In the Korean stock market, investors must wait, on average, for about 3 weeks after subscription day to trade initial public offering (IPO) stocks. At subscription day, the underwriting firms receive orders for the IPO stocks from various investor groups such as employees of the issuing firm, institutional and individual investors. This results in creating time lapse, or non-trading period between subscription and listing day in Korea. This is quite different from the IPO process in the U.S. Such non-trading period may be a risk factor on the IPO investments. Because investors cannot trade the IPO stocks during the non-trading period, they cannot gain the trading profits when positive information for an IPO stock is obtained and vice versa. This paper examines how the existence of non-trading period affects the IPO underpricing as a new risk factor that can increase the expected losses of the IPO investments in the Korean stock market. To answer this research question, we first define non-trading period effect as the upper bound of dis- count rate for the offering price or the expected losses that can occur during the non-trading period. We then measure an option value of the non-trading period effect with option pricing model and analyze the effect of the option values on the IPO underpricing. Longstaff (1995) provided an option pricing model to estimate losses by implementing restrictions on stock trade. Assuming a perfect timer who knows the optimal selling point of the IPO stocks, a variant of Longstaff model is applied to estimate the maximum expected losses resulting from the restriction on trading IPO stocks as option values. To estimate the option value, we should measure the volatility of each IPO stocks. But that is impossible because IPO stocks have no past return data. As a proxy for the volatility of an IPO stock, we used the average of volatilities of the matching firms. Matching firms, which had to be listed at least 3 years prior to the listing day of the concerned IPO stock, were selected within the same industry of the IPO stock. Using the data of 602 IPO stocks listed in Korea exchange (KRX) with KOSPI and KOSDAQ boards from February, 2000 to July, 2007, we first analyzed the degrees of IPO underpricing using various initial return measures. We found that the IPO underpricing in the Korean stock market was very high and sustainable. The average initial return at listing day (= closing price of listing day/offering price -1) was 57.60%. And the averages of the other 3 initial returns, the average of holding period returns (HPR) and cumulating abnormal returns (CAR) calculated for 20 trading days and HPR from listing day to first negative stock return day after the listing day, were higher than the average returns at the listing day. The amount of initial returns at listing day is more than double the amount of that in the U.S market, which was reported by Ritter and Welch (2002). We found the evidence that non-trading period effect is one of the important factors which determine the IPO underpricing. The option values for non-trading period effect were 11.98% of the offering price on average, which is about 21% of the initial returns at the listing day. And the IPO underpicing increases as the option value of the IPO stock increases. When the 602 IPO stocks were divided into five groups by their option values, the average initial returns of the highest group fell in the range of 25~49%, significantly higher than those of the lowest group, and the highest group always showed higher initial returns among subsamples before and after the abolition of the market making rule and among those of listed markets, KOSPI or KOSDAQ. These results are consistent with the results of the regression analysis. Various factors were controlled to evaluate the effect on the initial returns of IPO stocks. The factors considered were the change of the market making rule, market returns before and at the listing day, and other IPO-related characteristic variables -IPO amount, competitive rate of subscription, firm age, ROA, leverage ratio, etc. The option value for non-trading period effect was significantly and positively related with initial returns of IPO stocks in all models. However, the effects of the non-trading period on the IPO underpricing decreased after the abolition of the market making rule. Previous papers suggested that the expected losses of underwriting firms resulting from the IPO regulations such as compulsory market making rule or put-back option rule are the main cause for the IPO underpricing. Above and beyond this cause, this paper suggests that the expected losses of investors during non-trading period are another important cause for the IPO underpricing in the Korean stock market.