')
 
     elif is_bool(val):
@@ -826,6 +826,7 @@ def trans(x):  # noqa
 
 def _maybe_convert_string_to_object(values):
     """
+
     Convert string-like and string-like array to convert object dtype.
     This is to avoid numpy to handle the array as str dtype.
     """
@@ -837,6 +838,20 @@ def _maybe_convert_string_to_object(values):
     return values
 
 
+def _maybe_convert_scalar(values):
+    """
+    Convert a scalar scalar to the appropriate dtype
+    This avoids numpy directly converting according to platform preferences
+    """
+    if lib.isscalar(values):
+        dtype, values = _infer_dtype_from_scalar(values)
+        try:
+            values = dtype(values)
+        except TypeError:
+            pass
+    return values
+
+
 def _lcd_dtypes(a_dtype, b_dtype):
     """ return the lcd dtype to hold these types """
 
diff --git a/pandas/core/internals.py b/pandas/core/internals.py
index c21e78f547988..463a2da529b5d 100644
--- a/pandas/core/internals.py
+++ b/pandas/core/internals.py
@@ -18,6 +18,7 @@
                                 is_datetime64_dtype, is_datetimetz, is_sparse,
                                 array_equivalent, _is_na_compat,
                                 _maybe_convert_string_to_object,
+                                _maybe_convert_scalar,
                                 is_categorical, is_datetimelike_v_numeric,
                                 is_numeric_v_string_like, is_internal_type)
 import pandas.core.algorithms as algos
@@ -1201,6 +1202,7 @@ def where(self, other, cond, align=True, raise_on_error=True,
                              "like")
 
         other = _maybe_convert_string_to_object(other)
+        other = _maybe_convert_scalar(other)
 
         # our where function
         def func(cond, values, other):
diff --git a/pandas/tests/test_common.py b/pandas/tests/test_common.py
index c77d71be7c9c9..905816081f0c5 100644
--- a/pandas/tests/test_common.py
+++ b/pandas/tests/test_common.py
@@ -770,6 +770,36 @@ def test_maybe_convert_string_to_array(self):
         tm.assert_numpy_array_equal(result, np.array(['x', 2], dtype=object))
         self.assertTrue(result.dtype == object)
 
+    def test_maybe_convert_scalar(self):
+
+        # pass thru
+        result = com._maybe_convert_scalar('x')
+        self.assertEqual(result, 'x')
+        result = com._maybe_convert_scalar(np.array([1]))
+        self.assertEqual(result, np.array([1]))
+
+        # leave scalar dtype
+        result = com._maybe_convert_scalar(np.int64(1))
+        self.assertEqual(result, np.int64(1))
+        result = com._maybe_convert_scalar(np.int32(1))
+        self.assertEqual(result, np.int32(1))
+        result = com._maybe_convert_scalar(np.float32(1))
+        self.assertEqual(result, np.float32(1))
+        result = com._maybe_convert_scalar(np.int64(1))
+        self.assertEqual(result, np.float64(1))
+
+        # coerce
+        result = com._maybe_convert_scalar(1)
+        self.assertEqual(result, np.int64(1))
+        result = com._maybe_convert_scalar(1.0)
+        self.assertEqual(result, np.float64(1))
+        result = com._maybe_convert_scalar(pd.Timestamp('20130101'))
+        self.assertEqual(result, pd.Timestamp('20130101').value)
+        result = com._maybe_convert_scalar(datetime(2013, 1, 1))
+        self.assertEqual(result, pd.Timestamp('20130101').value)
+        result = com._maybe_convert_scalar(pd.Timedelta('1 day 1 min'))
+        self.assertEqual(result, pd.Timedelta('1 day 1 min').value)
+
 
 class TestConvert(tm.TestCase):