## D.10 Defining the Boundaries of Direct Investment

## Source details

**Canonical URL:** [D.10 Defining the Boundaries of Direct Investment](https://www.imf.org/-/media/files/data/statistics/bpm6/approved-guidance-notes/d10-defining-the-boundaries-of-direct-investment.pdf)

## Other formats

- [Markdown version](/-/media/files/data/statistics/bpm6/approved-guidance-notes/d10-defining-the-boundaries-of-direct-investment.pdf.md)
- [Structured JSON version](/-/media/files/data/statistics/bpm6/approved-guidance-notes/d10-defining-the-boundaries-of-direct-investment.pdf.json)

---

### Overview
- The guidance note (GN) examines three boundary issues for direct investment (DI):
  - (A) potential inconsistency in subsidiary classification criteria among the Balance of Payments and International Investment Position Manual, sixth edition (BPM6), the Benchmark Definition of Foreign Direct Investment, fourth edition (BD4), and the 2008 System of National Accounts (2008 SNA);
  - (B) re‑examination of the DI threshold value (shifting the DI threshold from 10 percent to 20 percent); and
  - (C) whether direct investment relationships can be achieved other than by economic ownership of equity (e.g., through warrants or repos).
- GN recommendations:
  - maintain the numerical‑only DI threshold of the current BPM6 definition with no changes;
  - support the aims of GN G.2 “Treatment of MNE and Intra‑MNE Flows” to bring the SNA definition into alignment with BPM6; and
  - remove from the research agenda the question of whether DI relationships can be achieved other than by economic ownership of equity and retain the current treatment of warrants and repos.

### Issue A — Definition of a Subsidiary Corporation
- Context and definitions:
  - BPM6 defines DI as investment where an investor holds 10 percent or more voting power in its investee (BPM6, paragraph 6.12).
  - BPM6 notes that “Control or influence may be achieved directly by owning equity that gives voting power in the enterprise, or indirectly through having voting power in another enterprise that has voting power in the enterprise.” (BPM6, paragraph 6.12).
  - BPM6 acknowledges that “effective control or influence may arise in some cases with less than these percentages” but recommends strict numerical application for international consistency (BPM6, paragraph 6.13).
  - BD4 similarly uses the 10 percent voting‑power criterion as evidence of a lasting interest (BD4, paragraph 117).
  - 2008 SNA allows an option to define a subsidiary where an investor has the “right to appoint or remove a majority of the directors” even if shareholding is below 50 percent (2008 SNA, paragraphs 4.73 and 4.75).
- Key concern:
  - 2008 SNA’s option introduces a governance‑based, non‑numerical criterion that goes beyond the BPM6/BD4 numerical approach.
- DITT survey findings and positions:
  - The majority of DITT members were unable to identify cases where the SNA option had been used in practice; a few cited golden shares and certain investment‑fund arrangements.
  - Seventy‑nine percent (11 members) did not agree with adding a subjective criterion and preferred retaining a numerical threshold.
  - The drafting team and DITT favored alignment with GN G.2, which recommends the SNA adopt BPM6/BD4 definitions of control to avoid subjective judgments and ensure consistency.
- Alternatives proposed:
  - Alternative 1 (A1): Status quo — retain the numerical threshold as in BPM6 and support GN G.2 alignment of SNA with BPM6.
  - Alternative 2 (A2): Modify BPM6 subsidiary criterion to include 2008 SNA references (i.e., the “right to appoint or remove a majority of the directors”).

### Issue B — DI Threshold (10 percent versus 20 percent)
- Historical background:
  - The 10 percent criterion has been applied since at least the third and fourth editions of the Balance of Payments Manual.
  - In 2004 the Direct Investment Technical Expert Group (DITEG) endorsed moving the threshold from 10 to 20 percent, but the IMF Committee on Balance of Payments Statistics rejected the proposal due to concerns about data discontinuity.
- Contemporary evidence and considerations:
  - Expanding Portfolio investment (PI) means some investors hold 10 percent or more voting power without intending control; such holdings are conceptually PI but recorded as DI under BPM6, producing potential data fluctuations around the 10 percent cutoff.
  - Financial accounting (IAS) commonly treats 20 percent or more voting power as evidence of significant influence, aligning with investments that are less volatile and less likely to reverse in a crisis.
  - Regulators’ thresholds were deemed not relevant for the statistical distinction between DI and PI.
- OECD 2016 metadata survey (BD4) summary:
  - 20 out of 34 respondents strictly apply the 10 percent voting‑power criterion.
  - Nine respondents apply a value threshold in addition to the voting‑power criterion.
  - Five respondents include enterprises that do not meet the 10 percent voting‑power criterion but have influence on management and/or exclude enterprises that meet the 10 percent criterion but do not have influence on management.
- DITT consultation feedback:
  - Thirty‑six percent (five members) said moving to a 20 percent threshold would have little impact.
  - Sixty‑four percent (nine members) had not undertaken analysis on the topic.
  - Of four members commenting on merits of change, three wished to maintain the status quo and one supported a shift to 20 percent for greater coherence with company rules and realistic influence.
- Alternatives proposed:
  - Alternative 3 (A3): Status quo — retain 10 percent threshold (holding 10 percent or more voting power).
  - Alternative 4 (A4): Apply a 20 percent threshold (holding 20 percent or more voting power).

### Issue C — DI Relationships Other Than by Equity (warrants, repos, debt)
- BPM6 and BD4 research agenda items:
  - Explore whether DI relationships can be achieved other than by economic ownership of equity (e.g., through warrants or repos, or acquiring voting power without purchasing equity).
- BPM6 current treatment:
  - “Voting power is not recognized if temporarily obtained through the holding of warrants (because the warrant holder does not possess voting power until the warrants are exercised)” (BPM6, paragraph 6.19).
  - “Voting power is not recognized if temporarily obtained through repurchase agreements (because no change in the economic ownership of the shares has occurred)” (BPM6, paragraph 6.19).
  - Legal transfer of voting power can occur without change in economic ownership in cases of repurchase agreements or securities lending, but BPM6 does not recognize temporary holdings as voting power for DI.
- DITT findings:
  - Responses unanimously indicated DI relationships could not be practically identified beyond economic ownership of equity; such cases are rare or difficult to recognize within current compilation frameworks.
  - The drafting team noted potential for debt holders to influence management but questioned whether such influence is sufficiently prevalent or significant to be treated equivalently to equity ownership for DI purposes.
- Alternatives proposed:
  - Alternative 5 (A5): Maintain status quo and remove the item from the research agenda (i.e., retain current treatment of warrants and repos).
  - Alternative 6 (A6): Retain the item within the research agenda for future evaluation.
- DITT consensus:
  - Unanimously agreed to remove the question from the research agenda and retain current treatment of warrants and repos; recommended further investigation by the national accounts community of SPVs in orphan entity structures where a charitable trust, rather than the sponsor, owns the equity of the SPV (2008 SNA Research Agenda, paragraphs A4.14 and A4.15).

### Outcomes and Recommendations
- Issue A (Subsidiary definition):
  - The drafting team and the large majority of DITT members recommend retaining a numerical threshold and aligning 2008 SNA with BPM6/BD4 definitions of control as promoted by GN G.2.
  - Rationale: ensure a practical, unambiguous methodology for DI compilers; reduce responder burden; enhance comparability across countries; minimize bilateral asymmetries; and aid reconciliation with AMNE, CDIS, and CPIS statistics.
- Issue B (DI threshold):
  - The drafting team recommends maintaining the current 10 percent threshold (status quo, Alternative 3) given limited appetite and limited analysis among DITT members to support a change to 20 percent.
  - Considerations for 20 percent (Alternative 4) were noted (alignment with IAS and potentially less volatile capital flows), but consensus did not support change.
  - DITT feedback: 36 percent (five members) expected little impact from shifting to 20 percent; 64 percent (nine members) had not analyzed the implications.
- Issue C (Non‑equity DI relationships):
  - DITT unanimously recommended removing the question from the research agenda and retaining the current BPM6 treatment of warrants and repos (status quo, Alternative 5).
  - The drafting team noted continued interest in related national accounts research on SPVs and orphan entity structures.
- Committee consultation:
  - A wide majority of Committee members supported the GN recommendations as reflected in the written consultation outcomes.

### Section 2 — Summary of Decisions and Agreement
- Agreement:
  - Maintain the status quo with respect to the definition of DI.
  - Drop from the research agenda the topic of DI relationships achieved through other than economic ownership of equity.
- Additional note:
  - One member suggested raising the threshold for DI relationship to 50 percent noting that this would help in addressing some concerns on the impact of the RIE imputation.

### Section 2 — Recommendations (GN recommends)
- Accept A1: Status quo – A1 retains a numerical threshold as the current definition in BPM6 and supports the aims of GN G.2 with respect to bringing the SNA definition into alignment with the BPM6.
- Accept A3: Status quo (holding 10 percent or more voting power) – retain a 10 percent threshold.
- Accept A5: Maintain status quo and remove the item from research agenda – remove from the research agenda the question of “whether direct investment relationships can be achieved other than by economic ownership of equity (e.g., through warrants or repos)”.

### Annex II — Proposed Changes to Statistical Manual
- Balance of Payments and International Investment Position Manual, sixth edition: 1.43 A research agenda has been identified for possible future work. It includes the following:
  - Remove
    - (b) whether direct investment relationships can be achieved other than by economic ownership of equity (e.g., through warrants or repos) (see paragraph 6.19)

*Source: Approved/Final version of the guidance note D.10 by the Direct Investment Task Team (DITT), BPM6 Update, IMF Committee on Balance of Payments Statistics.*

### Section 1

### D.10 Defining the Boundaries of Direct Investment

### Overview
- The guidance note (GN) explores three issues related to the boundaries of direct investment (DI):  
  - (A) potential inconsistency in subsidiary classification criteria among the Balance of Payments and International Investment Position Manual, sixth edition (BPM6), the Benchmark Definition of Foreign Direct Investment, fourth edition (BD4), and the 2008 System of National Accounts (2008 SNA);  
  - (B) re-examination of the DI threshold value (shifting the DI threshold from 10 percent to 20 percent); and  
  - (C) whether direct investment relationships can be achieved other than by economic ownership of equity (e.g., through warrants or repos).  
- The GN recommends:  
  - maintaining the numerical-only DI threshold of the current BPM6 definition with no changes;  
  - supporting the aims of GN G.2 “Treatment of MNE and Intra-MNE Flows” to bring the SNA definition into alignment with BPM6; and  
  - removing from the research agenda the question of whether DI relationships can be achieved other than by economic ownership of equity and retaining the current treatment of warrants and repos.

### Issue A — Definition of a Subsidiary Corporation
- Context and relevant definitions:  
  - BPM6 defines DI as investment where an investor holds 10 percent or more voting power in its investee (BPM6, paragraph 6.12).  
  - BPM6 notes that “Control or influence may be achieved directly by owning equity that gives voting power in the enterprise, or indirectly through having voting power in another enterprise that has voting power in the enterprise.” (BPM6, paragraph 6.12)  
  - BPM6 acknowledges that “effective control or influence may arise in some cases with less than these percentages” but recommends strict numerical application for international consistency (BPM6, paragraph 6.13).  
  - BD4 similarly uses the 10 percent voting-power criterion as evidence of a lasting interest (BD4, paragraph 117).  
  - 2008 SNA allows an option to define a subsidiary where an investor has the “right to appoint or remove a majority of the directors” even if shareholding is below 50 percent (2008 SNA, paragraphs 4.73 and 4.75).
- Key concern: 2008 SNA’s option goes beyond BPM6/BD4 numerical approach by allowing a non-numerical governance right to define a subsidiary.
- DITT survey findings and positions:  
  - The majority of DITT members were unable to identify cases where the SNA option (appoint/remove majority of directors without 50 percent equity) had been used in practice; a few cited golden shares and certain investment-fund arrangements.  
  - Seventy-nine percent (11 members) did not agree with adding a subjective criterion and preferred retaining a numerical threshold.  
  - The drafting team and DITT favored alignment with GN G.2, which recommends the SNA adopt BPM6/BD4 definitions of control to avoid subjective judgments and ensure consistency.
- Alternatives proposed by the drafting team:  
  - Alternative 1 (A1): Status quo — retain the numerical threshold as in BPM6 and support GN G.2 alignment of SNA with BPM6.  
  - Alternative 2 (A2): Modify BPM6 subsidiary criterion to include 2008 SNA references (i.e., the “right to appoint or remove a majority of the directors”).

### Issue B — DI Threshold (10 percent versus 20 percent)
- Historical background:  
  - The 10 percent criterion has been applied since at least the third and fourth editions of the Balance of Payments Manual.  
  - In 2004 the Direct Investment Technical Expert Group (DITEG) endorsed moving the threshold from 10 to 20 percent, but the IMF Committee on Balance of Payments Statistics rejected the proposal due to concerns about data discontinuity.
- Contemporary evidence and considerations:  
  - The drafting team noted changes in the economic environment, including expanding Portfolio investment (PI) where some investors hold 10 percent or more voting power without intending control — such holdings are conceptually PI but recorded as DI under BPM6. Repeated institutional investor activity around the 10 percent cutoff can produce data fluctuations.  
  - Financial accounting (IAS) commonly treats 20 percent or more voting power as evidence of significant influence, which aligns with identifying investments that are less volatile and less likely to reverse in a crisis.  
  - Regulators’ thresholds were deemed not relevant for the statistical distinction between DI and PI.
- OECD 2016 metadata survey (BD4) summary:  
  - 20 out of 34 respondents strictly apply the 10 percent voting-power criterion.  
  - Nine respondents apply a value threshold in addition to the voting-power criterion.  
  - Five respondents include enterprises that do not meet the 10 percent voting-power criterion but have influence on management and/or exclude enterprises that meet the 10 percent criterion but do not have influence on management.
- DITT consultation feedback:  
  - Thirty-six percent (five members) said moving to a 20 percent threshold would have little impact.  
  - Sixty-four percent (nine members) had not undertaken analysis on the topic.  
  - Of four members commenting on merits of change, three wished to maintain the status quo and one supported a shift to 20 percent for greater coherence with company rules and realistic influence.
- Alternatives proposed by the drafting team:  
  - Alternative 3 (A3): Status quo — retain 10 percent threshold (holding 10 percent or more voting power).  
  - Alternative 4 (A4): Apply a 20 percent threshold (holding 20 percent or more voting power).

### Issue C — DI Relationships Other Than by Equity (warrants, repos, debt)
- BPM6 and BD4 research agenda items: explore whether DI relationships can be achieved other than by economic ownership of equity (e.g., through warrants or repos, or acquiring voting power without purchasing equity).
- BPM6 current treatment:  
  - “Voting power is not recognized if temporarily obtained through the holding of warrants (because the warrant holder does not possess voting power until the warrants are exercised)” (BPM6, paragraph 6.19).  
  - “Voting power is not recognized if temporarily obtained through repurchase agreements (because no change in the economic ownership of the shares has occurred)” (BPM6, paragraph 6.19).  
  - Legal transfer of voting power can occur without change in economic ownership in cases of repurchase agreements or securities lending, but BPM6 does not recognize temporary holdings as voting power for DI.
- DITT findings:  
  - Responses unanimously indicated DI relationships could not be practically identified beyond economic ownership of equity; such cases are rare or difficult to recognize within current compilation frameworks.  
  - The drafting team noted potential for debt holders to influence management but questioned whether such influence is sufficiently prevalent or significant to be treated equivalently to equity ownership for DI purposes.
- Alternatives proposed by the drafting team:  
  - Alternative 5 (A5): Maintain status quo and remove the item from the research agenda (i.e., retain current treatment of warrants and repos).  
  - Alternative 6 (A6): Retain the item within the research agenda for future evaluation.
- DITT consensus: unanimously agreed to remove the question from the research agenda and retain current treatment of warrants and repos; recommended further investigation by national accounts community of SPVs in orphan entity structures where a charitable trust, rather than the sponsor, owns the equity of the SPV (2008 SNA Research Agenda, paragraphs A4.14 and A4.15).

### Outcomes and Recommendations
- Issue A (Subsidiary definition):  
  - The drafting team and the large majority of DITT members recommend retaining a numerical threshold and aligning 2008 SNA with BPM6/BD4 definitions of control as promoted by GN G.2.  
  - Rationale includes ensuring a practical, unambiguous methodology for DI compilers, reducing responder burden, enhancing comparability across countries, minimizing bilateral asymmetries, and aiding reconciliation with AMNE, CDIS, and CPIS statistics.
- Issue B (DI threshold):  
  - The drafting team recommends maintaining the current 10 percent threshold (status quo, Alternative 3) given limited appetite and limited analysis among DITT members to support a change to 20 percent. Considerations for 20 percent (Alternative 4) were noted (alignment with IAS and potentially less volatile capital flows), but consensus did not support change.  
  - DITT feedback: 36 percent (five members) expected little impact from shifting to 20 percent; 64 percent (nine members) had not analyzed the implications.
- Issue C (Non-equity DI relationships):  
  - DITT unanimously recommended removing the question from the research agenda and retaining the current BPM6 treatment of warrants and repos (status quo, Alternative 5).  
  - The drafting team noted continued interest in related national accounts research on SPVs and orphan entity structures.
- Committee consultation:  
  - A wide majority of Committee members supported the GN recommendations as reflected in the written consultation outcomes.

*Source: Approved/Final version of the guidance note D.10 by the Direct Investment Task Team (DITT), BPM6 Update, IMF Committee on Balance of Payments Statistics.*

### Section 2

### Defining the Boundaries of Direct Investment — Section 2

### Summary of Decisions and Agreement
- The GN reports agreement to maintain the status quo with respect to the definition of DI and to drop from the research agenda the topic of DI relationships achieved through other than economic ownership of equity.
- One member suggested raising the threshold for DI relationship to 50 percent noting that this would help in addressing some concerns on the impact of the RIE imputation.

### Recommendations (GN recommends)
- Accept A1: Status quo – A1 retains a numerical threshold as the current definition in BPM6 and supports the aims of GN G.2 with respect to bringing the SNA definition into alignment with the BPM6.
- Accept A3: Status quo (holding 10 percent or more voting power) – This alternative suggests retaining a 10 percent threshold.
- Accept A5: Maintain status quo and remove the item from research agenda – This proposal seeks to remove from the research agenda the question of “whether direct investment relationships can be achieved other than by economic ownership of equity (e.g., through warrants or repos)”.

### Annex I — Supplementary Information (Referenced documents)
- IFRS Foundation (2018) “Use of IFRS Standards around the world (2018)", London
- IFRS Foundation “IAS28”
- IFRS Foundation “IFRS10”
- IMF (2008), Balance of Payments and International Investment Position Manual, sixth edition, Washington, DC.
- OECD (2008), Benchmark Definition of Foreign Direct Investment, fourth Edition, Paris
- OECD and Luxembourg (2004) "DITEG Issues Paper # 2 Direct Investment – 10 Percent Threshold of Voting Ower/Equity Ownership, Employment"

### Annex II — Proposed Changes to Statistical Manual
- Balance of Payments and International Investment Position Manual, sixth edition: 1.43 A research agenda has been identified for possible future work. It includes the following:
  - Remove
    - (b) whether direct investment relationships can be achieved other than by economic ownership of equity (e.g., through warrants or repos) (see paragraph 6.19)

### Annex III — DI TT Survey Questions (DI TT members responded during preparation)
- Q1 Have DITT members observed cases when an investor acquires influence by purchasing warrants that would convert to more than 10 percent equity ownership once exercised?
- Q2 Have DITT members observed the following cases related to repurchase agreements or securities lending? If yes, please mention the frequency of the cases observed (e.g., very rare to frequently observed).
  - a) The case where an investor accumulates more than 10 percent voting power of an entity via repurchase agreements or securities lending and vice versa.
  - b) Any other cases involving influence as a result of warrants and repurchase agreements.
- Q3 Moving beyond equity, have DITT members observed cases when an investor, regardless of equity holdings in an enterprise acquires significant influence through the purchase of other financial instruments?
  - If yes
    - a) Did the investor company have an up or down stream customer relationship with the investee company?
    - b) Did the investor company have an up or down stream customer relationship with the investee company and lend on favourable terms?
    - c) Did the investor have a traditional commercial (arm’s length) relationship (e.g., where the investor counterparty is a financial institution)?
    - d) Did the investor company already have a minor equity stake (i.e., below 10 percent)?
    - e) Any other scenario? Please describe:

---


_Source: https://www.imf.org/-/media/files/data/statistics/bpm6/approved-guidance-notes/d10-defining-the-boundaries-of-direct-investment.pdf_