diff --git a/lib/lexer.js b/lib/lexer.js
index 161ae53b..fab26120 100644
--- a/lib/lexer.js
+++ b/lib/lexer.js
@@ -471,17 +471,21 @@
             tokens.push(['IDENTIFIER', interp]);
             i += group.length - 1;
             pi = i + 1;
-          } else if (((expr = this.balanced_string(str.substring(i), ['${', '}']))) && expr.length > 3) {
-            inner = expr.substring(2, expr.length - 1);
-            nested = lexer.tokenize("(" + inner + ")", {
-              rewrite: false,
-              line: this.line
-            });
-            nested.pop();
+          } else if (((expr = this.balanced_string(str.substring(i), ['${', '}'])))) {
             if (pi < i) {
               tokens.push(['STRING', quote + (str.substring(pi, i)) + quote]);
             }
-            tokens.push(['TOKENS', nested]);
+            inner = expr.substring(2, expr.length - 1);
+            if (inner.length) {
+              nested = lexer.tokenize("(" + inner + ")", {
+                rewrite: false,
+                line: this.line
+              });
+              nested.pop();
+              tokens.push(['TOKENS', nested]);
+            } else {
+              tokens.push(['STRING', quote + quote]);
+            }
             i += expr.length - 1;
             pi = i + 1;
           }
diff --git a/src/lexer.coffee b/src/lexer.coffee
index 74c96d97..30ba0d9b 100644
--- a/src/lexer.coffee
+++ b/src/lexer.coffee
@@ -349,12 +349,15 @@ exports.Lexer: class Lexer
           tokens.push ['IDENTIFIER', interp]
           i += group.length - 1
           pi: i + 1
-        else if (expr: @balanced_string str.substring(i), ['${', '}']) and expr.length > 3
-          inner: expr.substring(2, expr.length - 1)
-          nested: lexer.tokenize "($inner)", {rewrite: no, line: @line}
-          nested.pop()
+        else if (expr: @balanced_string str.substring(i), ['${', '}'])
           tokens.push ['STRING', "$quote${ str.substring(pi, i) }$quote"] if pi < i
-          tokens.push ['TOKENS', nested]
+          inner: expr.substring(2, expr.length - 1)
+          if inner.length
+            nested: lexer.tokenize "($inner)", {rewrite: no, line: @line}
+            nested.pop()
+            tokens.push ['TOKENS', nested]
+          else
+            tokens.push ['STRING', "$quote$quote"]
           i += expr.length - 1
           pi: i + 1
         i += 1
diff --git a/test/test_string_interpolation.coffee b/test/test_string_interpolation.coffee
index 031ad543..5c454544 100644
--- a/test/test_string_interpolation.coffee
+++ b/test/test_string_interpolation.coffee
@@ -28,7 +28,8 @@ ok "Escaping \$last" is 'Escaping $last'
 ok "Escaping \${last}" is 'Escaping ${last}'
 
 ok "$$" is '$$'
-ok "${}" is '${}'
+ok "${}" is ''
+ok "${}A${} ${} ${}B${}" is 'A  B'
 ok "\\\\\$$" is '\\\\\$$'
 ok "\\\${}" is '\\${}'