It is a pleasure for us to comment on the paper by Ghosh, Hasking, and Natarajan. We would like to thank the editors for their invitation. This is a very interesting paper because it approaches the modeling and forecasting of the daily load curve taking into account the main cyclical features, special days and holiday adjustments, and temperature effects that affect the electricity demand, and at the same time, includes dynamic price-incentive effects. With the latter characteristic, the model turns to be potentially useful for operators in a smart grid. In fact, it seems to be the first paper using dynamic price-incentives in modeling the load curve, and we could expect that it will exert an influence on future research. This paper has addressed two important issues regarding the modeling and forecasting of daily load curves. On one hand, the paper is inspired by the approach in [1], in which the aim is to model the hourly consumption using a separate model for each hour of the day. This approach has the inconvenient of lacking a structure which provides smoothness and continuity between successive hours but has the advantage of allowing a dynamic and covariate structures that can be specific for the hour of the day if data require it. On the other hand, the paper allows for a smooth transition between the hours of the day, improving the model proposed by [2], by using B-splines basis (which are low-rank basis), instead of cubic splines (where the number of parameters is equal to the number points), in this way, the total number of parameters is also greatly reduced even when the time granularity of the data increases. In fact, the paper presents applications for hourly and 15-minute data formulating models which can account for varying covariate effects on load at different times of the day. The following comments can be seen as a way to obtain a better compromise between the flexibility provided when one model for each moment of the day is built, and the restrictiveness of the spline smooth approach presented in this paper.In load curve forecasting, an important issue is the forecasts of special days and holidays. The paper takes into consideration these days in a too simplistic way, by introducing a load curve for these days in the formulation of the coefficients of the basis functions. However, there is empirical evidence, see for instance [3] and [4], that the load curve changes along the year mainly in summer and Christmas vacation periods, and that there are multiplicative interaction effects between special days and holidays with temperature and the weekly cycle. The authors used hourly contemporaneous temperatures, but temperature information beyond the maximum and minimum temperatures of the day could be almost irrelevant for consumption decisions. Furthermore, the effect of temperature in daily (hourly) electricity consumption is dynamic [3] and [4], because it is affected by temperatures in both previous days and current one.In the model proposed, the key point is the relationship between the basis c...