Archive:Inflation – methodology of the euro area flash estimate

Data from August 2011-August 2012, most recent data: Further Eurostat information, Main tables and Database.

Changes in the price of consumer goods and services are usually referred to as the inflation rate. Such changes measure the loss of living standards and are one of the most well-known economic statistics. This article describes the methodology used in the euro area inflation flash estimates which are usually released at the end of the reference month.

Table 1: Flash estimates and HICP annual rates for the all-items and main components - September 2011-September 2012

Accuracy of the flash estimates

Figure 1: Mean absolute deviation measured in percentage points, September 2011-September 2012

The aim of the inflation flash estimates is to predict as accurately as possible the actual inflation released later. Since accuracy is the degree of closeness of measurements of a quantity to that quantity's actual (true) value, the flash estimate is considered to be fully accurate if its value, rounded to one decimal place, is the same as the HICP annual rate, rounded to one decimal, released two weeks later. Table 1 puts side by side the flash estimates and the HICP annual rates released two weeks later. Over the last 12 months, the maximum difference between the flash estimate and the HICP annual rate was 0.4 (recorded in November 2011 for the energy component and in September 2012 for the non-energy industrial goods component).

Another way to measure accuracy is by the mean absolute deviation. The mean absolute deviation is the average of the absolute differences between the flash estimate and the HICP annual rate over time. Figure 1 shows the mean absolute deviation over the last 12 months for the all-items figure and the four main components.

The energy component has recorded the highest mean absolute deviation (around two times the food and services mean absolute deviation) over the last 12 months.

A third way to measure the accuracy is by evaluating its usefulness in predicting the direction of inflation. The flash estimate correctly predicts the direction of inflation if the difference between the flash estimate and the previous month inflation has the same sign as the difference between the actual inflation and the previous month inflation which is known two weeks later. For example, if the flash estimate is pointing to an increase on inflation and that increased is confirmed two weeks later, then the direction of inflation was accurately predicted.

There are three possible outcomes for the direction comparison:

the direction is accurately predicted in case the predicted increase / decrease / no change in inflation is confirmed (which will be denoted by );

the flash wrongly predicts the direction of inflation, e.g., predicting an increase when in fact there is a decrease instead (which will be denoted by );

it is not able to spot a change when there is a slight change and vice versa (which will be denoted by ).

Over the last 12 months the flash estimate accurately predicted the inflation's direction in 43 out of 60 estimates, and in only 3 out of 60 estimates the direction was wrongly predicted.

Figure 3: Inflation's direction analysis - October 2011-September 2012

Note: the first flash estimate breakdown into four main components was published in September 2012. Prior to that, flash estimate figures for food, non-energy industrial goods, energy and services were computed to test the accuracy of the algorithm in predicting their respective inflation rates, but only the all-item estimate was published at the end of each reference month.

Computation of flash estimates

The flash estimate breakdown procedure uses early information from the euro area countries supplemented with an econometric one-step ahead forecasting model for missing country data that combines timely energy price information with HICP back series and the HICP data of other countries available for the target reference month. The model estimates national HICPs and their main components (if not provided by a Member State) and aggregates them to the euro area level, together with the available flash estimates provided by the countries, to produce the flash estimates for the reference month.

There are five main steps that are processed in a sequential order:

Defining model specifications;
Processing energy prices data;
Processing available data;
Nowcast;
Calibration.

STEP 1: Defining model specifications

In this step, the model specifications are checked in order to find out if there is a model that performs better than the one used in the last round of flash estimate production. If a better model is found, it is then used in Step 4.

STEP 2: Processing energy prices data

For improving the prediction power of the model, the HICP data are combined with timely price data on fuels and heat energy. The information is received from DG ENER weekly; they publish prices of diesel, petrol and heat energy in their Weekly Oil Bulletin. At the time of the flash release, the energy price data of the last week of the month are normally missing; thus an automatic forecast for the missing week is made for each euro area country, and the monthly averages calculated from these weekly price data are used in the nowcast.

STEP 3: Processing available data

In this step, countries are separated into two groups: group 'A' countries which had provided preliminary data for the current month and group 'N' countries which had not. For each group, the available data for HICP and energy prices are aggregated into a single index; i.e., for the all-items aggregate, one index is created for group 'N' and one for group 'A'; for 'food', one index is created for group 'N' and another for group 'A', and so on. The HICP index for group 'A' has an extra observation (i.e. the preliminary data for the month in question) compared to than the index series for group 'N'.

STEP 4: Nowcast

This is the main step of the flash estimate procedure in which the 1-step ahead forecast and aggregation are done. For each of the euro area sub-component indices, an aggregate of countries 'N', i.e. without preliminary data, is forecasted using the model specified in Step 1 and the energy prices data and / or available preliminary data provided by other countries processed in Step 2. These 1-step forecast indices are aggregated with the aggregate of 'A' countries preliminary data, producing the (non-calibrated) nowcast for each sub-component of the flash estimate.

STEP 5: Calibration

The aggregation of the four sub-components 'food', 'non-energy industrial goods', 'energy' and 'services' has to correspond to the 'all-items' index. Since the nowcasts are calculated independently for each individual special aggregate, it is almost certain that their aggregation is not equal to the all-items nowcast. In Step 5, the 'un-balanced' component nowcasts are calibrated to fulfil the consistency requirement.

Figure 4: Flash estimate procedure

Models used

Prior to any model fitting, the first differences of the log are taken:

where is the aggregated index of the set of countries that have not provided any preliminary data. This will be the dependent variable to be forecasted;

where is the aggregated index of the set of countries that have provided preliminary data;

where is the weighted average of 'diesel' prices of the set of countries that have not provided preliminary data;

where is the weighted average of 'petrol' prices of the set of countries that have not provided preliminary data;

where is the weighted average of 'heat' prices of the set of countries that have not provided preliminary data.

For each aggregate that needs to be forecasted there are four possible regressors. Depending on the special aggregate, some regressors might not be significantly correlated with the dependent variable, so they are not likely to be used in the forecasting model.

Nowcasting the all-items aggregate

The model used to fit the 'all-items' is a regression model with SARIMA errors, which is given by:

The parameters for the SARIMA were previously determined in Step 1 of the flash estimate procedure and they are described in the model specifications file.

The forecast is given by:

where is a forecast from the SARIMA model shown above. The final step is to produce the nowcast index for the all-items, by aggregating with .

Nowcasting the main components

Although the number of main components that will be published is four, i.e., 'food' (FOOD), 'non-energy industrial goods' (IGOODSXE), 'energy' (ENRGY) and 'services' (SERV), in the estimation ENERGY is split into ELGAS, which is the aggregation of electricity, gas, solid fuels and heat energy, and FUELS, which is the aggregation of liquid fuels and fuels & lubricants for personal transport equipment. The reason for the split is that it results in improved final accuracy. Since the two sub-components have significantly different dynamics, the indirect forecast (forecasting first the sub-components and aggregating the forecasts later on) performs on average better than the direct forecast (forecasting 'energy' directly).

Calibrating the main components

As final step in the estimation procedure, the main components nowcast are calibrated in such a way that their aggregation equals the nowcast for the all-items, making the five nowcast consistent. The calibration factor is given by: