Task Force on the Valuation and Measurement of Equity –Unquoted shares

- Note by Canada (document for agenda item 4 b-c-d)

Introduction

1. The first TFVME meeting was, to a large extent, taken up with valuation and measurement of unquoted shares. In particular country experiences were compared and discussed, with presentations by France, Spain, United States and Canada. The format of the first meeting was more that of a workshop, where cross-country developments and issues were discussed in an open form. While this was a useful first step that provided considerable background information, the TFVME must now focus on outcomes at this second and final meeting.

2. One observation that can be made is that country experiences are varied. The question then becomes how best to reflect this in SNA93, while respecting important work undertaken under by WGUS and as part of the forthcoming revisions to BPM.

General Approach

3. On the matter of unquoted shares, SNA93 was quite clear but also quite limited:

Shares and other equities should be valued in the balance sheets at their current prices when they are regularly traded on stock exchanges or other organized financial markets. The value of shares in corporations that are not quoted on stock exchanges or otherwise traded regularly should be estimated using the prices of quoted shares that are comparable in earnings and dividend history and prospects, adjusting downward, if necessary, to allow for the inferior marketability or liquidity of unquoted shares. Equity in quasi-corporations should be valued as equal to the value of the quasi-corporations’ assets less the value of their liabilities. (13.74)

4.SNA93 is first and foremost a conceptual document. It is clear that market valuation of assets and liabilities – including equity – is the principle to which all countries should adhere. SNA93 does an awkward job of linking the market value of corporate assets and liabilities to corporate equity / net worth, largely side-stepping the issue. However, it then becomes very prescriptive on the matter of valuing unquoted equity (13.74). The issue is whether this prescriptive approach should be retained in an SNA93 re-draft. The AEG has endorsed that “… outputs from the TFVME that are potentially going to be used in an updated SNA should focus on the underlying concepts and not be too prescriptive” (TFVME agenda item 2b paragraph #460).

5. The current SNA93 prescriptive approach to valuing unquoted shares is largely mirrored in the WGUS work and recommendations. This suggests that it would be widely applied in certain jurisdictions, in particular in Western Europe and selected other developed economies. WGUS recommendations will likely find itself in a re-draft of ESA95. Looking further, comparability will be enhanced by adopting as common a method as possible internationally.

6. On the other hand, there are a couple of disadvantages to retaining a fixed approach to valuing unquoted shares in SNA93. First, some countries may not be able to follow an approach that values unquoted equity with reference to quoted equity (CMV), largely because the stock market is very thin or non-existent. This is an important consideration in an international standard, and would argue for a more flexible approach. Second, more flexibility in the SNA93 re-draft would allow for harmonization of the emerging standards in BPM, which (at this point) has not adopted one single recommendation but rather a host of options.

7. On the issue of comparability, it could be argued that international comparability in many areas of the SNA, and in particular with reference to valuing unquoted equity, is not an achievable goal except across groups of similar economies. In this respect, flexibility in adhering to the SNA93principle of market value may be more important than achieving comparability in methods recommended. Adherence to standards may be more thoroughly achieved across by incorporating some degree of flexibility in recommendations.

8. It is proposed that the TFVME adopt the AEG recommendation (clear on the concept, less prescriptive on the method) and consider a less prescriptive approach for unquoted equity, specifically:

-That the current method of valuing unquoted equity with reference to quoted equity be retained at the (i) main approach or (ii) widely-accepted approach in SNA93; but, that

-That other (second-best) methods be considered in SNA93 in those cases where the preferred approach is not feasible.

Methods

9. As was clearly evident at the first TFVME meeting, country experiences are varied with respect to equity measurement and valuation, and can also differ between balance sheet accounts data and the international investment position (IIP) data. In a number of cases quoted shares, or at least quoted portfolio investment shares, are estimated at market value. Unquoted shares are to a large extent, currently left at book value. Household sector estimates are typically estimated on the basis of residual derivation, which is dependant on:

-the accuracy of the control totals for equity liabilities; and,

-the accuracy of the holdings of the major institutional investors in the economy.

10. In some cases, there seems to be a mixture of CMV and book value estimates in country balance sheet accounts. In a few cases – most notably, the U.S. IIP – there are three estimates of direct investment: book value, current cost (a version of NAV) and CMV.

11. A NAV type estimate is a feature – though possibly not highlighted – of some countries’ balance sheet accounts. Some countries only produce partial balance sheet accounts (financial assets and liabilities only) so that NAV is likely not an option for these countries, at least in the short-term.

12. Given the complexities related to valuing corporate equity, SNA93 should probably contain an expanded discussion of approaches to valuation.

Valuing unlisted equity with reference to listed equity

13. The CMV approach has wide support, and this likely be clearly reflected in a re-draft of SNA93. It is currently in use to some extent and its application was included in the recommendations of WGUS. In addition, a version of CMV emerged from the last DITEG meeting.

14. Essentially, the CMV approach (as drawn from the April TFVME paper by France) can be summarized as follows:

-Unquoted shares are valued as own funds at book value times a “capitalisation ratio”, calculated as the market value of quoted shares divided by their own funds at book value

-The capitalisation ratio should be discounted for differences in liquidity between quoted and unquoted companies

-Other equities are valued at own funds at book value.

15. The feasibility of this approach is related to the propensity to list in various economies. There is an issue of excluding very large and very small companies in certain industries so as not to bias the resulting capitalization ratios. There is an issue of whether a cutoff (based on size) should apply in the CMV approach, below which another approach (in particular, OFBV) should apply. There is a further issue of how to adjust (or whether to adjust) for differences in liquidity between quoted and unquoted shares. This issue will be discussed further below.

From an accuracy and comparability point of view there is the issue that the estimates are quite sensitive to assumptions made with respect to the application of the CMV method.

Other methods

Own funds at book value

16. In the work to date on balance sheet accounts, there appears to be a general sense that OFBV should be avoided if possible, except in the case of below a certain cutoff in the CMV approach. In IIP-FDI statistics this was a favoured approach for many years, and was considered early on as a an option for unquoted equity by DITEG; however, DITEG moved to endorse a current/market value methods at its final meeting (does this make OFBV more of a fall-back option).

17. It remains to be decided whether TFVME supports the retention of OFBV in a limited way.

Net asset value

18. Net asset value is total assets at current/market value less total liabilities (excluding equity) at market value. NAV is a conceptual construct in SNA93. In theory it seems that NAV should reconcile to CMV, but this is not the case in practice. A number of countries have NAV estimates and 2 versions of NAV(NAV and current cost) were supported in the last DITEG meeting.

19. NAV has some shortcomings. These include:

-PIM approach for produced assets;

-Inconsistent current/market valuation across the different asset-liability categories;

-Incomplete coverage of asset (e.g., especially of intangible assets);

-Retention of certain assets at face value (e.g., loans)

20. Having said this, it is likely very difficult not to accept a role forNAV in the valuation of unlisted equity. It may be all that is feasible in certain countries. It may be a reasonable substitute for OFBV in the cases where a size cutoff for the CMV approach is seen as important. It may be a useful memorandum item in the corporate sector.

21. TFVME should consider an expanded discussion of NAV and its role in supporting the measurement of unquoted equity.

Discounted present value

22. The DPV method for unquoted shares is essentially the discounted present value of future earnings. The usefulness of this method is that it may be better adapted to the certain economies, where a small number of securities account for a very sizable percentage of stock market capitalisations. Notably this method is consistent with SNA93, which indicate that the valuation at current prices of an asset can be approximated by calculating the present or discounted value of the flow of future returns generated by that asset.

23. In Spain, the DPV is estimated by discounting the profits obtained, using as a discount factor a rate inferred from the implicit discount rate obtained for quoted corporations and from other adjustments. This method always has at its heart the issue of the discount rate.

24. TFVME should consider DPV as an option for unquoted shares. Further, similar to NAV and OFBV, DPV could be considered as an option for valuing unquoted shares in the cases where a size cut-off for the CMV approach is seen as important.

Combination of methods

25. Does TFVME want to endorse the notion of a combination of methods in the valuation of unquoted shares (as a second-best option) allowing for more flexibility in methods across countries? Where does DITEG stand on this?

Liquidity discount in the CMV approach

26. There is theoretically a basis for assuming that unquoted shares should be adjusted to reflect the lower liquidity. However, the problems with the liquidity discount are not insignificant. First, it is difficult to come up with a reasonable discount factor. Second, that discount may vary over time and by industry. Third, the application of one discount factor may lead to unacceptable results in some industries and/periods.

27. The following paragraph is an excerpt from the paper by France delivered at the April TFVME meeting.

The difference in the prices of two assets, which are similar in all aspects except liquidity, is called the “liquidity premium”. A “liquidity premium” should be added up to the price of an unquoted share to obtain the price of a quoted share. Conversely, a “liquidity discount” should be deduced from the price of a quoted share to obtain the price of an unquoted share. Unless we have some information on the transaction price of unquoted shares, it is difficult to evaluate the general liquidity discount. In a study carried out for the French “Commissariat au Plan” and the September 2002 meeting of the Working Group on Unquoted Shares of Claude Picart of the Department of Business Statistics of the INSEE proposed a method for calculating the liquidity premium. The basic data are provided by LIFI, a database on financial links between parent and daughtercompanies, and by SUSE, a database on balance sheets based, among other sources, on tax reports. In a first step, the date of acquisition of shares by a parent company is identified. At this date, the shares are recorded on the asset side of the balance sheet of the parent company at the transaction price. In this way a market value is observed for an unquoted company. If the acquisition concerns only a part of the capital, the information on the percentage of control helps to assess the market value of the whole acquired company. After consistency controls, the sample contains 16,000 acquisitions from year 1994 to 1999.

In a second step, the own funds at book value of the acquired company are extracted in the year of acquisition, and then a capitalisation ratio is calculated for each acquired company. After elimination of small companies (own funds less than FRF 10 million) and of companies whose financial assets are more than 10% of total assets, the median ratio for quoted companies is 1.26, while the median ratio for unquoted companies is approximately 1. This would lead to a liquidity discount of 20%. In a third step, a liquidity discount is computed on price earning ratios (PER). The PER is preferred to the capitalisation ratio because it is supposed to refer to the standard textbook formula of valuation of financial assets and directly provide the actualisation rate, which includes the liquidity premium/discount. Moreover, the availability of financial links helps to overcome the usual shortcomings in the use of PER, i.e. the double counting of distributed profits in groups of companies due to the use of individual accounts. With LIFI, the companies with foreign subsidiaries are eliminated and the profit of the remaining parent companies are corrected of double counting of distributed profits using the profit and loss account of their resident subsidiaries. A multi-linear regression of the PER on the type of company gives a difference of 3.96 between quoted companies not included in an index and unquoted companies. Given a PER of 16 for quoted companies not included in an index, the liquidity discount would then be 25%. Thus, for the 2005 version of the financial accounts, a 25% discount was applied to the capitalisation ratios.

28. The TFVME should consider the option of supporting a liquidity discount in principle, but allowing countries the flexibility to decide whether it is feasible to adopt a liquidity discount in practice. This may be an important consideration, given that liquidity discount was not considered by DITEG. TFVME should also consider reporting specifically on this issue to BOPCOM.

Corporations versus quasi-corporations

29.SNA93 does not explain why the valuation of equity of small unlisted incorporated entities should be different than that of quasi-corporations (13.74), where a NAV estimate is recommended. Given that small incorporated entities may have more in common with unincorporated entities, this issue should likely be discussed further in an expanded re-draft. This relates to using a NAV estimate for smaller corporations (discussed above)

30. TFVME should probably consider this point.