Independent report

The report of the Taskforce on Pension Scheme Voting Implementation: Recommendations to government, regulators and industry

Published 20 September 2021

Foreword by the Chair

I would like to thank the many stakeholders -too many to name individually- who generously lent their time and their insights to the taskforce in the course of many conversations and email exchanges. COVID prevented any face-to-face work. The UK responsible investment sector will be directly affected by our recommendations; our engagement with it confirmed that the knowledge and potential to adopt and develop them is huge.

Finally, I thank my fellow members of the taskforce – Sarah Wilson of Minerva Analytics who was Vice-Chair, Leanne Clements of the Association of Member Nominated Trustees, Maria Nazarova-Doyle of Scottish Widows and Janice Turner of the Association of Member-Nominated Trustees. Our particular thanks go to David Farrar, previously of DWP, who provided the secretariat. All of them provided advice, ideas and just the right amount of challenge.

Simon Howard
Chair of the Taskforce on Pension Scheme Voting Implementation

Executive summary

The Taskforce on Pension Scheme Voting Implementation (“TPSVI” or “the taskforce”) was set up by the Minister for Pensions and Financial Inclusion (“MfPFI”) in December 2020 to address problems in the voting of equity shares by pension schemes. It reflects the importance of voting in broader stewardship.

The taskforce has focused on how to facilitate more voting and better-quality voting by occupational pension schemes by encouraging them to set voting policies and by making recommendations that will support the changes in behaviours needed from service providers to meet this objective.

Part 1 of the report considers context and key problems: complexity in pension and investment structures; complexity in how voting is delivered; issues with splitting the vote in pooled funds; and stakeholder attitudes and asymmetry of power. We conclude that legal structures and particularly those linked to investment through insurance platforms are an issue; that as others have found, IT is much cited as a problem but cannot be allowed to slow progress; that there are no material or insuperable problems with splitting votes; and that there are significant problems with stakeholder attitudes and asymmetry of power in relationships.

In part 2 we consider the steps an owner might take in setting a voting policy, namely policy development, policy implementation and reporting and monitoring, and we make our recommendations. Key recommendations are referenced there, the full list is on page 41. We found that few owners have set voting policies and instead many rely on fund manager policies which are often found wanting. In our consultation we found that 72% of managers have additional internal voting policies which are not shared with clients. We make a series of recommendations, notably that schemes need not set a voting policy but must take and demonstrate “ownership” of the policies carried out on their behalf, and fund managers should disclose policies more fully. In terms of implementation, we recommend that fund managers of pooled funds should voluntarily offer investors the opportunity to set “expressions of wishes” and we ask that the Financial Conduct Authority (FCA) should confirm the legality of aspects of that process. In respect of reporting and monitoring we note a mix in quality in current practice and a lack of consistency in who defines “significant vote” for implementation statements.

We conclude that there is much to improve in terms of vote reporting and monitoring with action required from fund managers. We recommend that the Department for Work and Pensions (DWP) promote a vote disclosure reporting template and that the FCA give guidance on a key set of aggregate data that asset managers should be required to report. More broadly, we suggest that the Occupational Pension Schemes Stewardship Council help pension schemes in the stewardship of service providers and that Department for Business, Energy and Industrial Strategy (BEIS) bring forward proposals to lower the thresholds for filing shareholder resolutions as recommended by the Asset Management Taskforce.

In part 3 we look forward. Because of some of the trends and problems identified in part 1, we recognise that our work has not future-proofed pensions against developments linked to voting. It cannot be denied for instance that savers who bear the risk have a right to understand and have a say over the stewardship of their assets. This will promote better saver engagement. Our interim solution is the adoption of expressions of wish, but looking forward, if adoption by managers is slow, we recommend the issue is referred to the Law Commission to propose structures that give owners the necessary rights. More broadly we suggest 4 principles covering: expression of wish, “form over substance” in respect of the nature of relationships, transparency over voting entitlements in products, and co-operation in the hope that stakeholders will recognise the direction of travel and act pragmatically. We also note some important innovations which have implications for expression of wish and split voting. We end by making the point that the UK should embrace changes of the kind we recommend such that UK financial services offer products that will support and grow UK pensions and be attractive to investors around the world.

Introduction and structure of the report

1. The Taskforce on Pension Scheme Voting Implementation (“TPSVI” or “the taskforce”) was set up by the Minister for Pensions and Financial Inclusion (“MfPFI”) in December 2020 to address problems in the voting of equity shares by pension schemes. It reflects the importance of voting in broader stewardship. Stewardship itself is now recognised as a core responsibility of asset owners, particularly large asset owners. The taskforce has focused on voting and deliberately avoided considering wider stewardship issues such as engagement: we have looked at the plumbing, not the water. Our focus has been: first on how to facilitate more voting and better-quality voting by occupational pension schemes by encouraging them to set voting policies; and second, making recommendations that will support the changes in behaviours needed from service providers to meet this objective.

2. We think that voting policies can and will evolve. For the purposes of this report, we use the term voting policy to mean a set of guidelines or instructions issued by a pension scheme to an agent to inform voting on their shares. These need not be precise and rigorous, they may be indicative and general. We expect many to be set on a “comply or explain” basis. We definitely do not envisage pension schemes instructing votes on particular motions; we do envisage them saying how they would like their votes cast in respect of particular topics and themes. The area is changing rapidly, with many observing the rapid growth in interest in social aspects of investment since the COVID-19 pandemic struck in 2019. We would expect policies to be either flexible enough to cope as changes unfold, or to be reviewed regularly. Above all, it seems to us that if the agents involved seek to understand what matters to their clients, any real difficulties will be minimised.

3. The taskforce was given the following terms of reference:

• help drive solutions to voting system issues, with specific reference to addressing present obstacles
• increase the number of asset managers who are prepared to engage with their clients’ voting preferences
• recommend regulatory and non-regulatory measures to ensure the convergence of asset managers’ approaches to voting policy and execution with trustees’ policies

4. There are 3 parts to this report:

• part 1 sets the context and examines some key background problems
• part 2 makes recommendations by examining the current problems facing a pension scheme seeking to set and implement a voting policy
• part 3 makes brief comments on the future

Part 1: context and key problems

Why voting matters

5. Some explanation for the focus on voting implementation is necessary. It is the exercise of votes^{[footnote 1]} at General Meetings that gives the risk-bearing capital providers the ability to hold directors, as the appointed agents of the providers of capital, to account for the governance and stewardship of investee companies.

6. Historically, much of this business has been considered as routine and of negligible interest to the markets. Indeed, such was the lack of interest that voting turnout was routinely below 30%. However, since the days of the Cadbury Report^{[footnote 2]} through the Myners Principles^{[footnote 3]} to the Kay Review^{[footnote 4]}, there have been increased demands for voting and stewardship rigour, whether through the appointment of auditors or election of suitably qualified and diverse boards of directors. Today, votes are seen as an important tool in the stewardship toolbox and, used wisely, can send important messages on a variety of issues from the appropriateness of executive remuneration to the rigour of net zero carbon strategies. With turnouts at the UK’s largest 350 companies now routinely in excess of 70%^{[footnote 5]} it’s clear that the Myner’s Report message about voting volume has been heard, although some might claim that the same cannot always be said about voting quality.

Figure 1: UK Top 350 Voting Turnout 2010-2020

Source: Minerva / Manifest

7. Today, resolutions on topics like climate are sometimes requisitioned by shareholders, the owners of the company. It is our view that motions of that type are likely to increase in number and to gain increasing public attention. It is making sure that the voting system can cope with increased scrutiny from stakeholders and the public that has shaped much of our work. In particular, we think that despite aspects of current legal structures that may restrict wider input into voting, it is probably naïve for industry participants to think that the views of asset owners and savers can be ignored.

8. In the year of COP 26 recent voting instances linked to climate show the good and the bad of the system. In Exxon, the large US oil company, investors have been seeking to force a better approach to climate change since 2016 when a shareholder resolution urging action on the issue received only 38% of votes. Resolutions have since been passed but the Board was not acting quickly enough and this year shareholders used their votes to install 3 new directors committed to change. In contrast, a report by Majority Action showed that 3 large passive managers voted respectively for 3, 2 and none of 12 climate resolutions filed by the ClimateAction100+ investor coalition. The resolutions were filed by asset manager leads not activist groups.^{[footnote 6]}

9. So, voting matters in terms of how companies are run, who is appointed to run them, and how they act. Good voting, used as part of broader stewardship, can help secure better financial returns for pension savers and that is a significant factor in seeking to improve the system.

10. In terms of how pension schemes can be more engaged with voting and stewardship, more transparency in respect of policies and actions by their advisers and managers will, we believe, foster a better market for stewardship. It will be easier for owners and savers to appoint the best advisers and fund managers and that will improve the quality of stewardship the system applies to firms. More broadly, the FCA business plan draws attention to the need for “investor stewardship by institutional investors, including voting at Annual General Meetings” in order to reach environmental and social goals and says it will regulate if needed.^{[footnote 7]}

Our assessment of the current position

11. Several trends are currently visible in the approach of pension schemes to stewardship and thus voting:

• increased focus on the need for stewardship including voting, and increasing demands being placed on schemes, particularly in relation to climate change and ESG issues;
• increased focus on the processes and outcomes of stewardship, and thus in part voting
• reporting issues connected to disclosure, data and forms of reporting

12. These trends are putting pressure on the current system of de facto delegation of stewardship and voting by schemes to agents – principally asset managers. We are in a position where the trustees of pension schemes feel they are being called on to do more whilst beginning to sense that their agents are not supporting them adequately and are not giving them the data necessary to review their performance. There is a gap between demand from clients – which needs to grow further – and supply from the industry.

13. In relation to demand, our aim is to stimulate it; there is latent demand stemming from the factors listed above, but it needs to be crystallised. In relation to supply, we want to make interventions which will make it easier for clients to implement voting polices or better review those being exercised on their behalf.

14. There are difficulties in moving forward in both areas, in part because of some legal and structural issues which make it easy for incumbent service providers to refuse reasonable requests. These are joined to some cultural issues, with many incumbent service providers showing considerable resistance to change.

Fundamental problems limiting the development of pension scheme voting

15. As work progressed, several key themes emerged. These were frequently factors in more than one area and they are best covered on a top-down basis as a way of showing the complexity of the background against which voting must develop. Readers should recognise that some key elements, notably legal structures, substantially influence the shape of all the recommendations.

16. But we do not have a magic wand. The complexity of the issues discussed here means that commonsense recommendations we might like to make are impossible.

Complexity in pension and investment structures

“If I were you, I wouldn’t start from here”

17. The apocryphal phrase could have been written for UK pensions. For instance, 5 different types of pension scheme each have over £100 billion invested in them, and some significantly over. These 5 differ fundamentally in the degree and form of governance methods, and in the level of risk to which the person in the scheme is exposed. And indeed, in what the “person in the scheme” is called in technical terms. The table shows some of the main features of the 5 different types of schemes:

Type of scheme	Does the person pay into the scheme?	Key governance body “protecting” the person in the scheme	Is a given level of pension certain?	Does the person in the scheme bear investment risk?	What is the person in the scheme called?
Trust-based defined benefit	In some no, but generally yes	Trustees	Yes	No	Beneficiary
Trust-based defined contribution	Yes	Trustees	No	Yes	Beneficiary
Public service defined benefit	Yes	Managers	Yes	No	Member
Workplace personal pension	Yes	Arguably the Independent Governance Committee of the pension provider has some role	No	Yes	Saver
Non-workplace personal pension	Yes	No-one	No	Yes	Saver

18. The point is this: objectively a vote and the consideration of it should be the same whatever the legal form of pension scheme on behalf of which it is cast, but the structures make the governance and execution processes for stewardship and voting very different. Some of these differences have material impacts in terms of where the responsibility for voting sits. The first 2 types of scheme have trustees. A board of pension trustees has many duties, but their key feature is that they operate to benefit the beneficiary. In the case of trustees, stewardship is a statutory responsibility. We deem stewardship to encompass voting, but the extent of trustee duties in respect of voting has not been fully defined. A case has been made by competent parties that trustees can wholly delegate all responsibility for voting to agents. We are sceptical as to that and we adopt what we think is the practical and commonsense approach of saying schemes have responsibility for voting policy, but they are not expected to “do” the voting.

19. However, some types of pension schemes do not have trustees and it is difficult to identify who “protects” the saver in those structures. While every firm or agent providing services to those types of pensions sits under a range of regulatory and legal obligations, and those obligations have often been drafted with the interests of the saver in mind, none of these firms owe the saver a fiduciary duty of undivided loyalty.

20. What further complicates matters is that the equity assets of all these types of schemes can frequently end up being invested alongside those under other types of ownership in the same fund, a pooled fund. The “ABC asset management equity fund” can have money invested representing the savings or other rights of a variety of pension scheme types and potentially other types of savers as well such as charities and retail investors. In investment terms this is usually a good thing, and the economies of scale may further benefit investors in the form of lower costs.

21. But in terms of voting (and stewardship) there may be material disadvantages from the aggregation of equity assets from at least 2 causes. First, the process of aggregating assets to capture benefits of scale extends beyond investment into administration. Typically, an asset manager holds all their equity assets with one custodian with the holdings in different funds aggregated. When a vote is cast it is not differentiated by fund but cast as part of a single block all voting the same way. This means the investee company having the vote will see a potentially massive block vote from a single name – typically not that of the manager but of the custodian. The company will require the co-operation of the custodian and asset manager to establish who owns the shares further up the chain if it wishes to understand more about who is investing in its shares.

22. The second problem is in the phrase “who owns the shares” itself. In legal terms it is always clear who owns shares. The problem for pension schemes in voting is that the legal owner is often NOT the pension scheme (or saver) who invests, because the chain of ownership, driven by the desire for economies of scale and outsourcing, is heavily intermediated. The Law Commission summarised the position of ownership in intermediated securities chains in its recent scoping paper on the subject:

“Under the law of England and Wales, it is now “reasonably well settled” that the arrangements between parties in an intermediated securities chain are characterised as a ‘series of trusts and sub-trusts’ ”^{[footnote 8]}

and further:

“As well as trusts and sub-trusts between the parties, the relationships in the chain are regulated by individual contracts entered into between the relevant parties at each stage in the chain. For most investors, the contract will reflect the intermediary’s standard terms and conditions, rather than a bespoke arrangement according to the ultimate investor’s preferences.”^{[footnote 9]}

23. One key area in the pensions chain where these complexities are evident is where a pension scheme invests through an insurance platform into what are commonly termed “unit-linked contracts of insurance”. These contracts typically pass money from the insurer’s customers – including pension schemes – to pooled funds that are either managed by an asset manager linked to the insurer or one completely independent of the insurer.

24. If a pension scheme uses such a platform, it has a contract with an insurer. In simple terms the insurer contracts to provide the scheme with the return from the nominated fund(s). The insurer buys units in the fund from the asset manager which buys shares with the money. The return is paid to the insurer who pays the equivalent to the pension scheme. But because of the terms of the contract between the scheme and the insurer the pension scheme does not own the shares, they are owned by a party further “down” the investment chain. Nor does the pension scheme even own units in the pooled fund – those are owned by the insurer. What the pension scheme “owns” are rights stemming from a contract. This position exists whenever one of the pension scheme types outlined above purchases through a platform.

25. There is absolutely no doubt that this is the legal position and indeed some trustees that have asked questions about voting in unit-linked contracts or in pooled funds have been told they do not own shares and thus have no voting rights.

26. The shares which are purchased do have voting rights. Those rights are exercised by parties other than the pension scheme, usually with the asset manager central to decisions. The votes are exercised through the “the trusts and sub-trusts” the Law Commission referred to and are typically subject to the contracts between the various parties using the “intermediary’s standard terms and conditions”. Since as stated above the pooled fund involved may have a very wide range of investors going beyond pension schemes, it is likely that the standard terms will have been drafted on a broad, lowest common denominator basis which confers few if any rights on the investor.

27. In the pensions chain under discussion here there are 3 actors: the pension scheme; the insurer; and the fund manager running the pooled fund. The pension fund cannot instruct on voting because it owns no voting rights. The insurer may be able to instruct if it has the right contractual relationships with the fund manager, but in most cases it hasn’t. As the value of stewardship is more widely recognised, some insurers are unhappy that they are unable to exercise more control over the rights attaching to shares in the pooled fund. Our understanding is that many of the insurers, whilst benefitting from the structures would like to see change and to be able to exercise more control^{[footnote 10]}. There may however, be resistance to passing that control further up the chain to the pension schemes, and that may become an issue in the future.

28. It is worth noting that investing via an insurer is attractive to some pension structures for various reasons. In certain circumstances schemes would qualify for higher compensation if they had invested through an insurer, for instance in the case of a failed investment intermediary. Investing through a platform tends to make administration easier. The approach is so popular that some firms have set up insurance subsidiaries solely to attract pension business and operate as facilitators. That fact may suggest the structures should be seen as increasingly of purely administrative utility which in turn would suggest that the investment effects should be disentangled, rather than remaining something driven by administrative factors alone. There are some signs of this happening, see page 37, and our view is that changes should be made.

29. Some stakeholders suggest that using an insurance structure may help to “insulate” fund managers from clients.

30. The legal structures of pension schemes and their governance models, and the form of investment funds lie at the heart of many of the problems of pension scheme voting. They are the result of evolution and the search for economies of scale. These offer benefits, but in the case of voting there are disadvantages, namely the limits on the power of investors to argue for their voting policies and wishes. In effect, the service benefits of pooling are bundled with the limitations on the ability to set a voting policy. There is no automatic reason why this must be so. Many in the asset management industry have simply decided that it should be.

Complexity in how voting is delivered

Structure and processes

31. The delivery of equity votes in the UK is also extremely complex. Some of this complexity stems from the legal structures point above, some from attempts to vote “efficiently”. This is not the pension schemes’ fault; it is the result of the processes of aggregation and search for economies of scale in the investment chain.

32. One of the key links is the asset manager. Most asset managers outsource some or all of their voting activity, usually to boost efficiency. Typically it is the administrative processing of votes.

33. In addition to vote execution administration, asset managers may also outsource some or all of the consideration of how to vote. Outsourcing is not an all or nothing process, rather it is a continuum. Managers may buy data and information, they may buy opinions or recommendations on how to vote or indeed they may delegate the voting decisions completely. The nature and processing of the opinion varies. Some managers may follow the house recommendations received from their service provider or they may use the guidance as input to an internal review process. Others may have a bespoke voting policy and custom recommendations.

34. Asset managers review vote recommendations received to differing degrees. Our consultation carried out between June and August 2021 showed that 24% of asset managers review less than a quarter of the recommendations they receive. 41% review all of them. One asset manager reviews none, and one said the information was confidential. One manager reviewed 100% of recommendations in active funds but relied very heavily on unreviewed recommendations in index funds.

35. Some asset managers seek to use their expertise to offer “overlays” or policies to owners and asset managers. They are of course another link in the chain.

The role of IT in voting

36. Good technology, and fit for purpose procedures are indispensable for smooth investment operations. However, it is clear from comments made to us that systems issues are a key factor in the resistance to innovation in voting. Professor Iain Clacher looked into this for the Association of Member Nominated Trustees (AMNT) last year and his interviewees in turn cited under-investment and this was further echoed in a comment made to us:

“People won’t touch the legacy systems, the people who know them have left”.

37. Our interviews with stakeholders have left us in no doubt that there are things owners and the responsible staff in asset managers would like to do in terms of voting and disclosure which they cannot because of inadequate systems.

We suspect that some of this resistance is driven by financial factors such as the under-investment Prof Clacher cited, and that many players are patching systems to cope with change on an incremental basis if at all. Why stewardship and voting has been such an orphan in this respect is beyond our remit, but there does appear to be a chronic resistance to investment in stewardship and voting systems which would not be acceptable anywhere else in the investment chain. However, the cumulative combination of lack of investment, open standards, and competition has stunted innovation leaving both clients and stewardship teams with difficulties.

38. As the FCA has identified, asset managers are heavily reliant on robust and reliable technology to ensure the smooth operation of their businesses and the protection of client assets. It therefore expects managers to ensure technology is managed correctly, including appropriate oversight of third-party firms and intra-group service providers. Despite this expectation, very few people in the investment industry would deny that vote plumbing has suffered due to a combination of increased expectations undermined by a chronic lack of investment in systems and technology. Straight through vote processing therefore remains as elusive as it does for other corporate actions.

39. Some have suggested that it is problems with registrars or companies that prevent split voting, however the facts would say otherwise. Voting levels have risen over the past 30 years from substantially below 25% to over 70% in the UK’s largest companies.

40. As per our remit, our focus has been on the barriers to asset owner participating in the voting process from policy through to reporting. We have therefore looked at the upstream plumbing – the parts of the system between the asset owner and manager, to uncover the technological and administrative barriers which make split voting apparently so daunting for managers.

41. The key area of complexity in the upstream area is ownership of assets in pooled funds. A pooled fund is an investment vehicle such as a Common Contractual Fund, unit-trust or SICAV where assets from different owners are combined. These structures can generate economies of scale, tax efficiency, and are generally a positive development. However, there is rarely a direct legal or operational connection between the investment product that a pension saver buys and the legal ownership records which the company sees. This therefore requires asset managers to undertake complex manual reconciliation tasks between separate record keeping systems to ensure accurate voting entitlement. The complexity is compounded when products are sold through multiple intermediaries or platforms. Most importantly, it is highly unlikely there are any forms of unique client or product identifiers that would make the necessary connections between what has been bought by the client and what is owned by the fund which could enable automated reconciliation. This does not just impact vote execution, but also implementation statement reports and other forms of reporting such as costs and charges. Looking ahead, when pension funds need to be able to provide the Task Force on Climate-Related Financial Disclosures (TCFD) reporting, accuracy in ownership will be even more crucial to manage client risk on behalf of their beneficiaries and clients. This reporting will be even more crucial if we are to achieve our Climate Change commitments.

42. Stewardship teams are extremely dependent on multiple disparate -and often creaking - systems elsewhere in their organisation. For many decades, it has been argued that improving voting is a desirable public good and that it should be made easier for pension schemes to have influence over the stewardship of their investments. We therefore have made some specific recommendations which will help define a better end state. In general, we take the view that adequate investment in IT can deliver what is needed for our recommendations. We do not, and the industry must not, regard technology as a limiting factor.

43. Some argue that market forces will effect change, and indeed recent innovations show that the problems are not insuperable. Read our comments on market forces and issues of asymmetry.

Issues with splitting the vote in pooled funds

44. In the light of evolving pension scheme regulatory requirements, a reasonable request from owners to the managers of pooled funds is an expression of wish – “please vote this way on certain issues”. In general – but not universally – such requests are refused. Four reasons for refusing are usually cited:

• legal barriers in terms of ownership
• IT and operational problems
• a weakening of the manager’s voice when it comes to dealing with investee companies
• regulatory barriers linked to the operation of the vehicles

45. However, despite these objections we are aware that several pension funds, have been offered pooled fund split voting services on some or all of their investments and split voting is routinely taking place. It is therefore disappointing that the asset management industry has routinely and publicly denied what is happening on a day-to-day basis. We do not underestimate the potential operational issues associated with split voting in pooled funds at scale, however, considerable time and effort would have been saved if the issues had been more openly discussed and the issues addressed.

46. Investment structures in the UK have become extremely complex in recent years. Some of this complexity stems from the legal structures point above, some from attempts to achieve greater investment administration efficiency. This is not the schemes’ fault; it is the result of the processes of aggregation and search for economies of scale and increased profitability in the investment chain.

47. We recognise the issues of ownership discussed above, and do not repeat that discussion here. Readers might wish to note the first 2 quotes in the next section

IT and operational problems

48. When we asked fund managers whether operational difficulties would prevent splitting votes, 38% said it was “practically impossible”, 50% said “somewhat, but can be overcome”, and 12% said there were “no barriers”.

49. We received comments from fund managers in response to these and other questions linked to splitting votes and we show some below:

“We do not believe there is a legal barrier to implementing this but there are legal consequences which are difficult to overcome such as disclosures.”

“There are some [legal barriers] in some markets, but the main concern is operational and philosophical, plus cost to pension schemes.”

“We can split the votes if desired/required”

“There are technical barriers and cost barriers. Plus risk and control challenges for asset managers.”

“To enable client-directed voting in pooled funds at scale, technological developments are required to calculate pro-rata share ownership as well as other additional operational developments to build an interface with vendors in the proxy voting chain. In addition, there are certain markets where other hurdles exist, such as share-blocking or where split voting is not permitted.”

“Depending on the number of clients in the pool, it’s operationally time consuming to spilt out clients’ holdings and vote accordingly.”

“There are several operational barriers that prevent split voting. Firstly, current systems (including third party systems used in vote administration) do not have the ability to efficiently accommodate split voting at the fund level. Secondly, there is not currently a mechanism for gathering vote instructions from clients. This would likely require a sophisticated system to translate individual policies into votes on individual ballot items. Lastly, the complexity and costs of monitoring the entire process would increase.”

“[It] would be very time consuming with current IT systems”

50. The comments confirm the presence of IT and operational issues, but none seem to us sufficient to block this development should clients want it. Two important groups of clients do want it. First is pension trustees. This taskforce was set up by the Pensions Minister Guy Opperman in response to the Association of Member Nominated Trustees’ report on barriers to split voting in pooled funds last year. The AMNT, whose trustees govern collectively over £1-trillion of assets, has been calling for several years for the right of pension trustees to have their voting policies respected by fund managers on a comply or explain basis.

51. The second group is the Association of British Insurers (ABI). As mentioned above, some insurers are unhappy with their limited rights in respect of pooled funds into which they invest their clients’ funds. We received this comment from the ABI in respect of split voting:

“Our members who hold customer’s funds in pooled schemes with the asset manager would like to see the ability to split the vote. This would enable voting preferences throughout the investment chain to be better heard. Pension providers have a stronger relationship to their scheme members, and their voting preferences (which are detailed in their Environmental, Social, and Governance (ESG) and stewardship policies) will often take into account customers’ views, collected through extensive customer research.”

Weakening of “voice”

52. The argument is that a manager’s engagement impact is weakened if it is known or suspected that the manager cannot vote all the shares they manage.

53. We disagree with this argument for several reasons. First, it assumes that the manager is right and any of their customers wanting to vote the other way are wrong. This will clearly not always be the case. Indeed, it is not unusual for there to be important votes where the asset management community is divided. For instance, at the Shell 2021 annual general meeting (AGM) 2 climate-change resolutions, one proposed by management and one by shareholders each received considerably different levels of support despite both being concerned with the same fundamental issue. Data for the 2021 voting season shows that 196 resolutions on remuneration received dissent of 20% or more, the highest number since remuneration policy votes were introduced in 2013. Set against the prevailing level of support for management on such resolutions of over 96%^{[footnote 11]}, there are clear differences of opinion as to what is, or is not, contentious.

54. We see 4 further flaws with this argument:

• just as there is not a universal consensus among asset managers on what is “right”, there can be no assumption that clients’ views should be “wrong”
• asset managers that offer segregated mandates where the client can direct voting will already be splitting the votes they control if a segregated client differs in their view
• as noted above, several large funds are already able to instruct on voting in pooled funds
• we have been told some asset management firms allow considerable autonomy to individual fund managers to decide individual votes with clients

55. These factors suggest that a significant number of asset managers engage with investee companies despite splitting votes. Common sense would also suggest that investee firms know full well that on contentious issues at least 2 legitimate views will exist – that is why the issue is contentious.

Regulatory and other barriers linked to the operation of the vehicles.

56. This argument has been put to us regularly and our detailed thoughts on the regulatory challenges are shown below. In short, we do not believe they are well founded, and we believe our engagement with regulators supports this position.

Our understanding of the regulatory position

The legal arguments made to us against the ability of trustees or other investors to set a voting policy in a pooled fund appear to rest on 6 arguments. We summarise and respond to each below – the viewpoint is our own, but it has been informed by engagement with the FCA. Our recommendations call for the FCA to make their position public that there is no barrier to setting an expression of wish in a pooled fund.

1. Section 235 of the Financial Services and Markets Act 2000^{[footnote 12]} – section 235 refers to the requirement that the participants in collective investment schemes do not have day-to-day control over the management of the property. However, this does not exclude the exercise of voting rights. This is self-evident given that section 235 is itself qualified by “…whether or not they have the right to be consulted or to give directions”. Therefore, the “right to be consulted or to give directions” alone is not intended to imply day-to-day control. The FCA’s response to Q 9 of Perimeter Guidance (PERG) chapter 10 makes this explicit. In our view, setting out an expression of wish in relation to voting some or all shares in a pooled fund does not constitute day-to-day control, just as setting a voting policy in relation to a segregated mandate does. This is in alignment with guidance issued by the FCA.

2. Collective Investment Scheme handbook (COLL) chapter 6.6A.2^{[footnote 13]} – This section requires that an authorised fund manager of an Undertakings for the Collective Investment in Transferable Securities (UCITS) scheme must refrain from placing the interests of any group of unitholders above the interests of any other group of unitholders (there are corresponding measures for alternative investment funds (AIFs) in the Alternative Investment Fund Managers Directive (AIFMD) level 2 regulation which are incorporated into UK law via a Statutory Instrument, rather than FCA rules). This requirement is an overarching principle, but it does not articulate Regulator expectations at the level of voting. There are many circumstances where different services are offered, for example share classes are priced differentially according to investments or the type of client (advised, non-advised, institutional, or part of the corporate group) or a different basis for responding to requests. These are not challenged by the FCA. Allowing pooled fund voting is simply providing a service to some clients. The principle of refraining from placing the interests of any group of unitholders above the interests of another group is a very general principle, and allowing some investors to set an expression of wish – in an existing or new share class – would not in our view be a breach of that rule.

3. COLL 6.6A.6^{[footnote 14]} – This rule requires that an authorised fund manager of a UCITS scheme must develop adequate and effective strategies for determining when and how voting rights are to be exercised, to the exclusive benefit of the scheme concerned, and the strategy must determine measures and procedures for ensuring that the exercise of voting rights is in accordance with the investment objectives and policy of the relevant scheme. The manager must have such strategies and investment objectives and policy. But it would not be contrary to strategy, objective or policy to invite expressions of wish or to respond to expressions of wish which were put to the managers. The investment objectives and investment strategy are set out in the fund prospectus. In our view accepting an expression of wish is wholly compliant with the exercise of voting rights in accordance with the investment objectives and policy of the scheme. For many or most existing pooled funds, a change to investment objective to signal a willingness to accept an expression of wish would at most be a notifiable change, of which managers could simply inform investors.

4. COLL 6.6.13(2)^{[footnote 15]} - This FCA rule sets out that where the scheme property of an authorised fund contains units in any other scheme managed or otherwise operated by the authorised fund manager, the depositary must exercise any voting rights associated with those units in accordance with what he reasonably believes to be the interests of the unitholders in the authorised fund. This provision is solely made to protect other unit holders from asset managers who may have large holdings in a fund through its position as an underlying fund in a fund owned by those managers. It ensures that the manager holding units through an underlying fund cannot independently vote their units [the units, not the votes associated with the underlying shares] by entrusting them with a depositary. It is not intended to bite on the exercise of voting rights associated with sharers.

5. “Treating customers fairly” – FCA principle 6- Some managers have sought to argue that if they gave a service to some investors (for example occupational pension scheme trustees) and not others (for example, individual retail investors), that would not constitute treating customers fairly. As with COLL6.6A.2, this is a high-level principle and is not articulated at the level of voting. Also, as with COLL6.6A.2, it does not serve as a bar for providing a service to some clients – there are many other circumstances where different services are offered for different clients, for example in relation to costs or reporting.

6. “We’d have to have a meeting of the investors in the fund” – this argument is predicated on the suggestion that allowing some investors to set an expression of wish would be a fundamental change to the fund and therefore needs to be voted on. In our view, such a change is highly unlikely to be a fundamental change to the fund prospectus. Taking a step back, implementing an Expression of Wish is just one possible option managers will engage with when making a decision, or – looked at another way – just a change of process in instructing on votes. Managers would not trigger a fundamental change to a fund needing an investor meeting in switching from one source of voting advice to another. They do not need an investor meeting to set an expression of wish.

The one persuasive argument put to us is that the Conduct of Business Sourcebook (CoBS) 6.6A.6 prevents managers from following voting instructions which are perverse and clearly against unit-holders’ interests – for example in relation to the wind up of a company where under one option investors would receive a share of the firm’s assets whereas under the other they would receive nothing. However, in the vast majority of votes it is arguable that to vote in either way would not be contrary to the interests of unit-holders and is therefore entirely consistent with FCA rules. This echoes the point made above that there is no monopoly of wisdom anywhere in the voting chain. The use of an “expression of wish” would be a way in which trustees could set a voting policy without fettering the discretion of their fund managers in such rare instances.

Stakeholder attitudes and asymmetry of power

57. A recurring comment when we have raised pension scheme concerns with voting issues has been “if they don’t like what they’ve got they should move their assets”. This remark is both correct – people can fire a service provider – and shockingly wrong. It speaks to attitudes of incumbency and complacency so worrying as to require comment.

58. In the pension world scale matters. Large pension schemes are courted and can get bespoke services. But there are very few of them- only 20 UK schemes have more than £15 billion in assets, which is perhaps big enough to bang the table^{[footnote 16]}. In contrast 900 of the 920 pure defined contribution (DC) schemes^{[footnote 17]}, and 5200 of 5500 defined benefit (DB) schemes^{[footnote 18]} have less than £1 billion in assets. The large schemes can more than hold their own – these are the type of schemes that can already instruct votes on pooled funds. The much larger number of small schemes generally get an off the shelf, one size fits all service. We have mentioned above how the Law Commission identified that in the intermediated securities chain, for most clients “the contract will reflect the intermediary’s standard terms and conditions, rather than a bespoke arrangement according to the ultimate investor’s preferences.”^{[footnote 19]} Much of the feedback we have received speaks strongly to this being the case when voting issues are raised: clients are told that they don’t own shares, and that they have the standard service and no chance to flex it.

59. But the problem extends far beyond voting – and this is beyond this taskforce’s remit. Many fund managers seem to offer very poor client service. We give examples in part 2 in respect of fund manager voting polices and vote reporting.

60. The problem is that it can be hard and expensive for pension schemes to change key suppliers like asset managers and insurance platforms: fiduciaries are rightly required to take advice; there are administrative costs; there are investment costs in selling and buying assets; and in many cases we know there is the feeling that “the new lot will be just the same” – something which is highly likely to be the case, given the unwillingness of the vast majority of managers to accept a trustee voting policy and the vagueness and lack of specificity in many manager policies^{[footnote 20]}. With significant frictional costs in changing supplier^{[footnote 21]}, the uncertainty over the offerings of alternatives and no opportunity to suggest a voting policy without losing the efficiencies of pooled funds, the line from incumbents that unhappy clients should move is rather disingenuous.

61. Disingenuous too is the idea that schemes are knowingly accepting a less comprehensive set of rights (including on voting) by investing in a pooled fund. This idea seems new to most people involved in responsible investment. It was referenced in the Law Commission report.^{[footnote 22]} The argument can be made- but seems not to have been: as far as we know no service provider has ever done so. We have heard from scheme trustees who say they have “knowingly” delegated voting to their asset manager, but we have heard none that say they have “knowingly” accepted “a less comprehensive set of rights”. The difference is material. If this idea is an ex post reaction from industry players to the emergence of valid concerns on voting (and other) issues it is very disappointing.

Conclusions to part 1

62. In this section we have outlined our views on some key, recurring topics. We sum them up as follows:

• Voting is of growing importance in securing value for pension savers of all kinds.
• There is enormous complexity in pension and investment structures which makes it hard for pension schemes to exercise control over voting.
• There is enormous complexity in how voting is delivered which again works against the interests of pension schemes.
• We are told that there are many issues with splitting votes in pooled funds, but none appears material or insuperable
• There are significant problems with some stakeholder attitudes and the asymmetry of power between pension schemes and their agents.

63. These factors and our views on them inform our analysis and recommendations in part 2.

Part 2: recommendations and supporting evidence

64. In this section we focus on ways to enable and encourage greater focus on voting by schemes, and on ways of addressing issues and difficulties stakeholders are finding in practice. We move away from the top-down approach used in part 1, and we use a structure that mirrors what a “good” process in setting and operating a voting policy for a pension scheme would look like, namely:

• policy development
• policy implementation
• reporting and monitoring

65. For each area we comment on what the current position is, informed by our many conversations and by responses to our consultation carried out in summer 2021. We then summarise our recommendations with the relevant full text recommendation noted in brackets in bold thus: (12). The complete list of full text recommendations is at Annex 2.

Policy development

Few asset owners have set voting policies

66. The issues covered in part 1 give rise to an environment where few but the largest and/or most determined pension schemes consider in detail how the votes attached to their assets are used.

67. In our consultation only 31% of asset owners said they had set their own voting policy. Since the responses were generally from large and/or more “active” owners we think this probably overstates the overall percentage by some margin. In a DWP survey of large schemes in early 2021, 5 schemes had their own voting policy, 7 delegated to a third party other than a fund manager, and 25 delegated to a fund manager.

68. In our consultation some 41% of owners said they relied on the fund manager’s policy, which we feel probably understates the reality, especially among the long tail of medium and small sized pension schemes. 28% of owners relied on an overlay provider or used a trade association policy.

69. Given the reliance on fund manager voting policies and the number of schemes with multiple managers there is a risk of managers having duplicated holdings and voting them in contrary ways. In our consultation we asked about the perception of this risk. 59% of asset owners were concerned about it as were 60% of fund managers. In contrast, 89% of advisers shared the concern. It may be that the advisers with their broader view are noticing it more than the other 2 groups but that is a supposition.

70. We have been told that trustees are not in breach of the law if votes cast by their managers conflict, but from the perspective of voting and stewardship it is a concern. When reviewing data on votes and especially “significant votes” as defined in DWP’s disclosure regulations^{[footnote 23]}, schemes should be aware of when their managers have voted contrary to each other and should consider the implications. These conflicts might be linked to defensible differences in manager approach – but they may well be harder to explain. This is one of the aspects we would expect trustees to report on under recommendation 1 below.

Fund manager voting policies

71. Given the number of owners that rely on fund manager voting policies it is important to understand their quality. Two bodies have surveyed fund manager voting policies. The Association of Member Nominated Trustees^{[footnote 24]} (“AMNT”) reviewed the voting policies of 42 fund management firms in 2018 in respect of climate change, women on boards, ethnic diversity, and executive pay. The AMNT found amongst other things:

• 20 voting policies did not refer at all to climate change or 2 degree scenarios
• 3 referred to climate change but had no voting guideline (which would set out how they may vote at the AGM)
• 16 fund managers referenced gender diversity on boards but with no voting guideline
• 28 did not refer to ethnic diversity in their voting policy
• few fund managers (approximately 25% of those that published their policy) had a voting guideline on tackling inappropriate remuneration in some form

72. In 2020 ShareAction^{[footnote 25]} considered voting policy as part of a broader process of ranking 75 fund managers’ approach to responsible investment. In respect of voting policy, the findings echoed those of the AMNT:

• most assessed voting policies included no specific commitments regarding shareholder proposals on climate, human and labour rights and biodiversity
• most asset managers’ voting policies lacked commitments on human rights due diligence, remuneration structures and non-discrimination
• only a few of the assessed voting policies contain explicit guidance for biodiversity-related resolutions

73. Bearing in mind the current regulatory environment in the UK which addresses the integration of ESG and stewardship factors in the investment process, these 2 reports suggest strongly that some fund manager voting polices are misaligned with their ESG policies – or perhaps even worse, their stated ESG policies do not reflect the reality of firm investment practice. We note that some asset managers recently failed to receive Stewardship Code accreditation from the Financial Reporting Council (FRC) because “The organisations that did not make the list commonly did not address all the Principles or sufficiently evidence their approach, instead relying too heavily on policy statements.”^{[footnote 26]}

74. In our consultation we asked about the level of disclosure, and 72% of fund managers said they had additional internal voting policies which were not disclosed to clients. The most usual reasons given were linked to wanting the freedom to evaluate each situation on its merits. But this has the effect of making it less clear to the client what action will usually be taken. We have heard anecdotal evidence that some clients are told more about voting polices than others.

75. Coincidentally or not, evidence suggests that assessing voting policy may play a limited role in the appointment of fund managers. In a question with defined answer options in the consultation, only 35% of advisers said that when working with a client to appoint a fund manager that they explicitly considered stewardship or voting in the overall appraisal. 15% used a service provider and 10% said it wasn’t considered. (The question format means answers do not sum to 100%).

Support for asset owners wanting to set a policy may not always be readily available.

76. If a scheme wants to set a voting policy, it isn’t clear how much support is readily available. Of the schemes that had sought to set a policy, 50% said their advisers had been helpful, but 50% said advisers had only helped “somewhat”, “when prompted”, or had not helped at all. Only 35% of advisers said they proactively sought to help clients develop voting policies, and 70% of advisers said they did not have a current default policy in respect of voting which informs work with clients. One interesting comment from an asset owner was:

“….voting and engagement was not a priority to the investment consultant; [but] it was a priority to the trustees.”

77. Anecdotal evidence supports the idea that investment consultants are not always supportive of trustees wishing to be more robust on voting issues or indeed other aspects of their relationship with fund managers. We think this may be a function of 2 things: the consultants’ fee-driven business model, and their wider position as intermediary between other stakeholders which means signals of interest from trustees, and of receptiveness from asset managers can get lost.

78. Consultants exercise significant, and in some cases potentially excessive, influence especially among smaller clients. The Competition and Markets Authority (“CMA”) reviewed the investment consultancy industry in 2018 and recommended that it be regulated by the FCA^{[footnote 27]}. But as important as regulation will be increased competition and culture change in consultants. A focus on delivering to pension schemes will enable those schemes to better comply with their legal duties.

79. However, there are encouraging signs that attention is beginning to focus on voting with 81% of asset owners that responded saying they anticipated reviewing their voting policy within 12 months.

Our observations and recommendations

80. These factors lead us to make the following observations and recommendations.

81. We do not recommend that owners be forced to draw up their own detailed voting policy. Whilst we think their doing so would be very desirable, our hesitancy reflects the limited resources available to many schemes. One scheme commented that:

“We do not have resources to develop a policy and (most importantly) implement and monitor it.”

We note however, the requirement under DWP’s investment regulations^{[footnote 28]} for schemes to include in their SIP and their default SIP a stewardship policy to encompass the trustees’ policy in relation to voting, engaging, and monitoring. We also note the availability of “off the shelf” voting policies such as those of the Pensions and Lifetime Savings Association (PLSA)^{[footnote 29]} and the AMNT^{[footnote 30]}.

Therefore, we recommend to the DWP that if a scheme does not set its own policy, it be obliged to accept responsibility for the policies carried out on its behalf. We recommend that schemes report as appropriate on differences between their own policy and that of their fund manager(s), or on differences between the polices of their various fund managers. These reports should be informed by the guidance in the UK Stewardship Code 2020 (Recommendation 1).

82. We recommend also that DWP should promote voting through all aspects of its work with a focus on linking to Stewardship Code guidance (4). In addition, DWP or TPR should provide guidance on what good quality voting policies look like. As stated above we think this area is likely to evolve rapidly. (7)

83. We recommend that investment advisers and consultants be regulated as recommended by the CMA. From our perspective, this regulation should focus on disclosure of how they resource and undertake work related to stewardship and voting (15). Linked to this, we recommend that in Stewardship Code submissions consultants and advisers describe how they will review and judge service providers that fall short in respect of principles 6 (client needs) and 12 (exercising rights and responsibilities) (23).

84. We recommend that when giving advice on setting and implementing investment strategy, investment consultants and advisers should flag the importance of voting and voting policies (24). We recommend that the Investment Consultants Stewardship Working Group should promote both the importance of voting to member firms and others giving advice in this area, and the importance of the industry working both as individual firms and as a group with other parties on voting issues. (22)

85. To encourage more focus on voting and reflecting our view of its importance as a driver of leading practice, we recommend that the FRC consider increasing the profile of voting as a distinct activity in stewardship as it reviews the effect of the 2020 code (17). We also urge the FRC to tier service providers by Stewardship Code submissions as soon as practicable (18).

86. We recommend that the FCA should ensure full disclosure of fund manager voting policies (11).

Policy implementation

87. Where trustees of schemes set policies – which we encourage them to do – we make recommendations to ease implementation based on informal stakeholder feedback and responses to the consultation. In part 1 we outlined some difficulties in owners implementing voting policies with respect to law, IT, and attitudes to splitting the vote in pooled funds. Our view on the last 2 was that the problems were far from insuperable, and we supported that with evidence from the consultation responses. Here our recommendations and discussion are focused on what is possible under the current law.

Segregated funds

88. Most fund managers say they will consider accepting voting polices from pension schemes if negotiated for a bespoke or segregated mandate. In such mandates the “standard terms and conditions” may be varied, although the willingness of managers to do so is far from universal. A scheme’s ability to get the terms it wants is greatest in the selection and negotiation phase of manager selection. There is advice available on how to do this available from various sources.^{[footnote 31]}

89. In the case of schemes with segregated mandates the issue is the willingness to ask for what they want. Our view is that we can’t force pension schemes to make such requests, but we think it is desirable.

Pooled funds: “Expression of wish”

90. There are undoubtedly legal problems that mean pension schemes investing through an insurance platform cannot “force” an instruction or policy on service providers further down the voting chain. However, in respect of non-insurance assets, several firms have recently introduced new products and services whereby they will act on what is termed an “expression of wish” from the pension scheme.^{[footnote 32]} These firms have concluded, following detailed legal and regulator engagement, that they are not breaching their legal obligations by doing this. The feedback we have received supports this view.

91. By setting the expression of wish the pension scheme indicates how they would like the shares to be voted. We see no reason why the expression of wish should not be drafted to represent a voting policy on any issue or group of issues.

92. Views on expression of wish were perhaps predictably divided in our consultation. When we asked stakeholders if they thought fund managers should be required to offer an “expression of wish” or proportional/split fund voting in pooled funds 62% of asset owners said yes, but 70% of fund managers said no. As much as this indicates potential demand, it shows a seriously sharp divide between clients and their service providers.

93. We would not be surprised if the form and content of expressions of wishes became in time relatively standardised. Stakeholders tell us that managers would need some precision in the expression; we would expect that market practice and experience should allow the degree to become known rapidly. However, to be clear, we do not expect schemes to give extremely precise expressions. We see this as a tool whereby instructions at the strategic and policy level can be set. If service providers are as client-centric as they often claim, they will facilitate it.

94. Some may say that expression of wishes are illegal. We do not agree, and having spoken widely to stakeholders, lawyers and conduct regulators we do not think such objections are sustainable.

Our observations and recommendations

95. We recommend that all fund managers should voluntarily offer pooled fund investors the opportunity to set expressions of wish on request. (20)

96. To speed the process and avoid any questions on legality, we think the DWP and FCA should give their views publicly (3, 10).

97. If progress does not materialise at the pace required, we recommend the issue be referred to the Law Commission (2). This is discussed more fully in part 3.

Reporting and monitoring

98. Data and reporting by agents are crucial to improving pension scheme voting. They are tools whereby owners can hold their agents to account against the instructions they have given and if necessary, review them against peers if a change is needed. Importantly, reporting by the scheme can then inform and engage savers and beneficiaries. Clearly, reporting by schemes can be no better than the reporting to the schemes by their agents.

Current state

99. Current vote reporting generated critical comments in the consultation. When we asked asset owners about the main problems they faced in monitoring managers voting and engagement work, the most frequent answers from the list of prompts were “Insufficient information” “boilerplate” and the lack of connection to investment decisions. Advisers’ responses were the same.

100. We asked owners “In what circumstances do you receive information about why votes have been cast (‘vote rationales’)?” and prompted with examples. 61% of owners received rationales for votes against management and 43% for shareholder resolutions but 17% received no rationales. These are low numbers for votes for which signatories to the Stewardship Code are expected to provide rationales in line with principle 12.

101. We also asked about narrative reporting, “How would you describe the narrative disclosures, including the specific reasons for exercising votes?”. 61% of owners said they knew the reason votes were cast, but 31% answered “It is lacking in detail and so it is difficult to understand the reasons for voting”. One comment was revealing:

“When the fund manager is setting out these votes as case studies, they are bespoke and informative. Other narrative disclosures are too brief.”

102. Under new regulations all occupational schemes with over 100 members need to report on “significant votes”^{[footnote 33]}. We asked how easy it had been to get the necessary data. 58% of advisers, firms that frequently help with the gathering of data and the preparation of the reports, said that it had been “difficult” or “very difficult” to get the data; none said it had been easy. 48% of asset managers said it had been “difficult” or “very difficult” to provide clients with data.

103. In relation to defining “significant vote”, asset owners said 55% of the time it was being done by fund managers and 21% of the time by the investment adviser. Advisers said it was fund managers 68% of the time. It is certainly apparent that few owners are doing the defining. Some respondents did feel it was right that fund managers define what is significant, but it seems wrong to us that asset owners have such a small role in defining this: the owners must be able to say what matters to them, and we would expect it to be covered in voting polices they set or in those set out by their fund managers.

104. The link between data and reporting is important: to clients, data alone has little value unless they have the staff and skills to interpret it. Clients need the full story and not all of them get it. We asked how often reporting linked voting and engagement to investment performance, investment decision making and key risks. The most frequent answer was “never”, although some said reporting covering these aspects came “occasionally, unprompted” whilst up to a third could get such information when they prompted the manager. We think voting is a part of investing and certainly plays a part in risk management. These data suggest the system is not currently delivering in these areas.

105. Given the volume of voting data, we were interested in ways that narrative reporting could develop to ease comprehension and better show the effect of voting. We asked stakeholders if reporting significant votes and associated engagement under key themes would be helpful. This found widespread support: positive responses from asset owners, fund managers and advisers were all over 87%.

106. Whilst change in fund manager vote reporting is underway, driven by amendments to FCA rules implemented in PS 19/13^{[footnote 34]}, several problems remain.

107. Firstly, fund managers are only required to report at the firm level although some are now voluntarily reporting at fund or mandate level. Furthermore, they are required to report on all but insignificant votes (see below).

108. Trustees, on the other hand are required to report on the significant votes related to their investments,

109. The consequence is that if a manager does not report at the relevant mandate level, schemes will receive data or reports in relation to shares they have not owned for time periods that do not relate to the scheme’s reporting period.

110. Our consultation showed strong support for the idea that schemes should be able to request all data: nearly 90% of asset owners, advisers and fund managers agreed that schemes should be able to access all the voting data related to their investments.

111. In addition to the levels of reporting, one key area identified by service providers involved in vote reporting was the lack of consistent and standardised formats of disclosure which led to numerous problems in report preparation. We asked the stakeholders involved whether they felt there would be benefit in having i) technical standards and ii) service level agreements in place for voting and engagement reporting. The results were as follows:

Fund managers

Response	Technical standards	Service level agreements
“helpful”	80%	52%
“unhelpful”	20%	48%

Advisers

Response	Technical standards	Service level agreements
“helpful”	95%	79%
“unhelpful”	5%	21%

112. The great unknown in this area is whether the poor client service reported to us is the result of understandable problems as regulations change, or deliberate behaviour aimed at minimising costs. Both are possible - and certainly even “good” firms struggled at times with producing data for implementation statements. But we suspect some was the result of inadequate resourcing. The TPR has a third-party compliance regime which makes provision for sanctions on parties that frustrate the ability of trustees to meet their existing duties such as reporting on significant votes, and we think this should be used if necessary.

Observations and recommendations

113. There is clearly much to improve in terms of vote reporting and monitoring. More action is required from managers to support schemes. We recognise that there are complexities in the investment administration eco-system and that the transposition of the Shareholder Rights Directive into UK law was both hasty and literal, with little opportunity to develop any implantation guidance or standards. However, the difficulties in vote disclosure are essentially a symptom of a wider systemic problem with investment reporting. Consequently, there is more to be done and a greater good will be accomplished by more hygienic investment plumbing overall, for example by the ability to trace asset ownership end to end, in much the same way as we expect in our food supply chain, back from fork to farm.

114. We think DWP should encourage more and better reporting by schemes and should promote a vote disclosure reporting template. There is already precedent for this in, for example, the United States where Form N-PX specifies a standard disclosure format and so is used by many of the world’s largest asset managers. Such a format would remove the necessity for additional cost and customisation which may be used as a reason for non-disclosure^{[footnote 35]} (5,6)

115. Given that there is more effort involved in managers excluding so-called insignificant votes than in disclosing all votes, the FCA should set expectations in handbook guidance of fuller disclosure on vote rationales, (12), and it should carry out a review of manager use of the easement in respect of not reporting “insignificant” votes (13a). The FCA should also assess whether manager voting polices help or hinder scheme investment decisions (13b). The FCA should set out in handbook guidance a key set of aggregate data which managers should be required to report (14). The priority at all times should be to ensure open standards, consistent with both the UK Government^{[footnote 36]} and FCA principles^{[footnote 37]} on the benefit of data inter-operability to empower customers and ensure fair competition.

116. The DWP, FCA and TPR should monitor the delivery of data at fund and mandate level and intervene if it is too slow (8).

117. More broadly, asset managers and their trade bodies should sign up to the principle of answering reasonable requests about their voting and stewardship polices (21).

118. We urge TPR to use its powers against service providers that prevent trustees fulfilling their duties (16).

Other areas

119. We are worried about the imbalance in power between parties in the voting chain. Support is rightly given to collaboration between commercial rivals as a way of accelerating stewardship activity. A well-known example is ClimateAction 100+.

120. We think something similar is needed to help pension schemes – particularly the smaller ones in their work on stewardship and voting. These schemes not only need to keep their service providers up to the mark in respect of the stewardship of investee companies, but they should also help each other in the stewardship of the service providers themselves. We see no reason why a forum operating in this area could not facilitate the exchange of information about service providers. We suggest the Occupational Pensions Stewardship Council^{[footnote 38]} might fulfil the role.

121. There is a need for this: we have quoted reports on fund manager voting polices by 2 non-governmental organisations – schemes themselves should be supporting and carrying out similar work. We believe that a body with the aim of gathering and sharing information would help schemes better engage with service providers.

122. There are also technical changes that will help shareholders file shareholder resolutions.

Observations and conclusions

123. We recommend that the Occupational Pensions Stewardship Council consider the needs of smaller schemes in its work (9).

124. We recommend that BEIS bring forward proposals to lower the thresholds for filing shareholder resolutions as recommended by the Asset Management Taskforce^{[footnote 39]} (19).

Part 3: the future, a final recommendation, 4 principles and innovation

125. Our remit was to suggest improvements to the current system of pension scheme voting implementation, and we have done so. But these proposals are all made against the “as now” background rather than the situation which is likely to evolve in the next few years. We are quite realistic: our work has not future-proofed pensions against the pressures linked to voting that are already evident and which we think will only increase.

126. The central issues are the poor alignment of interests, coupled with the very limited say for newer savers who bear investment risk. In layperson’s terms, these mean that the person paying into the pension has no say in how the voting rights their payments generate are exercised. Some in the pensions chain say this is right because the issues are complex, and without their expertise votes will not be cast “correctly”. In our view these issues will need to be addressed soon. Failure to do so risks undercutting the perceived legitimacy of the pensions system. In the span of a career the prevailing pension system has changed from the paternal DB where the saver bore no investment risk and paid few if any contributions, to the generally DC system current where most workers pay into schemes and bear all the investment risk. Structures and attitudes which originated in the former system remain current now and are increasingly not fit for purpose. It surely cannot be denied that savers who bear the risk have a right to understand and have a say over the stewardship of their assets. Yes, they probably shouldn’t be encouraged to carry out portfolio construction or the day-to-day management decisions in relation to that money, but surely they should have the right to say for instance “use our votes against managements that lag on climate change or pollute”.

127. Some may say this is a political stance – we disagree. Politicians on all sides agree that people need to save more into their pensions. That will be easier to bring about if the saver feels engaged and listened to; it will be easier if the people they are paying are seen to be listening; it will be easier if the saver feels a degree of alignment and ownership. Well over 10 million people are now paying into DC schemes with most of the cash ending up in pooled funds where legal structures often limit voting rights.

A final recommendation

128. Our interim solution to some of these issues is recommendation (20), saying that certain owners, depending on the legal structure of their investments, should be able to make expressions of wish and have them acted on by fund managers. But there may be opposition and resistance. If that is the case, then we think there is no alternative but to change the law - a significant task, given some of the issues involved. Our final recommendation (2) is:

If the market does not move more rapidly to offer Expression of Wish or equivalents across all pensions investment structures, we would recommend the Minister for Pensions to ask the Law Commission to propose legal frameworks that give investors the necessary rights.

129. That recommendation is the nuclear option. An obvious follow on for instance is whether it would be just assets in pension schemes that are affected. If pension savers get rights, why not savers more broadly in ISAs or direct savers in pooled funds? We do not seek to answer those questions, but it all looks a bit like Pandora’s box.

Four principles

130. We think the investment community need to respond to these challenges and should start now. A lot could be done by simply adopting better attitudes. Culture and attitudes don’t lend themselves to regulation, and frankly, given the prevailing complexity, we have, at times, found it difficult to put the required attitudes into terminology suitable for actionable recommendations. Therefore, we are putting forward some principles -some of which could also be applied beyond pensions. They are deliberately high level, and they overlap. We suggest 4:

• owners of all types should be able to set expressions of wish on a comply or explain basis which are actioned through the chain
• in responding to reasonable requests from parties in the same pensions chain, all stakeholders should be willing to operate on a “form over substance” basis where the reality of who does what takes priority over legal forms
• all stakeholders should be transparent as to the entitlements and restrictions in respect of voting that each of their products offer: transparency will drive choice
• all stakeholders in the voting chain must work in tandem to improve and deliver to savers – the system is only as strong as the weakest link

Innovation

131. The problems we have addressed are system wide, but it does not mean that relevant innovation is not underway. A number of recent developments should be noted:

• AMX and partners^{[footnote 40]} announced a new pooled fund service in February 2021 which “…aggregates investor stewardship preferences and seeks to execute votes in alignment with their expressions of wish. Where aggregate investor preferences conflict within a pooled fund, voting instructions can be split accordingly.”
  
  
• Tumelo^{[footnote 41]} is working with Aviva and Legal & General among others to pass unit-holder’s opinions on topics like climate and pollution to fund managers ahead of AGMs. The feedback is considered as part of the manager’s voting decision-making. Tumelo reports that in trial schemes up to 30% of members have engaged, and in one scheme members placed 14,000 ‘votes’ for their fund manager’s consideration. Before using Tumelo’s platform, typically two-thirds of members did not know which funds they owned, let alone which companies were held or what issues their fund managers were voting on.
  
  
• Citizen Shareholders^{[footnote 42]} is building a global mechanism for triggering minority shareholder rights. They aggregate unit-holder support on curated shareholder resolutions and empower voting on board proposed resolutions. The former instructions are intended to be binding, the latter being subject to local jurisdictional statutory and fiduciary rules. Asset Managers are only able to join the project if they agree to roll this service out across all their funds containing equities.

132. What all these show is the idea of expressions of wish being transmitted from the unitholder to the asset manager or insurer who is required to consider them to a greater or lesser degree. AMX and Citizen Shareholders are explicit that votes in pooled funds will be split as required. Our understanding is that other providers accept that this is a likely outcome and some are preparing products to serve this growing market.

133. The UK is a world-leader in finance and fintech. However, the regulatory and cultural environment sometimes lags and holds back development. While we cannot regulate for culture, we should ensure that the regulatory environment supports and empowers both clients and service providers to deliver better solutions. Engaging clients by allowing them to express more opinions on stewardship and voting is clearly possible in IT terms: the comments and consultation responses from industry stakeholders in this report and the innovations listed immediately above show this. The legal and regulatory environment should continue to be kept up-to-date and changed to better support owner-directed and pension trustee voting. We can be nimble. We should be so now, and we should build an environment that offers products that will support and grow UK pensions and be attractive to investors around the world.

Annex 1: Key Principles for Voting

Government, regulators and industry should all sign up to the 4 principles below:

• owners of all types should be able to set expressions of wish on a comply or explain basis which are actioned through the chain
• in responding to reasonable requests from parties in the same pensions chain, all stakeholders should be willing to operate on a “form over substance” basis where the reality of who does what takes priority over legal forms
• all stakeholders should be transparent as to the entitlements and restrictions in respect of voting that each of their products offer: transparency will drive choice
• all stakeholders in the voting chain must work in tandem to improve and deliver to savers – the system is only as strong as the weakest link

Annex 2: Recommendations

Recommendation 1

Trustees should either:

• set their own voting policy; or
• if they have not set their own policy, acknowledge responsibility for the voting policies asset managers implement on their behalf

This should be done by trustees publishing, via the Statement of Investment Principles:

• their voting policy if set
• where another organisation or organisations decide how votes are cast, those policies in relation to at least 3 topics and an acknowledgement of the trustees’ acceptance of them

And trustees should publish, via their Implementation Statement:

• an assessment of the efficiency of implementation. This could be achieved by requiring trustees to report in their implementation statements on voting against the voting policies identified in their SIP in relation to at least 3 topics which trustees believe to be most significant to members’ best interests
• where there are differences between the trustees’ own policy and that of the asset manager(s), or between the differing policies of asset managers, their assessment of the importance of those differences, and the actions trustees are taking to address the differences, in accord with Stewardship Code principle 8 (monitoring managers and service providers).

Addressed by: DWP

Recommendation 2

If the market does not move more rapidly to offer expression of wish or equivalents across all pensions investment structures, we would recommend the Minister for Pensions ask the Law Commission to propose legal frameworks that give the owners the necessary rights

Addressed by: DWP

Recommendation 3

DWP should publish guidance on the ability of trustees to set an expression of wish over their own voting policy in pooled funds.

Addressed by: DWP

Recommendation 4

DWP should encourage trustees through all aspects of its work and through legislation if necessary to take more account of voting and engagement policies in their appointment and ongoing monitoring of investment managers. Specifically, DWP should refer to relevant Stewardship Code guidance for example principles 7 (stewardship, investment and ESG integration) and 12 (exercising rights and responsibilities)

Addressed by: DWP

Recommendation 5

DWP should encourage trustees – especially of larger schemes – to report more effectively on how votes relating to their investments were cast. Trustees should report in narrative form on the topics identified in their policy (see recommendation 1) and on votes related to those policies. Whilst they should recognise the primacy of financial materiality in selecting those votes, criteria should include the trustees view of significance and perceived member/saver interest.

Addressed by: DWP

Recommendation 6

DWP in conjunction with industry bodies should promote a vote disclosure reporting template. It should be extended to cover engagement activities carried out on the trustees’ behalf. On engagement, they should work with Investment Consultants Sustainability Working Group (ICSWG) who have already begun to develop a reporting template.

Addressed by: DWP

Recommendation 7

DWP or TPR should provide guidance on what good quality voting policies look like, including examples, both for the purposes of trustees wishing to set their own policies and for trustees wishing to assess asset manager voting policies. This guidance should give examples to help trustees report as recommended in recommendation 1.

Addressed by: DWP / TPR

Recommendation 8

The DWP, FCA, and TPR should closely monitor delivery of vote reporting at fund or mandate level. If managers do not deliver by the end of 2022 the FCA should legislate or issue handbook guidance to deliver fund and mandate level-reporting.

Addressed by: DWP/ FCA / TPR

Recommendation 9

Trustees from smaller schemes suffer from an imbalance in power and information compared to their service providers. The Occupational Pensions Stewardship Council is urged to consider the implications of this in its work, and the particular needs of small and under-resourced schemes that cannot deal on an equal footing with large service providers.

Addressed by: DWP / OPSC

Recommendation 10

The FCA should issue clarification – for example, via guidance – to indicate that there is no breach of fund rules in acting on an expression of wish.

Addressed by: FCA

Recommendation 11

The FCA should ensure that all asset manager customers are treated fairly, and that disclosures on “private” voting policies are made to all clients and not only a subset

Addressed by: FCA

Recommendation 12

The FCA should bring forward handbook guidance at the earliest opportunity to set the expectation that all asset managers explain on request and in a prompt manner their rationale for all voting decisions, in particular where:

• there was a vote against the Board
• votes against and for resolutions laid by shareholders
• a vote was withheld
• the vote was not in line with the manager’s public policy

or to publicly disclose a clear and reasoned explanation of why they have chosen not to.

Addressed by: FCA

Recommendation 13

(a) The FCA should carry out a review of the first year of asset manager vote reporting. If widespread use is found of the permission NOT to disclose votes that are deemed “insignificant”, FCA should produce handbook guidance clarifying expectations on insignificant votes or legislate to remove this easement.

(b) The FCA should also carry out a study of the quality of asset manager voting policies to assess the extent to which they facilitate or impede trustees in making effective appointment decisions. Where adverse effect on market competition is found, the FCA should produce handbook guidance clarifying expectations on asset manager policies.

Addressed by: FCA

Recommendation 14

The FCA should set out – working with wider industry and documenting in FCA handbook guidance if necessary – a key set of aggregate data which all asset managers should be required to report at firm level on their voting and engagement activity.

Addressed by: FCA

Recommendation 15

We recommend that Investment Consultants should be regulated as recommended by the CMA to HM Treasury. The regulation should focus on: how they resource and undertake stewardship and voting work; how they support clients’ integration of stewardship and voting into investment; and how they manage conflicts of interest between clients and asset managers

Addressed by: HMT / FCA

Recommendation 16

In the period before the outputs of some of our recommendations are delivered by the FCA, in particular numbers 8 and 12, TPR should use third-party compliance notices against asset managers that frustrate the ability of trustees to meet their existing duty of reporting on significant votes, or the fuller duties we envisage being placed on them from recommendations 1 and 5

Addressed by: TPR

Recommendation 17

The FRC should consider increasing the profile of voting in subsequent editions of the Stewardship Code.

Addressed by: FRC

Recommendation 18

The FRC should tier service providers by quality of Stewardship Code submission as soon as practicable, making sure that voting and engagement is given due attention in the analysis

Addressed by: FRC

Recommendation 19

In line with the recommendations of the Asset Management Taskforce, BEIS should bring forward proposals to lower the threshold for filing of shareholder resolutions from 5.0% to 2.5%, whilst maintaining the alternative test for 100 shareholders who hold on average £100 of paid up capital.

Addressed by: BEIS

Recommendation 20

All asset managers should offer pooled fund investors the opportunity to set an expression of wish in a fund at the cost of implementation. Any costs for doing so should only reflect the marginal cost of operationalising the individual policy, and not the costs of upgrading IT and voting infrastructure to enable vote splitting.

Addressed by: Asset managers

Recommendation 21

Asset managers and their trade bodies should sign up to the principle of answering all reasonable requests on their voting and stewardship activity. They should not work on the basis that reporting via the PLSA template, abiding by the Stewardship Code and compliance with FCA rules will be sufficient. They should be willing to provide answers to all reasonable requests from clients. Asset managers should ensure that links to all current voting policies are contained within their stewardship reports to clients

Addressed by: Asset managers

Recommendation 22

ICSWG should work with their respective firms and other service providers to highlight the importance of shareholder voting for pension schemes. This should include making the quality of asset manager voting policies and voting disclosure, as well as their willingness to enable trustees to set voting policies, a key factor in determining fund recommendations.

ICSWG should work together to hold fund managers to account for their unwillingness to accept client voting policies in pooled fund arrangements. They should promote the “expression of wish” TPSVI recommendation to both their client base and most importantly fund managers.

Consultant firms should be willing to engage in respect of the development of voting practices with all stakeholders on behalf of their own individual client bases and as an industry

Addressed by: Investment consultants and advisers

Recommendation 23

In their Stewardship Code submissions, consultants should fully describe the steps and actions they take in respect of service providers that do not meet the standards expected of them, particularly in respect of principles 6 (client and beneficiary needs) and 12 (exercising rights and responsibilities)

Addressed by: Investment consultants and advisers

Recommendation 24

The person trustees consult on setting an investment strategy should tell them that they should consider voting as part of it. They should flag the existence of template policies and guidance

Addressed by: Investment consultants and advisers

1. Whilst there are other types of “votes” in investment, we are considering the exercise of voting rights attaching to equity shares in public listed companies. Whilst the majority of votes will be exercised by owners of equity shares, bondholders will also be called on from time to time to exercise their votes. Given the shifts in capital raisings in recent years away from equity to fixed income, we feel it is important to emphasise that voting as a generality is an important investor protection. ↩

2. “Given the weight of their votes, the way in which institutional shareholders use their power to influence the standards of corporate governance is of fundamental importance. Their readiness to do this turns on the degree to which they see it as their responsibility as owners, and in their interest of those whose money they are investing, to bring about changes in companies when necessary, rather than selling their shares”. [Para 6.10] and “Voting rights can be regarded as an asset, and the use or otherwise of those rights by institutional shareholders is a subject of legitimate interest to those on whose behalf they invest. We recommend that institutional investors should disclose their policies on the use of voting rights.” [Para 6.12] Source: The Financial Aspects of Corporate Governance, December 1992 https://www.frc.org.uk/getattachment/9c19ea6f-bcc7-434c-b481-f2e29c1c271a/The-Financial-Aspects-of-Corporate-Governance-(the-Cadbury-Code).pdf ↩

3. Effective intervention, when appropriate, is in the best financial interests of beneficiaries. As such, it is arguably already a legal duty of both pension fund trustees and their fund managers to pursue such strategies. [para 5.89] Institutional Investment in the United Kingdom: A Review (The Myners Report) March 2001, https://webarchive.nationalarchives.gov.uk/ukgwa/20070506151732/http:/www.hm-treasury.gov.uk/media/2F9/02/31.pdf ↩

4. Kay Principle 1: All participants in the equity investment chain should act according to the principles of stewardship, based on respect for those whose funds are invested or managed, and trust in those by whom the funds are invested or managed. [Page 12] July 2012 The Kay review of UK equity markets and long-term decision making: final report)https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/253454/bis-12-917-kay-review-of-equity-markets-final-report.pdf ↩

5. Source: Minerva Analytics Voting Review 2021 ↩

6. https://shareaction.org/wp-content/uploads/2020/11/Voting-Matters-2020.pdf p21 ↩

7. https://www.fca.org.uk/publication/business-plans/business-plan-2021-22.pdf p44 ↩

8. Law Commission - Intermediated Securities - Scoping Paper para 2.63 ↩

9. Law Commission - Intermediated Securities - Scoping Paper - para 2.66 ↩

10. See comment from the ABI on page 19 ↩

11. Minerva Analytics 2021 UK Season review ↩

12. PERG 9.4 Collective Investment Scheme (section 235 of the Act) FCA handbook ↩

13. COLL/6/6A Duties of AFMs of UCITS schemes to act in the best interests of the scheme and its unitholder ↩

14. COLL 6.6A Duties of AFMs in relation to UCITS schemes and EEA UCITS schemes - FCA Handbook ↩

15. COLL 6.6 Powers and duties of the scheme, the authorised fund manager, and the depositary - FCA Handbook ↩

16. See UK: Consolidation remains controversial, Special Report, IPE ↩

17. See table 1.19 of DC trust scheme return data 2020 - 2021, The Pensions Regulator ↩

18. See figure 3.12 of The Purple Book 2020 (ppf.co.uk) ↩

19. Law Commission-Intermediated Securities- Scoping Paper para 2.65 ↩

20. See page 26 below for instance and AMNT review into asset managers’ voting policies and practices ↩

21. Hymans Robertson suggest costs of c10-20bps for a switch of £150 million of global equities depending on the approach taken, up to £300,000. Whilst some costs would fall for a smaller transaction, others such as advisory fees are not linked to size and would rise as a percentage. For a £10 million portfolio costs might be more than 25bps. ↩

22. Law Commission - Intermediated Securities - Scoping Paper - para 2.26 ↩

23. See paragraph 30 of schedule 3 to the Occupational and Personal Pensions (Disclosure of Information Regulations), S.I. 2013/2734. The duties in these regulations in relation to disclosure of significant votes apply to occupational schemes only. ↩

24. https://amnt.org/wp-content/uploads/2019/05/AMNT-final-review-for-FCA-22-May-2019.pdf ↩

25. https://shareaction.org/wp-content/uploads/2020/11/Voting-Matters-2020.pdf ↩

26. FRC Press Release re signatories 6th September 2021 ↩

27. Investment consultants market investigation - GOV.UK (www.gov.uk) ↩

28. Occupational Pension Schemes (Investment) Regulations SI 2005/3378. The Stewardship policy is covered in regulation 2(3)(c) (for the SIP) and regulation 2A(1)(b) (for the default SIP). ↩

29. https://www.plsa.co.uk/Policy-and-Research/Document-library/PLSA-Stewardship-Voting-Guidelines-2021 ↩

30. http://redlinevoting.org/download-red-lines/ ↩

31. See for example from the PLSA at https://www.plsa.co.uk/Portals/0/Documents/Made-Simple-Guides/2019/ESG-Made-Simple-2019.pdf ↩

32. See page 38 ↩

33. Significant vote reporting is required pursuant to the transposition of the Shareholder Rights Directive into UK law. The Directive itself did not define “significant”. ↩

34. PS19/13: Improving shareholder engagement and increasing transparency around stewardship, FCA ↩

35. Securities and Exchange Commission Mutual Fund Proxy Voting Records and Policies 2005 ↩

36. Open Standards principles - GOV.UK (www.gov.uk) ↩

37. FS21/7: Open finance – feedback statement, FCA ↩

38. The Occupational Pensions Stewardship Council is a body set up by DWP to promote and facilitate high standards of stewardship of pensions.https://www.gov.uk/government/groups/occupational-pensions-stewardship-council ↩

39. Investing with Purpose: report of the Asset Management Taskforce – recommendation 6. ↩

40. AMX ↩

41. Tumelo ↩

42. Citizenshareholders ↩