Wrap Text
Informative Notice – 20210831 White Paper Index Capping Methodology
To: Market Participants
From: FTSE/JSE
Date: 31 August 2021
Subject: White Paper: Index Capping Methodology
1. Introduction
FTSE/JSE is currently considering the most appropriate capping methodology to apply in its indices that is both
aligned with principles of good benchmark design, and also relevant for local investors. Market participants have
been consulted on a proposed amendment to the capping methodology that introduces the concept of Group
Entity Capping. This approach considers a company whose valuation is almost exclusively dominated by its holding
in another listed entity in the same benchmark, and where the combined benchmark weighting of the two
companies is a concentration consideration. In this scenario, the proposal is to consider the two companies
collectively when applying index weight capping.
This white paper is provided to share the key feedback from the consultation process with market participants,
complemented by appropriate market data and considerations from FTSE/JSE. FTSE/JSE acknowledges that the
issue under consideration is considerably more complex than it may appear on the surface, and raises several
questions around good benchmark design, public engagement and procedure, short term impacts and unintended
consequences.
FTSE/JSE has implemented an interim solution in the Capped SWIX Indices from the September 2021 index review.
This approach is necessary to allow the responses from the consultation process to be fully considered and
debated before a permanent solution is finalised, as well as to allow the market to re-calibrate after the exchange
offer transaction.
2. History of the Capped SWIX Indices
The FTSE/JSE All Africa Index Series was introduced in 2002 as a replacement to the existing JSE Actuaries Indices
of the time. This heralded the start of the South African index partnership between FTSE Russell and the JSE, and
introduced several modernisation elements to the indices, such as free float weighting, hierarchical sector
classification, and enhanced data products. The indices are intended to provide benchmark and investable index
products to the South African investment ecosystem.
In 2004 FTSE/JSE launched the Capped All Share Index and the Shareholder Weighted (SWIX) All Share Index. The
existing All Share Index exhibited weight concentrations in some counters and sectors, and the two new
approaches were co-created with market participants to provide performance benchmarks that were better
immunised from concentration risk. Over time, the SWIX approach was largely adopted by institutional managers
and investors as the local benchmark of choice.
Page 1 of 17
In 2016, FTSE/JSE launched the Capped SWIX All Share Index. This was triggered directly by requests from the
South African fund manager and asset consultant community, who were concerned with the increasing
concentration exposure to Naspers in the SWIX All Share Index. The index was launched to serve the benchmark
requirements of South African active general equity managers and was accompanied by a Top 40 tradable version
to support hedging and other product requirements. Over the course of 2017 and ahead of the Prosus listing in
2019, a significant number of institutional investors, asset consultants and fund managers moved their
performance benchmark to this new index, and the index is currently seen as the de facto performance benchmark
for South African general equity investors.
In September 2019, Prosus listed on the JSE and was added to the FTSE/JSE indices. The chart below shows the
progression of both the capped and uncapped SWIX index weights of Naspers and Prosus since the launch of the
Capped SWIX All Share Index:
40%
Naspers and Prosus Weighting in the FTSE/JSE SWIX All Share
35%
30%
SWIX All S ahre Index Weighting
25%
20%
15%
10%
5%
0%
Aug '20
Aug '16
Aug '17
Dec '17
Aug '18
Aug '19
Aug '21
Dec '16
Jun '17
Jun '18
Dec '18
Jun '19
Dec '19
Jun '20
Dec '20
Jun '21
Oct '16
Oct '17
Oct '18
Oct '19
Oct '20
Feb '21
Feb '17
Apr '17
Feb '18
Apr '18
Feb '19
Apr '19
Feb '20
Apr '20
Apr '21
Naspers (Capped) Prosus (Capped) Combined (Uncapped)
The implementation of the Prosus exchange offer for Naspers shares in the indices has made a large structural
change to index weightings, as holdings have shifted from Naspers into Prosus. It is essential to note that this shift
occurred across all market holdings in aggregate due to the corporate action, and that the index weights have
been adjusted on 18 August to align the index to the underlying market holdings. In other words, there was no
index rebalance triggering a trading event and subsequent portfolio rebalance. Rather, the corporate action
caused a change in portfolio weights that was then reflected by an index update.
At the close of trading on 24 August, the Capped SWIX All Share Index reflected index weights of 4.1% for Naspers
and 6.0% for Naspers, compared to 8.0% and 1.4% respectively immediately prior to the corporate action.
Page 2 of 17
Capped indices are designed to mitigate concentration concerns for index users. Capping is applied on a quarterly
basis for each listed company, allowing intra quarter weights to rise above the cap or fall below it, before being
reset to the capping level at the next quarterly review. This comes at the cost of increased turnover in the index,
as capping levels are reset each quarter, resulting in additional index rebalance churn. In quarters where the large
(capped) stock outperforms the index, the position is decreased for the next quarter, and conversely increased at
the end of a quarter where the stock underperforms the index. This is highlighted below for the Top 40 index, but
a similar pattern is observed in the All Share indices. On average, review churn (or one-way turnover) of the
Capped SWIX 40 is 2.9%, compared to 1.8% in the uncapped index.
6%
Review Index Churn
5%
4%
3%
2%
1%
0%
Mar Jun Sep Dec Mar Jun Sep Dec Mar Jun Sep Dec Mar Jun Sep Dec Mar Jun
'17 '17 '17 '17 '18 '18 '18 '18 '19 '19 '19 '19 '20 '20 '20 '20 '21 '21
Top 40 Churn SWIX 40 Churn Capped SWIX 40 Churn
Prosus is classified as a foreign company for index purposes. As such, its index free float is decreased to reflect
only those issued shares that are dematerialised on the South African share register, managed by the Central
Securities Depository, Strate. Equity shares are fungible and can be transferred between the South African register
and the Amsterdam register through an administrative process. From its September 2019 listing there has been a
steady monthly movement in shares from the local register to the European, except for July 2021.
12%
Prosus Holdings on Strate
10%
8%
6%
4%
2%
0%
Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul
'19 '19 '19 '19 '20 '20 '20 '20 '20 '20 '20 '20 '20 '20 '20 '20 '21 '21 '21 '21 '21 '21 '21
Page 3 of 17
3. Consultation Process
In September 2019, Naspers transferred its international assets into a new listing called Prosus, with a primary
listing on Euronext Amsterdam and a secondary listing on the JSE. Various market participants raised concerns
with FTSE/JSE regarding the resultant combined weight of Naspers and Prosus in the FTSE/JSE capped indices, and
particularly the FTSE/JSE Capped SWIX All Share Index. Two local market workshops were scheduled to further
discuss the issue and elicit views from index users.
On 5 November 2019, FTSE/JSE issued a market consultation providing background information on this issue, and
introducing the concept of the proposed group entity capping approach. This was a proposal to cap any two index
constituents jointly when their combined weights exceeded 10%, and the valuation of the parent company was
dominated by the value of its holding in the subsidiary company.
Following a consideration of the feedback received from this market consultation, FTSE/JSE announced in
December 2019 that the proposed amendment to the capping rules would not be implemented at that stage. This
notice also highlighted “the scenario that the combined weights increase significantly as a result of market
movements or future corporate actions”, which would be particularly relevant in the context of the primary
purpose of the FTSE/JSE Capped SWIX All Share Index.
Further to that announcement, some market participants have challenged the validity of the process that was
followed in coming to this decision. In particular, the key arguments against the implementation of the proposed
group entity capping were not published by FTSE/JSE and were therefore not subject to public debate.
The announcement of the Prosus exchange offer for Naspers shares triggered renewed interest from market
participants in the construction of the capped indices in general, and the concerns about the combined weighting
and corporate action impacts referenced in the December 2019 announcement in particular. Several market
participants and an industry body submitted new motivation to FTSE/JSE to review the decision taken in 2019. As
such, FTSE/JSE requested comment from market participants on whether the capping methodology required
amendment.
On 20 August, FTSE/JSE announced that it would apply an interim solution in the FTSE/JSE Capped SWIX All Share
and Top 40 indices, reducing the company capping level from 10% to 6% with effect from the September 2021
index review.
This white paper provides the key arguments received from both consultation processes, with supporting content
from bilateral discussions, internal considerations, and market data. It further details how the consultation process
will proceed during the interim solution period.
FTSE/JSE is grateful to all interested parties that have engaged on this matter. There have been many written
submissions received in response to the requests for comment, and in many cases market participants have taken
considerable time to draft submissions that are of exceptional quality. This has been supported by bilateral
discussions.
Page 4 of 17
4. Responses to the Consultation
A total of 38 formal submissions have been received in response to the 2021 consultation. This section provides a
numerical summary of the response to the question: “Do you believe that Group Entity Capping should be applied
in the FTSE/JSE Capped SWIX indices (J433 and J430)?”
4.1 Responses against the group entity capping proposal
FTSE/JSE received 15 responses against implementing group entity capping in the SWIX:
Category Count J433 AUM (Rbn)
Offshore Manager 7 R0
Local Bank 1
Local Fund Manager 5 R40bn (1 did not disclose)
Listed Company 1
Private Investor 1
There were no offshore investors or managers that were in support of the proposal. None of the offshore
respondents reported being subscribing clients of the FTSE/JSE Capped SWIX All Share Index or having any funds
that are directly benchmarked against this index. The AUM number requested in the consultation was in direct
reference to assets that are benchmarked against or tracking the Capped SWIX All Share Index.
Three of the five domestic managers stated in their response that they believe Naspers and Prosus are essentially
the same entity or share the same underlying assets but voted against the proposal on other grounds.
Furthermore, two respondents against the proposal had colleagues in their broader organisation that submitted
a response in support of the proposal.
4.2 Support for the group entity capping proposal
FTSE/JSE received 21 responses that supported implementing group entity capping in the SWIX:
Category Count J433 AUM (Rbn)
Local Fund Manager 15 R540bn (1 did not disclose)
Asset Consultant 2 n/a
Retirement Fund 1 R19bn
Local Bank 3 R47bn (2 did not disclose)
19 out of the 21 supporters are licensed subscribers to the Capped SWIX All Share Index. The combined supporters
for the proposal represent around R600bn worth of assets that are either benchmarked against the Capped SWIX,
or track it directly. Two respondents that supported for the proposal had colleagues in their broader organisation
that submitted a response against the group entity capping.
4.3 Unanswered Responses
FTSE/JSE received 2 responses that did not directly answer the question, both from local managers. However, both
respondents were in favour of capping the joint exposure in principle.
Page 5 of 17
5. Feedback from the Consultation and Market Engagement
This section of the white paper summarises the key arguments both in support and against the proposal for group
entity capping. It also includes some more generalised comments that are relevant to the broader discussion on
capping and benchmarks. This section is intended to represent the various submissions received from market
participants and does not necessarily reflect the views of FTSE/JSE.
5.1 Alignment of portfolio weights to benchmark weights
Almost all respondents agree that active fund managers can take active positions in individual stocks that are
either higher or lower than prevailing benchmark weights. There were however a range of views received
regarding:
• The extent to which the South African market in general “hugs” benchmark weights and are prepared
to take portfolio positions based on economic fundamentals rather than benchmark weights
• The proportion of South African investors that are benchmarked against the FTSE/JSE capped indices
• The extent to which market holdings in the long run can be driven by changing benchmark methodology
and weights vs. market forces transitioning to a benchmark that represents investor philosophy
A prevailing theme from respondents that do not support the proposal is that a change to index methodology will
directly result in a change to domestic portfolio holdings, causing a direct and materially negative impact to all
Naspers and Prosus shareholders.
5.2 Impact of the Prosus exchange offer
The implementation of the Prosus exchange offer to Naspers shareholders has already structurally changed the
aggregate position of Naspers investors, including those that are benchmarked against a capped FTSE/JSE index.
In short, aggregate holdings of Naspers shares have decreased by 45%, with a resultant increase in Prosus holdings.
FTSE/JSE indices have already been adjusted to reflect this new reality.
Any decision to not implement the proposed group entity capping approach would result in index weights that are
not reflective of this reality after the exchange offer, as index users are required to significantly increase their
portfolio holdings of Naspers in order to maintain their prevailing active weights (either under or over).
Respondents that believe the underlying economic assets of the group have not fundamentally changed object
strongly to this artificial index reweighting. In other words, any decision by FTSE/JSE to not implement the proposal
will lead to artificial changes in portfolio allocation.
5.3 Naspers and Prosus as different companies
There is no dispute that Naspers and Prosus are two separately listed companies in the legal sense. They have
different identifiers, registrations, shareholders, assets and liabilities. In index terms, they continue to be treated
as separate companies for index selection purposes – they are ranked independently and count as two distinct
constituents in the Top 40 Index. This index principle has remained unchanged since the Prosus listing and is not
currently being debated.
Page 6 of 17
Naspers invests in several companies, and lists Prosus, Media24 and Takealot.com as the top 3 on its website.
Naspers currently holds 57.2% of the share capital in Prosus and is its largest shareholder. Since both Naspers and
Prosus are listed on the JSE, it is possible to compare the market valuations of its Prosus shareholding to the
market valuation of Naspers itself on 25 August 2021:
• Market Capitalisation of Naspers: R1.1bn
• Market Capitalisation of Prosus x 57.2%: R1.5bn
As such, the valuation of the Naspers stake in Prosus exceeds the entire valuation of Naspers itself. Furthermore,
Naspers itself is not a meaningfully diversified holding company as confirmed by the dominance of its Prosus
valuation. Ultimately, an investment into either Naspers or Prosus gives the investor exposure to the same pool
of assets (except for Media24, Takelot.com and others). The companies describe this as Naspers having a 40.3%
exposure to the group’s underlying assets, and Prosus shareholders the balance of 59.7%. The Prosus asset base
is therefore represented in both the Prosus listing and the Naspers listing.
Any investor that has a hypothetical mandate to restrict exposure to Prosus to a fixed level of say 10%, would
therefore reasonably include both its Naspers and Prosus holdings in this calculation, despite them being separate
listed companies.
Respondents have made additional supporting comments about the commonalities between the companies,
including overlap of the management team and board of directors, as well as the fact that the 49.5% and 57.2%
crossholding strengthens the interconnectivity of the two companies.
One critical differential between the two companies stems from their geographies. Naspers is South African
domiciled and incorporated, subject to South African legislation, taxation and sovereign risk profile, and priced in
South African Rand. It is typically profiled as an emerging market company by global investors, and its investor
base is primarily emerging markets investors and South African investors. Prosus on the other is domiciled and
incorporated in The Netherlands, primary listed on Euronext, subject to European regulation, taxation and
sovereign risk profile. It is included as a constituent in developed market indices and tracked by the majority of
active and passive funds focusing on Europe, many of whom do not have a mandate to invest in South Africa. As
a result, it is majority owned by European and global investors.
As a general rule, global investors hold Naspers on an emerging markets mandate, whereas they hold Prosus on a
developed markets mandate. One result of this is that global investors are unlikely to hold both Naspers and Prosus
in the same portfolio for fund, and those funds that do hold both are not holding them at levels that cause
concentration concerns.
A final differentiator between the two companies speaks to the historic price performance. While the stock price
returns are very strongly correlated for long periods, there are significant periods where there is divergence
between the two share price movements.
Page 7 of 17
5.4 Management of concentration risk through capping
The consultation process has focused on capped benchmarks only. Given the separated nature of the two
companies, they are treated as separate companies in the uncapped indices and there is currently no debate on
the combined weight of the two stocks in the uncapped indices. The application of free float rules regarding
constituent cross holdings already prevents any double counting in this regard.
The Capped SWIX All Share Index in particular was created in 2016 as a direct response to the increasingly
dominant weight of Naspers in the SWIX All Share Index, and at the request of local managers using the indices as
a performance benchmark. Investors, consultants and managers drove a voluntary benchmark shift to this new
index, and it became the de facto performance benchmark for local managers long before the Prosus listing.
Many index users were unhappy with the way that the combined weight structurally increased in this index beyond
the 10% target on the initial Prosus listing, seeing this as contrary to the initial purpose of the index. In effect, the
capitalisation issue and Prosus listing increased the weight of the underlying Naspers/Prosus assets in the
benchmark without any corresponding change to fundamentals of those assets. Similarly, a further structural
increase of the combined weights to the 16%-18% range is further considered to be contrary to the specific
purpose of this index. A majority of respondents feel that a combined weight at this level would contradict the
purpose of a capped alternative to the SWIX All Share Index, thereby undermining its relevance as a benchmark
and driving a market transition to a different benchmark that is fit for purpose.
Supporters of the proposal therefore do not see it as a new change to the capping methodology. Rather, they
position that the existing methodology must be reviewed to ensure that the index continues to deliver according
to its original stated and agreed purpose.
Capping is intended to restrict index exposure to a single listed company, regardless of how diversified that
company itself may be. This is often positioned as a risk management tool to limit portfolio losses in the event of
the failure of a single company. It is in this context that the proposal is supported – a rapid and material
devaluation of Prosus is almost certain to also trigger a corresponding devaluation of Naspers given its 40.3%
exposure to the group’s underlying assets. This would not be an index capping concern if the combined weighting
of the two companies were not higher than the prevailing capping limit, or if Naspers itself was materially
diversified.
5.5 Group Entity Capping is not based on global best practice
The FTSE/JSE Africa Index Series is designed to be locally relevant for South African investors, and as such has
some idiosyncrasies when compared to global benchmarks. The two most relevant are the inclusion of foreign
companies and the resultant inconsistent weighting of foreign shareholders.
Some listed companies consider a dual listing structure, with single or multiple secondary listings on exchanges
other than their primary one. It is common practice for index providers to only include the company in the country
or benchmark index that captures their primary listing or domicile. This prevents double counting of the company
when looking across multiple benchmarks or sub-benchmarks. FTSE/JSE applies the non-standard treatment of
Page 8 of 17
considering all JSE-listed equities in its benchmark indices. This means that Prosus is eligible for FTSE/JSE index
inclusion, despite this being contrary to global best practice.
However, the weighting of foreign companies in the FTSE/JSE indices is based on the market capitalisation that is
held on the South African share register, and not the full global free float. Therefore, the index weighting of foreign
companies represents the local share register only, which is predominately domestic investors in the case of
foreign companies. Domestic companies are however included at their full market capitalisation, with their index
weight representing both domestic and international investors. In the case of Naspers, which is nearly 60% held
by non-residents, this creates a weighting misalignment in the indices, since its uncapped index weighting does
not truly reflect the aggregate portfolio holdings of domestic investors.
In practice, while the SWIX All Share weight of Naspers averaged 25% in 2020, the actual average All Share portfolio
weight considering only South African resident investors in aggregate was 16%. This differential is an essential fact
since the uncapped SWIX All Share weighting is therefore not reflective of the typical domestic investor’s holdings.
It also confirms that view that aggregate portfolio holdings in practice are not driven solely by capped index
weights.
There are several global precedents to be found of single asset investment companies being listed on the same
exchange as the underlying company in which they invest, and yet there are no other examples of two companies
of this nature being treated as a single entity for index capping purposes. However, FTSE/JSE has not found any
examples of this phenomenon occurring where the primary local benchmark is capped, both companies are
benchmark constituents, and the combined weight of the two companies exceeds the index capping level.
The Naspers and Prosus corporate structure is of itself a fairly unique structure by global standards. It is uncommon
to have two listed companies on the same exchange that provide such a comparable entry point into the same
basic assets. It is even particularly rare for this scenario to occur in a counter that so clearly dominates the domestic
benchmark in terms of market capitalisation.
As such, the implementation of a group entity capping approach appears to be a unique solution in comparison to
global index standards. However, it appears equally that the underlying concern that has triggered the proposal is
itself a unique scenario is comparison to global stock markets.
5.6 Index capping will cause value destruction for shareholders
Arguably, FTSE/JSE has allowed the combined weight of Naspers and Prosus to climb to levels beyond 10% due to
the December 2019 decision taken on capping. A sudden shift of this level back to 10% could cause short term
shocks in the market and temporary value destruction. Similarly, allowing the combined index weights to step up
to 16%-20% have the potential to cause similar value destruction for all other stocks in the market, as managers
sell off their portfolio to fund the increase in Naspers and Prosus weightings.
It is estimated that 10%-15% of local market activity is driven by pure index tracking. Not all funds are index
trackers, and not all funds are even benchmark cognisant. Furthermore, there are multiple indices that are widely
referenced in any given market, including South Africa.
Page 9 of 17
Respondents argue that capping destroys value for shareholders for a stock that is consistently growing in value.
As the stock grows, the index weighting cap is regularly breached driven an index rebalance and consequent forced
down-selling of the stock by index users. This value destruction impacts all shareholders of the stock, and not just
index users. It must be recognised that this is not a criticism of a particular capping level, but of the capping
principle itself – the same phenomenon would occur with a capping level set at 5%, 10% or 15%.
As an extension of this concern, local investors, including retail and pension fund investors, would be restricted
from participating in a future Naspers/Prosus growth story if they need to continually trim their holdings down to
10% of their portfolio. This could result in a sub-optimal outcome in terms of collective wealth generation. On a
cumulative annual basis, the wealth of South Africans invested in capped index tracking products will greatly be
reduced if Naspers continues to innovate and grow, whilst having their exposure capped.
A further extension of this argument by some respondents is that implementation of the proposal will lead to
changes in portfolio allocation resulting in a net outflow of multi-billion dollars from Naspers and Prosus, to align
to a synthetic index that grossly underweights and thus hamstrings the realisation of the full value of Naspers and
Prosus.
Of course, no return comes without some risk, and past performance should not be used as an indicator of future
performance. Risk-aware investors may look to diversify their portfolios away from single stock exposure to limit
the maximum amount of risk from a single point of failure.
Local managers report that they switched to a capped benchmark precisely due to concerns around increasing
concentration risk in Naspers. Portfolio management resulted in the application of a limit on their Naspers
portfolio weights. This flips around the causation concept - that the introduction of the Capped SWIX All Share
was the reason that managers decided to decrease their holdings in Naspers. Rather, some domestic managers
have a natural ceiling on their single stock concentration and identified a benchmark that reflected this philosophy.
While some respondents believe that the introduction of the Capped SWIX All Share Index created selling pressure
and contributed to the widening of Nasper’s discount, FTSE/JSE is not aware of any research that concludes the
introduction of the index itself caused a forced and sustained sell-off of Naspers shares for local investors. In fact,
there is evidence that aggregate domestic portfolio holdings in Naspers increased steadily from 2013 to 2017.
5.7 Index impacts are broader than just index users
Index data and licensing rights are delivered as a commercial service rather than a regulated function. The JSE
does not require any market participant to use a performance benchmark, nor does it stipulate which index should
be used as that benchmark. Market participants are free to choose the index that best fits their investment
mandate or strategy, and they do so.
A commonly held view is that those clients that have chosen to utilise a particular index should have a voice when
changes to the index methodology or purpose are considered. Many believe that those entities that are not direct
users of the index should not have an equal voice and are potentially conflicted in their views. Conversely, there
is a view that the proposed change could benefit a few investors to the detriment of many, and that all investors
Page 10 of 17
should be considered in managing the indices and decisions regarding them should be made in the best interest
of all investors and market participants.
This is even more important if there is a risk that an index methodology decision could be materially prejudicial to
international investor interests in South Africa. Even if they are not directly benchmarked to it, an unexpected
change in domestic index methodology may put foreign investors off from investing into Naspers and Prosus.
One challenge to FTSE/JSE is how to effectively balance the demands of all stakeholders. The aggregate investment
behaviour or risk preferences of local fund managers will have an impact on asset prices and liquidity which may
result in positive or negative outcomes for other market participants. While investor activity may be driven by
index construction in the short term, in the longer term this is not true. Market indices measure aggregate market
holdings and performance, and these are driven by the collective actions of all participants who will ultimately
select benchmarks that are most closely aligned to their own strategy.
5.8 Market dynamics will naturally decrease the weight of Prosus on the JSE
When Prosus listed, nearly 10% of issued shares were registered on Strate, resulting in a SWIX All Share Weight of
3.3%. Since then, there has been evidence of a gradual transfer of shares between the South African and European
registers, with shares moving off the Strate register on a consistent basis. The Strate holdings prior to the exchange
offer represent 3.6% of the total shares issued. As Prosus has seen share price appreciation over time, this
counteracting decrease in free float has put downward pressure on the index weighting.
It is expected that this trend will continue after the exchange offer completes. This creates the risk that an index
adjustment now will structurally decrease the weight of Naspers and increase the weight of Prosus compared to
historic (pre exchange offer) levels, but over time the Prosus weight still steadily decrease due to the register
dynamics. This may ultimately trigger a situation where the Naspers index weight has been artificially lowered,
without the corresponding and appropriate Prosus weight contribution.
5.9 Policy uncertainty may drive away foreign investment on the JSE
The group entity capping proposal was considered recently in December 2019, only 18 months ago. Global
investors who are not in tune with the local asset management industry may therefore feel surprised an issue
previously considered to be closed has been reopened for debate so soon. However, local industry participants
will appreciate that the December 2019 announcement was poorly received by many in the market, and that it
has been a topic of regular debate since then. In addition, the December 2019 announcement explicitly highlighted
the scenario that the combined index weights could increase significantly as a result of a future corporate action.
It must also be stressed that index administration is not a regulated function of the JSE, nor is it a regulated
financial service in South Africa, and that a number of providers calculate indices on a competitive basis in South
Africa. FTSE Russell is the benchmark administrator of the FTSE/JSE Africa Index Series, as defined in the European
Benchmark Regulation, and the index series complies with the recommendations made by the International
Organization of Securities Commissions (IOSCO), as laid out in the Principles for Financial Benchmarks. FTSE
Page 11 of 17
Russell has an established reputation for transparent, robust, rules-driven index construction methodologies and
is committed to leading global best practice standards in index governance.
Some respondents raised concerns that the implementation of the proposed group entity capping could lead to a
severe loss of trust by international investors resulting in material outflows from the South African capital market.
Certainty is an important tenet of index rules and shareholders have taken positions based on the existing rules
and on the basis that the publicly communicated treatment of Prosus and Naspers in the FTSE/JSE indices will
persist. Introducing a new rule will create a view of indexation in South Africa in that policy certainty is
unpredictable, and a progression of consultations or reversals of historical decisions would have negative
implications towards the JSE/FTSE for the credibility or assuredness of decisions being made.
5.10 Basing index methodology on a single company is bad practice
In general, index methodology decisions should not be based on individual companies. Changing index
methodology purely in response to a specific company’s situation should ideally be avoided. However, Naspers
and Prosus are a unique aspect of the South African equity market, both in terms of their corporate structure, and
their dominant market capitalisation. Any discussion around their valuations, benchmark treatment or corporate
action impact will have a proportionately large impact on all market participants. This makes it difficult to optimise
index methodology in a purely generic way, without due consideration for this local market dynamic that is
materially impactful on any local portfolio.
By transferring its assets into an offshore company, Naspers may have set a precedent that other companies could
consider following in the future, particularly if they experience rapid growth to an outsized position on the
exchange.
5.11 The exchange offer will have a structural impact on global index weights
Most global investors that have a South African or Emerging Markets mandate would not use the FTSE/JSE Africa
Index Series as their benchmark. Since around 55% of Naspers is held by non-residents, it is relevant to consider
the behaviour of this investor segment.
As a result of the corporate action structure, the free float of Naspers in global emerging markets and South Africa
indices will fall by nearly half, with a resultant large drop in index weight, and with no commensurate increase in
weighting for Prosus. This will cause a structural shift in the demand for Naspers from global investors, potentially
driving value destruction for shareholders. This impact is likely to be significantly larger in the long term than any
capping methodology impact given the comparably larger shareholder base using global indices compared to the
Capped SWIX All Share. The impact may be offset by benchmark incognisant global investors who are attracted to
Naspers.
The decreased benchmark weight of Naspers for global investors is likely to be accompanied by a corresponding
decrease in the total index weighting of South Africa, particularly in emerging markets benchmarks.
Page 12 of 17
5.12 Proposals to amend capping are driven by performance considerations
Given the importance of Naspers/Prosus to the performance of most funds that are benchmarked against the
capped indices, it is essential to recognise that strict limits around benchmark weights will not apply equally to all
funds. The ability to overweight the benchmark may therefore allow some funds to keep outperforming
competitors who will have to reduce their exposure to some of the best performing assets on the exchange.
In a similar vein, capping the weight of stocks that have historically outperformed the rest of the market serves to
artificially lower performance hurdles for active managers, making is easier to earn performance fees. If market
participants believe that Naspers and Prosus will consistently outperform the rest of the JSE market over the long
term, it would be self-serving to keep their benchmark weights as low as possible but their actual portfolio holdings
as high as possible. This would help them to create “synthetic alpha” and boost performance fees, raising the risk
that the proposal will benefit the asset management industry at the expense of the end investor.
Other respondents noted that benchmark selection is typically not in the hands of the asset manager, particularly
for institutional mandates, but rather is allocated by the asset owner, often supported by an asset consultant.
5.13 Impacts to capped indices and uncapped indices should be aligned
One of the intended outcomes of the exchange offer as indicated by Naspers and Prosus is to reduce the weight
of Naspers on the JSE and to increase the index weight of Prosus. This is playing out as intended in the uncapped
indices. However, in the absence of any changes to capping methodology, in the capped indices the weight of
Prosus increases significantly while the weight of Naspers remains at or near the 10% cap level. This overall
increase in benchmark weighting is not linked to any fundamental drivers or aligned with the actual economics of
the exchange offer for the investor but is rather a technical consequence of the interaction of technical index rules
and the corporate action design. Put differently, an existing Capped SWIX All Share Tracker that tendered and
received the 45.33% on offer, would need to trade on the open market to re-purchase all the Naspers shares that
it had just exchanged for Prosus shares, selling down the rest of the diversified portfolio to fund this.
5.14 All capped indices should be treated the same
The overwhelming response from the consultation was that all capped indices should be subject to the same
capping methodology. There was extremely limited support to have a different approach applied to the Capped
SWIX indices compared to the Capped All Share and Capped Top 40 index.
5.15 Reducing the capping factor impacts other companies
Reducing the capping factor to a lower level and continuing to cap Naspers and Prosus separately will have
unintended consequences on other index constituents. For example, a permanent shift to a 6% cap at a company
level will result in the Naspers and Prosus group weight having an upper cap of 12%, but all other index
constituents having a cap of 6%. In a scenario where Naspers and Prosus are in practice considered jointly for
concentration risk purposes, this approach would unfairly restrict index weightings for growth companies in
comparison to Naspers and Prosus.
Page 13 of 17
5.16 10% Capping is not the right level
The single stock concentration limits speak to 10% for CIS funds and 15% for pension funds. As detailed in
paragraph 5.5, the average domestic investor in 2020 held 16% of Naspers in their All Share portfolio, and many
individual funds that benchmark the Capped SWIX All Share Index hold Naspers weights in excess of the 10%
benchmark weight. There was some support from respondents to increase the base capping level of all FTSE/JSE
capped indices from the current 10% base to a level between 12% and 15%. Respondents observed the 15% limit
highlighted in Regulation 28 in support of this argument.
One counter to this is that a fair benchmark should allow investors to be overweight on stocks they are bullish on,
and underweight on stocks they are bearish on. If there is a regulatory cap in play that prevents portfolio weights
exceeding 15%, then a benchmark weight at 15% would preclude the manager from acting on their convictions
for that particular stock. Other managers may also have internal risk appetites around concentration that are more
prudent than the regulatory maximums, although this should not drive benchmark design.
Several respondents that supported the proposal were strongly opposed to a change to the 10% capping level.
Any change would be arbitrary in nature and would render the historic index track records unusable.
5.17 Index rules should not be arbitrary or ambiguous
The index capping levels of 10% are a long-standing convention in South Africa. However, there can be no research
that unequivocally confirms that 10% is the optimal level for all classes of investor. Some investors have indicated
a preference for a lower cap such as 8%, and some for a level as high as 15%.
One solution to this problem is to introduce a range of indices, each with a different capping level. Users would
then be free to select the benchmark that is most aligned with their investment strategy or mandate. However,
the JSE does believe that a widely accepted performance benchmark is a sign of a healthy local market since it
improves comparability and transparency. Furthermore, the speed with which such a large portion of the local
market changed from an uncapped to a capped benchmark confirms the underlying support for the approach.
A 10% capping level is not necessarily fully representative of aggregate holdings by all investors in the market – a
common test for benchmark replicability and representativeness. At the same time, the historic uncapped Naspers
weight is also not a true reflection of aggregate portfolio holdings. This is because the index construction excludes
foreign shareholders in dual listed stocks (held offshore) from the index, while including foreign shareholders of
JSE primary listed stocks such as Naspers. This creates a tendency to overweight local stocks, especially those that
see global investor demand through globally-benchmarked investing. The opposite effect is seen for a foreign
company like Prosus, which would have typically have an aggregate domestic portfolio weight greater than its
uncapped SWIX weighting.
Respondents feel that the wording of the group entity capping rule should be carefully reviewed to ensure that it
is unambiguous, and as far as possible future proof. Certainty in index rules and their implementation is essential
to the integrity of benchmark provision. As such, the 75% and 65% hurdles around which the proposed rule is
Page 14 of 17
constructed can also be challenged as arbitrary. One suggestion from the consultation was to review these levels
on an annual or ongoing basis to ensure their ongoing relevance.
Some respondents requested that any references to “parent” and “subsidiary” be removed from the rule wording.
The intention of the rule is to manage concentration in a way that reflects the economic value of the constituents’
derived value, and this is independent of whether the relationship between the two entities results in control.
6. Interim Solution
FTSE/JSE has implemented an interim solution in the FTSE/JSE Capped SWIX All Share Index and the FTSE/JSE
Capped SWIX Top 40 Index, reducing the capping level from 10% to 6% for all constituents, with effect from the
September 2021 review. The principle behind the interim approach is two-fold:
• To allow effective and fair procedure and engagement regarding the implementation of a more
permanent solution
• To allow the Naspers/Prosus transaction to fully settle in the market, and market positions to reach
their new equilibrium, without undue influence or dislocation caused by changing benchmark weights
This white paper and the continued consultation process detailed in the Next Steps paragraph will deliver on the
first principle. With respect to the second, any index action that significantly changes the benchmark weighting of
the Naspers and Prosus complex either above or below the prevailing levels is a concern, particularly in the
absence of associated underlying economic drivers. Such an index change may result in artificial forces impacting
equilibrium prices.
6.1 Capped SWIX Indices
When developing a view on “prevailing” weights, consideration must be given to:
• The local register holdings of Prosus have steadily decreased since listing, thereby impacting index free
float and weighting. This appears to have stabilised at an equilibrium level from April through June 2021
• Recent price volatility as a result of China volatility as well as the transaction itself have impacted price
levels of the two stocks
Considering the four-month period before the Prosus exchange offer of April to July 2021, the combined weight
of Naspers and Prosus in the FTSE/JSE Capped All Share Index ranged between 10.2% and 13.2%, with an average
and median weight of 11.5%. The table below shows the weight progression over key dates:
Date Naspers Prosus Combined
31 December 2020 9.4% 3.0% 12.4%
01 April 2021 10.9% 2.3% 13.2%
03 May 2021 9.9% 2.1% 12.0%
01 June 2021 8.8% 1.9% 10.7%
01 July 2021 10.2% 1.6% 11.8%
01 August 2021 9.5% 1.5% 11.0%
17 August (pre-offer) 8.0% 1.4% 9.4%
18 August (post-offer) 6.1% 4.3% 10.4%
25 August 2021 (current) 6.1% 4.2% 10.3%
Page 15 of 17
As a result of the application of the exchange offer, the weight of Naspers in the index has already decreased, and
the weight of Prosus increased. This index change reflects the change to aggregate portfolios in the market due to
the corporate action implementation. If the interim approach was not applied, modelling indicates that the
combined index weights would increase to around 15%-16% at the September index review. This would require
those managers who wish to maintain their active weights to effectively buy up the Naspers shares in the open
market that they had previously exchanged through the corporate action.
If there is no change to the capping methodology, then this is the correct approach, albeit somewhat non-intuitive.
However, if the proposed ground entity capping approach is in fact applied in the indices then managers will have
exchanged 45% of their Naspers shares in August, bought them back in the market in September, only to sell them
off again in response to the methodology change. This would result in unnecessary trading costs and market
volatility.
Introducing an individual cap at 6% is expected to result in the index weights for each company to be around 6%
at the September review, with a combined weight of 12%. Modelling the uncapped weights in the SWIX All Share
projects a Naspers weight of 8%-9% and a Prosus weight of 6%-7%.
It is not expected that any other companies will be impacted by the 6% capping level in the SWIX All Share.
However, when considering the FTSE/JSE Capped SWIX Top 40, it is likely that between one and two other
companies may be impacted by the change to a 6% cap, although not at a significant level in benchmark terms.
6.2 Capped All Share Indices
In the FTSE/JSE Capped All Share Index (J303), which is not SWIX-weighted, there are several other companies
besides Naspers that are capped at the 10% level. As a result of this and other index dynamics, the uncapped
weights of Naspers (6%-7%) and Prosus (4%-5%) modelled at the September review are far closer to prevailing
index levels. This means that there is no index dislocation expected at the September review based on the current
10% cap, and it is not necessary to implement an interim solution in this index, or the Top 40 version (J300).
6.3 Other market variables
As the market re-calibrates following the large corporate action, it is ideal that there are no artificial impacts driven
by changing index weights. Two other key variables that must be carefully monitored are:
• The behaviour of global investors into South Africa and Emerging Markets, following the down-
weighting of both Naspers and South Africa in global benchmarks
• Stabilisation around the proportion of Prosus shares that are held on the South African share register,
and therefore the Prosus index free float
FTSE/JSE expects a period of between three and six months following the corporate action is required for the
market to fully re-calibrate.
Page 16 of 17
7. Next Steps
This white paper presents a detailed background to the capping question, qualitative feedback both in support
and against the introduction of group entity capping, and a motivation for the application of the interim approach
to the SWIX indices. Overall, the majority of respondents support the proposed group entity capping approach
and indicate that they consider Naspers and Prosus jointly when evaluating concentration risk in their portfolios.
The majority is significantly higher when only considering direct users of the index.
Market participants are invited to carefully consider all the views expressed in this document. FTSE/JSE further
invites any additional feedback or views on this matter to be submitted by 31 October 2021. Written comments
can be submitted to indices@jse.co.za. In further support of effective market engagement, any interested party is
invited to schedule a bilateral meeting with the team to discuss any of the content of the white paper.
Over and above the central proposal around group entity capping, the following key points require further
consideration:
• Consideration for raising the overall capping level from the current base of 10%
• The most appropriate construction of the potential group entity capping rule to ensure that the wording
is appropriate, unambiguous, and as future-proof as possible
• How to balance the views of direct index users with indirect impacts to other investors
• The perception of regulatory uncertainty around benchmark design and index rules
• Consideration for any unintended consequences of changes to benchmark weights
In addition, FTSE/JSE will carefully monitor the market dynamics over the remainder of 2021, with particular
attention to the geographic split of the Prosus share register.
FTSE/JSE wishes to thank all those that have submitted a response to the request for comments. Market
engagement and world-class index governance are critical components of the FTSE/JSE Africa Index Series, and
will continue to drive any potential future changes to index methodology.
Page 17 of 17
Date: 01-09-2021 08:02:00
Produced by the JSE SENS Department. The SENS service is an information dissemination service administered by the JSE Limited ('JSE').
The JSE does not, whether expressly, tacitly or implicitly, represent, warrant or in any way guarantee the truth, accuracy or completeness of
the information published on SENS. The JSE, their officers, employees and agents accept no liability for (or in respect of) any direct,
indirect, incidental or consequential loss or damage of any kind or nature, howsoever arising, from the use of SENS or the use of, or reliance on,
information disseminated through SENS.