The consensus data in Valuatum system consists of two elements:
These two elements are combined in Valuatum consensus system so that consensus estimates are added on the top of complete research data produced by one analyst. Therefore consensus estimates replace all the data produced by one analyst where consensus data is available.
Consensus estimates are presented with their own color like in the picture below:
When you set your mouse over a colored number, you can see a tooltip that tells how many banks have been used for producing the consensus estimate.
Combining consensus estimates and basic research data means that almost all of the key figures are calculated from consensus estimates and are mostly based on them, although large amount of basic research data is also used in these calculations. Some examples may clarify this better:
P/E (Price/Earnings)
P/E figure consists of two elements: earnings per share and share price. Earnings per share is based on the consensus estimates in the Valuatum consensus data, and the share price is acquired every day from an external data provider. Thus, the P/E value of the Valuatum consensus data is directly based on the consensus and the actual price (the closing price of the previous day).
P/BV (Price/Book value)
The price in the P/BV value still is the closing price of the previous day, but BV i.e. substance (own capital's book value) is taken from the balance sheet, which is from the basic data of a single analyst. In practise, however, also this consensus estimate includes a dominating element. Let's take for example the book value of the current year: in the Valuatum model, it is calculated from the book value of the previous year, paid dividend, and the earnings of the current year. Two foremosts are historical figures, and thus every analyst that has carefully prepared his data has exactly the same figures.
In the Valuatum system, the correctness of the history figures is ensured by its users as well as the competition between analysts. The analysis is shared with a vast amount of users and with the company in question. Normally users and analysts are quick to point out mistakes and therefore there should not be any. In the other hand, the third component of the P/BV value, i.e., the earnings of the current year, is taken directly from the consensus figures. So none of the parameters are directly from a single analyst in the end. Hence, although the P/BV value isn't directly derived from the consensus estimates, it is still totally based on both the consensus and non-changing history figures.
To majority of other key financial ratios, for example return on capital (ROI-%, ROE-%) and valuation figures (EV/EBIT, EB/EBITDA, dividend yield, PEG etc.), applies the same as to the P/BV-ratio. These ratios contain many computational elements and thus data from the Valuatum system, produced by an individual analyst, is widely used. In spite of this, they practically never include discretionary elements (earnings forecasts) from any individual analyst. The ratios are based on historical realization data, objective figures (e.g. number of shares) and consensus estimates.
In some cases the ratios include estimates from the analyst. However, the effect of these estimates is very small. For example regarding the balance sheet, while calculating the return on investment (ROI-%) company’s invested capital is used in division of the (consensus) operating profit. Even though it is mainly derived from historical data and the consensus estimates, to what extent the Valuatum analyst considers the balance sheet to include non-interest bearing debt has some effect on the ratio. In these cases the estimates seldom differ much from the historical realization (proportion of non-interest bearing debt from the balance sheet) and even if the analyst would have forecasted a 30% change from the current level (rarely the case) the effect on the ratio itself, herein ROI-%, remains marginal.
The forecasts produced by an individual analyst, as mentioned above, are in practise produced and revised by several competing analysts. However, the figures are always taken from the model of the number one analyst of each company. Consequently, one given analyst is at all times responsible for each estimate.
Consensus estimates from outside never include divisional estimates, and sometimes even quarterly figures from some companies are not available. In a case like this, the figures in question are generated on basis of higher level consensus estimate. This means, that divisional figures are being calculated on the basis of consensus estimates of corporate level so that basic data's divisional level figures are being used in allocation to divisional level.
Thus, if estimations of operating profit in the basic data for three divisions of a single company are for example 100, 200, 300 (600 in total) and consensus estimate on corporate-level is 660, the estimation for divisional level will be generated: 110, 220, and 330 so that difference of 60 (10%) in basic estimate will be shifted on to divisional level also. Just as well, "allocating in relation to old estimations" -method is also used in generating results for quarter level, if the consensus estimates would for some reason not be available for certain companies.