Results are disseminated, adjusted for seasonality and calendar effects, of the main aggregates of the national economy and of the rest of the world and of the main balances and operations of the Non-financial Corporations, General Government and Households sectors.
The preparation of the results, adjusted for seasonality and calendar effects, include both the signal extraction process, which includes the seasonal and calendar adjustment process, as well as the procedures that guarantee the necessary annual consistency between the raw data and the non-seasonal data and between operations and the balances of the accounts system.
The procedures applied are those described in the general methodology of the series of main aggregates of the Spanish Quarterly National Accounts (SQNA). These therefore conform to the recommendations published in the manuals of quarterly accounts (Eurostat, 1999/2013 and the International Monetary Fund 2001/2016) and of seasonal adjustment (ESS guidelines on seasonal adjustment, 2015 and with the INE Standard for the correction of seasonal and calendar effects in short-term series, 2013).
In this way, the seasonal adjustment procedures are implemented with a parametric focus, based on regression models with ARIMA errors, identifying and evaluating a model a priori that adequately adjusts to the observed series and deriving appropriate models from this for each of the series components (cycle-trend, seasonal and irregular). The latest version of the TRAMO-SEATS software is used (Gomez and Maravall 1994).
It should be taken into account that the ARIMA model is chosen once a year when the annual series are also revised. Such models remain fixed during the rest of the year. However, these are monitored in every quarter. In addition, the parameters of the ARIMA model are recalculated whenever a new observation is available.
The consistency between raw data and seasonally adjusted data is also maintained, that is, the annual total of the seasonally adjusted series coincides with the annual total of the original series. This preference for consistency in optimality in the sense of seasonal adjustment must be understood in the context of the coherence which should exist in all national accounts systems, which is indispensable for the analysis of both short- and long-term economic developments. For this, use is made of methods based on models for the disaggregation of time series and the adjustment of total quarterly data.
On the other hand, some of these CNTFSI results series are arithmetic balances of others. If these balance series are directly adjusted for seasonal and calendar effects, the result would not be identical to the balance obtained through the seasonally adjusted elements, mainly due to their differing seasonal pattern and to the number of non-lineal operations involved in the process of adjustment for seasonal effects. Given the need for coherence in the balances with the previously seasonally adjusted series, an indirect estimate of these series is performed as a result of the operations or flows involved.
Finally, an exhaustive check is performed to verify that there is no residual seasonality in the seasonally adjusted series, a check that is especially important as the indirect method applies to some series, as indicated previously.
Ajuste estacional en CTNFSI_en.xlsx