We welcome the efforts undertaken by the writing team in order to prepare this chapter about data sources and collection strategies.

During OG4 UNSD presented the annotated outline of this chapter under the title “Data sources and data compilation strategies”.

However, when reading this draft, it is obvious that the chapter is about data collection in addition to data sources and data compilation. We suggest having a discussion on the title when it is clearer what is going into the chapter.

This chapter must bee seen especially in relation to chapter 6. In chapter 6 tables and reference lists for what kind of data items to collect from statistical units are defined. But, how to collect these data items is however not described in the draft chapter 6. We suggest that data collection strategies should be described in chapter 6, or it should come first in chapter 7.

The relationship between data collection and data compilation as wee see it is:
Statistical units → data collection → data sources → data compilation → basic energy statistics.

The aim of data collection is to get the data items described in Chapter 6. These data items can be collected using data collection strategies such as census or sample surveys. The data sets thus obtained constitute data sources that the energy statistician can use to compile energy statistics. Data can also be retrieved from existing administrative registers. In the data compilation process data is taken (retrieved, but not collected) from the data sources, and weighted, imputed, adjusted, balanced etc using the recommended data compilation methods described in this chapter and put together to obtain basic energy statistics.

1.. Substantial parts of this chapter deal with statistics production in general. I can imagine that are existing (UN-)documents on professional production of statistics in general (?). This chapter could refer to these type of documents brieflly and then concentrate on the situation on energy statistics (more specific: the data items identified in chapter 6). This could make the chapter shorter and more transparant.

2.. Administrative data sources
In 24 'long-term accessibility' is mentioned as one of the advantages.
Our experience in the Netherlands shows this is not always the case. Adminstrative data (registrations) exist by the sake of some kind of legal procedure. If this legal procedure changes the corresponding registration may dissapear at all! And then you have nothing and very little influence in the decision to stop the registration (not meant for statistical purposes)!

Concerning chapter 7 from my point of view it should be clearly pointed out that it is practically impossible to get reliable energy data from non energy specific surveys only. The can be used as hints or additional information for understanding trends but clearly not as exclusive data source for energy statistics. So for instance we conduct two (very comprehensive) energy consumption surveys in the household sector to get nearly the data mentioned in
(a)Household surveys and mixed household-enterprise surveys and their roles in energy statistics (para 9 to 13) and the listening in these paragraphs gives the impression that it can be obtained from general household surveys e.g. household expenditure survey.