diff --git a/src/mog_commons/__init__.py b/src/mog_commons/__init__.py
index 7525d19..66a87bb 100644
--- a/src/mog_commons/__init__.py
+++ b/src/mog_commons/__init__.py
@@ -1 +1 @@
-__version__ = '0.1.4'
+__version__ = '0.1.5'
diff --git a/src/mog_commons/string.py b/src/mog_commons/string.py
index 321c6a2..bf9e53c 100644
--- a/src/mog_commons/string.py
+++ b/src/mog_commons/string.py
@@ -131,10 +131,11 @@ def unicode_decode(data, encoding_list):
     assert encoding_list, 'encodings must not be empty.'
 
     xs = distinct(encoding_list if isinstance(encoding_list, list) else [encoding_list])
-    init, last = xs[:-1], xs[-1]
-    for encoding in init:
+    first_exp = None
+    for i, encoding in enumerate(xs):
         try:
             return data.decode(encoding)
-        except UnicodeDecodeError:
-            pass
-    return data.decode(last)
+        except UnicodeDecodeError as e:
+            if i == 0:
+                first_exp = e
+    raise first_exp
diff --git a/tests/mog_commons/test_string.py b/tests/mog_commons/test_string.py
index 399abc8..08366ba 100644
--- a/tests/mog_commons/test_string.py
+++ b/tests/mog_commons/test_string.py
@@ -106,5 +106,8 @@ def test_unicode_decode(self):
         self.assertEqual(string.unicode_decode('あいうえお'.encode('utf-8'), ['utf-8', 'utf-8', 'utf-8']), 'あいうえお')
         self.assertEqual(string.unicode_decode('あいうえお'.encode('sjis'), ['ascii', 'utf-8', 'sjis']), 'あいうえお')
         self.assertRaisesRegexp(
-            UnicodeDecodeError, "'shift_jis' codec can't decode",
+            UnicodeDecodeError, "'ascii' codec can't decode",
             string.unicode_decode, 'あいうえお'.encode('utf-8'), ['ascii', 'sjis'])
+        self.assertRaisesRegexp(
+            UnicodeDecodeError, "'shift_jis' codec can't decode",
+            string.unicode_decode, 'あいうえお'.encode('utf-8'), ['sjis', 'ascii'])