We investigate market-price convergence in the competitive Texas electricity market in the presence of large-scale wind generation using a large data sample of over 30 000 hourly observations for the period of December 1, 2010 to May 31, 2014. Hourly premiums vary by time of day and month. Simple univariate analysis suggests patterns related to the day of the week, although multivariate regression analysis reveals that this pattern is weak. The levels of the premiums are low and the forward premiums for any given hour exhibit serial correlation across days. An increase in wind generation tends to increase the premiums. This effect is significant for six of the twenty-four hours for non-West zones and half of the hours for the wind-rich West zone. The extent of the effect of rising wind generation on the West zone's premium is greater than its effect on premiums in other zones. However, an increase in wind generation tends to reduce the forward premium's volatility in nearly all hours. Taken together, these findings suggest that the Electric Reliability Council of Texas's day-ahead and real-time markets exhibit modest trading inefficiencies. However, making a sizable arbitrage profit on a consistent basis is difficult because of the unpredictable nature of wind generation. To be sure, accurate wind generation may arguably improve arbitrage profitability, especially for the West zone that contains most of Texas's wind farms. However, if the wind forecast accuracy also improves price convergence, the arbitrage profit diminishes as well.