The weak efficient market hypothesis can be interpreted as asserting that major market indices should lie near the Markowitz efficient frontier.  This is seen in many years, but not in years before large market downturns.  Most notably, it was not seen in most of 2007 and 2008 leading up the the crash of 2008.  Indeed, all the major indices were found on the Markowitz inefficient frontier right after the crash. More generally, we consider a frontier computed for Markowitz portfolios that hold only long positions, which lies to the right of the classical Markowitz frontier. For many years this so-called long frontier lies close to the Markowitz frontier, but not in years of market volatility. The question will be posed, but not answered, if this separation is a measure of a market exposed to large short positions that could contribute to a systemic downturn.