Load forecasting has been considered a powerful tool in managing and planning power systems. Several tecniques have been recently suggested for short-term load forecasting by a large number of researchers. This work studies the applicability of linear models in the area is intended to be a basis for a real forecasting application. The models were developed and tested on the real load data of a utility company located in the southeast of Brazil. All models are proposed for sectional data, that is, each hour's load is studied separately as a single series. This approach avoids modeling the intricate intra-day pattern (load profile) displayed by the load, wich varies throughout days of the week and seasons. Three models are studied, the first one a Dummy-Adjusted Seasonal Integrated Autoregressive Moving Average model - DASARIMA, acting as a benchmark, the second a two-step modeling that makes use of deterministic components to model trend, seasonality and calendar effects, called Two-Level Seasonal Autoregressive model - TLSAR; and the last one a Dummy-Adjusted Generalized Long Memory model - DAGLM. The test results showed that the hourly models are well suitable for forecasting application. The forecasting errors of the last two approaches were smaller than those of the DASARIMA benchmark. The work suggests that this kind of hourly models should be implemented in a through on-line testing in order to provide a final opinion on its applicability.