|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
About the Extel Financial database |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
What are the Extel Financials? The Extel Financials is a database containing fundamental accounting data for 26,500 world-wide companies with a history starting in 1985. The key feature that distinguishes this database from many others is that it is in a structured 'as reported' format. This is an important term and it will be explained in full below. The term as reported means that the data items in the database are the same as those reported in the annual report and accounts. This implies that there is a richness in the number and range of data items that means users can do just about anything, using the database, that could have done using the published accounts. In contrast, if a database is standardised it means that adjustments have been made to the annual report items before they are entered into the collection database by the editors. Some users prefer the detail of an as reported database. Usually these are fundamental analysts who use a great deal of the detail available in the annual accounts. Good examples are the Credit Analysts in banks, Equity Research Analysts or Strategic Consultants. Others prefer the standardised data set. Strictly, the Extel Financials is a structured as reported database because it structures the as reported data. In other words, it strikes a balance between providing enough data to satisfy sophisticated customers while ensuring that the data is still easy to use. If the database was truly as reported it would contain items for every different accounting description you could find in the annual report and accounts. This would probably number in excess of 300,000 items and would be impossible to use. The term as reported is used below instead of the more correct but cumbersome structured as reported . Detailed and logical pyramid structure The fact that the Extel Financials is an as reported database implies that it contains a large number of data items and a high level of detail. There are potentially 1,450 fields that can be populated. For example the total income or expense relating to exceptional items can be broken down into 14 different constituent parts. The whole P&L contains 210 fields for Commercial and Industrial companies. Working with a large number of data items is difficult and error prone. There are two main problems. The first is just remembering the names of the items. The second is knowing how the items relate to each other. The unique Company Analysis pyramid data structure provides an elegant solution to these problems. The data is organised as a series of pyramids that show the relationships between the data items. Each pyramid starts with a named grand total, such as total assets. The next level breaks this total into its constituents; for example total assets is broken into fixed assets and current assets. Subsequent levels break the components down into their components, and so on. Thus, at each level, the pyramid shows how the data items relate to those in the level below and the level above. For more detailed information about these pyramids click here or see the next section. In reality only a proportion of the 1,450 items will contain data for any one company. In the UK, where accounting standards require a high level of detail in the annual accounts, a large industrial company will have data for around 450 of these fields. There are two main reasons why the remaining 1,000 fields are not populated: • The Extel Financials is a global database. Given this, it must be able to cope with different accountancy standards around the world and a proportion of the fields will only be applicable to certain regions. For example, Untaxed reserves will be populated for Scandinavian companies where this item is common but not in the UK or the US. Goodwill written off reserves would be populated in the UK, France, Italy, Netherlands, Switzerland, Germany, Hong Kong and Singapore but not in other countries as this accounting treatment is not allowed. • There are different templates applicable to different types of companies. The Extel Financials adequate fields to correctly report data for commercial and industrial companies, banks, insurance companies, property companies and investment trusts. Each of these templates will have fields specific to the way that they disclose their information. For example, you will not see Sales Revenue for a bank, but you will see Interest Income. However, where comparable fields do exist, such as earnings or fixed assets, it is possible to compare and aggregate companies from different templates.
The file CAdata31 lists all the data items available in the Extel Financials database. It also reveals how the accounting data items relate to each other and what adds up to what. To follow these notes it would be helpful to have this file open in Excel. The accounting data items are organised into six sections and each section is further divided into sub-sections. The following table shows and outline of the sections and sub-sections.
In addition the database contains a business summary, sedol numbers, Sic information, key dates, address and contact details, key personnel and auditor name. The dictionary is organised as a series of pyramids that show the relationships between the data items. Each pyramid starts with a grade total, such as Total assets . The next level breaks this total into its constituents; for example Total assets , is broken into Fixed assets and Current assets . Subsequent levels break the components down into their components, and so on. Thus, at each level, the pyramid shows how data items relate to those in the level below and those in the level above. The quickest way to understand the pyramid is to open up CAdata27a.xls and have a look at the Balance Sheet Assets section. You will see how Total assets are made up, it is the sum of all the red items;
The items in green are footnote items and do not form part of the pyramid. They will be discussed in full below. The fixed assets total is made up of;
Next you can see how the Intangible fixed asset total is made up;
Goodwill also expands and this reveals the lowest level of the pyramid for this particular section; -
U&O stands for unidentified and other, it is a miscellaneous field. Data editors may have input data from the annual report and accounts into a miscellaneous field if there is no other appropriate description into which it will fit. Otherwise, companies do not always disclose enough detail to complete a pyramid level, so a balancing number has to be entered. This ensures that what is disclosed plus what is not disclosed adds up correctly. Throughout the database these balancing numbers are referred to as U&O. Pyramid items are sometimes sub-classified in alternative ways. For example, loans are sub-divided in three alternative ways: by maturity, by type of instrument and by type of collateral. When this happens, the pyramid will branch into alternative sub-levels. Look at Total liabilities and find the secton for Debt. If you expand debt you will see that there different coloured bands. The light yellow band shows the breakdown of debt by maturity, the light green by type of instrument and the blue by type of collateral. The dark green band represent footnotes. These allow useful components of the data items to be provided from a different perspective. For example Total Leases and HP provides additional information about Debt but does not form a part of the pyramid structure so it is treated as a footnote. In the dictionary every footnotes is linked to a specific data item, and is shown at the correct pyramid level. This makes it easier to see how footnotes relate to the main data items. Every data item in the dictionary has a descriptive name and a short name. The descriptive name is an everyday description of the item such as Total Assets or Fixed Assets. The short names are used in report definitions. They are unique and never more than 15 characters long. Commonly used items have very short names that are easy to remember; for example, the short name of Total assets is ta and of Fixed assets is fa . A complete list of descriptive names and short names are shown in the attached Excel file. Short names are like e-mail addresses, they are very precise and show the data item's exact position in the hierarchy. For example, the short name of Brands, patents etc is fa.i.br . The fa indicates that the item is a Fixed asset , the i that it is Intangible and the final br that it is Brands, patents etc. Components of short names, such as fa or br are used consistently throughout the dictionary so, with practice, it is easy to work with short names. Virtually all short names are abbreviated by dropping the first element of their path. For example, Brands, patents etc is part of Total assets ( ta ) so, strictly speaking, its full short name should be ta.fa.i.br . However, as it is so commonly understood that fixed assets are a component of total assets the initial ta is omitted. As explained above, every pyramid level contains a balancing U&O item for unclassifiable division of the item. For example, the fist subdivision of Debt ( dt ) is into short term and long term, that is dt.lt and dt.st . A different sub-division, based the type of instrument uses dt1 in the short names. Thus Bank loans becomes dt1.bnk and Financial leases becomes dt1.no1 . Notes to the data items have no at the end of their short name. When more than one note applies to the same data item, the subsequent notes are indicated with a number: no1, no2 etc. The previous part of the short name shows the item to which the note applies. Thus ivi.no is a note to Investment income ( ivi ) and ivi.no1 is a different note to the same item. All notes relate to a particular data item, so they are certain to double count items in the main pyramid breakdown. For example, Mortgage bank debt ( dt.no1 ) is debt issued by a mortgage bank. Since it must be either long or short term it will inevitably double-count dt.st or dt.lt . Also, it is likely to be a bank loan (although it might be a lease so it may double count dt1.bnk or dt1.ls). The pyramid structure of income Income statement disclosure varies enormously between countries and between different types of company. This makes a pyramid structure particularly valuable to ensure consistency. Unfortunately, income statement pyramids are less intuitive than balance sheet pyramids. The natural structure of income statement data resembles a cascade rather than a pyramid. It starts with sales; then moves down through a series of profit sub-totals until it reaches net income. However, in pyramid terms, net income is the grand total and must be the apex of the pyramid. Profit after tax resides on the second level with sales at the bottom. This is illustrated in CAdataA27, opening up the data structure from line 3 illustrates the point discussed above. Once the upside-down nature of the income pyramid is understood it presents few problems. However it may seem strange to find Net income at the top of the pyramid and Sales further down. The Extel Financials cover 26,500 companies from over 55 different countries with a back history of upto 20 years. This includes full quoted coverage for the UK, France, Netherlands and Switzerland in Europe, and Singapore, Hong Kong, Malaysia, Thailand and Australia in Asia Pacific. In regions where there is not 100% quoted coverage the majority of all the listed stocks in that country will be included. The structure of the data, the coverage and the detail are of no consequence unless the users believe that the data is of sufficient quality. Thomson Financial has an excellent reputation for quality data but elsewhere in the world this brand name is not so well known. The following points are relevant to any discussion of quality:
The database is updated via weekly CD ROM or by FTP. It is also possible to take a daily service by FTP. Typically, the annual report is available on the CD ROM is between 2 and 4 weeks after publication. |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
