This study examines whether the momentum phenomenon of the Korean stock market since the Asian financial crisis in the 90s can be explained by risk-based hypothesis or it is due to the delayed response induced by limit-to-arbitrage. The zero investment portfolios following Jegadeesh & Titman (1993) tradition do not enjoy the positive excess return after adjusting the risk factors from Fama-French (1993) methodologies. Furthermore there is no evidence that shows the momentum effect stems from the delayed response caused by market frictions. Limit-to-arbitrage proxy variables such as idiosyncratic volatilities, trading volume, illiquidity, recent price, number of zero daily returns, number of analysts following, cash flow volatility do not show any relationships with the momentum effect. Interestingly, the momentum effect is much more strongly pronounced for the stocks with high foreign investor ownership rate. If it is allowed to interpret the foreign investors are relatively more keen and rational to market information in the investment procedures compared to individual investors, this finding resonates with the previous rational expectation based explanations and the fact that the momentum strategies tend to show negative returns prior to 2000 when foreign investors ownership in the Korean stock market is institutionally-limited.