The come into effect of the Measuring Instruments Directive (MID) 2004/22/EC and the corresponding harmonized technical standards promotes the investigation of appropriate test and calibration procedures capable to assess the compliance of the measuring instruments for legal metrological control with the needs of the modern life. It is in this context that we focus our attention on the active (static) electrical energy meters (, annex MI-003) and in particular on their accuracy performance. The availability of a relatively simple model capable to predict the measurement error of an instrument as a function of its input quantities is essential for the design of the calibration plan and for the identification of possible critical aspects of the metrological confirmation process. It is indeed important to be aware that the adoption of different calibration plans might lead to different conclusions about the compliance of the same instrument with the error limits stated by the relevant standards. Although the model proposed relies on basic physical considerations about the typical architecture of static energy meters the scope here is not to identify the various sources of error and/or to quantify them starting from an a-priori analysis of the physical structure of these devices (as described, for example, in -). The identification of each source of error is a difficult task, especially considering that different non-ideal effects may lead to the same error contribution. We will therefore follow a quasi-black box and empirical approach leading to a simple model having few (three) parameters which will be quantified making use of the statistical analysis of series of measurements. Gain (multiplicative), phase and bias (additive) errors are considered. Gain and phase errors are associated to the gain and phase mismatch of the voltage and current channels of the energy meter, while the bias error is associated to a bias in both channels. Although measures are taken, in the energy meter architecture, in order to correct these errors the correction itself cannot be perfect and a residual error will unavoidably be present. All the parameters are assumed to be independent of load current. The plausibility of an additional contribution to gain error proportional to load current (originated from self-heating) is however tested through an appropriate statistical analysis. In the following we will refer to power meters (PMs) instead of energy meters because the principle of operation of these instruments is that of a PM and the conversion from power to energy is obtained through a count operation, which is intrinsically error free.