Docstoc

Consumer Panel - Risk Ratings

Document Sample
Consumer Panel - Risk Ratings Powered By Docstoc
					        EU latest proposals for
     “Key Information Document”
           and Risk Ratings

 Key findings of ABI/IMA Research

            23 September 2009


Julie Patterson
Director, Authorised Funds & tax
               What is the KID?
“Key Features” according to CESR:
   A tool to help retail investors make informed investment
    decisions
   Not primarily a marketing document
   Not an investor education document
   Not encumbered with info serving only legal or
    regulatory requirements
   Need not address investors’ information requirements
    on an ongoing basis
   Plain language and enabling comparisons between
    funds
   Generally no more than two pages/sides
           What will it tell you?
Key Elements:
   Names of fund, manager and promoter/group
   Fund objectives and investment strategy
   Material risk/reward factors likely to affect fund
   Past performance
   Charges (payable directly or indirectly by
    investors)
   Treatment of income
In two pages!
                    Risk Rating
The indicator should be:
 Applicable to as many funds as possible
 Based on robust methodology with no room for
  manipulation
 Clear that categorisation does not imply any guarantee
 Simple as possible
 Easy to implement for managers
 Allow easy and effective supervision by regulators
 Be stable and robust against normal changes in the risk
  of capital markets
“The indicator must not be misleading”
      Methodology proposed by CESR
 Historical volatility with use of a VaR measure for
  structured funds
 Based on investment returns over a 3-year period (or 5
  years for funds with returns less frequent than weekly)
 For new funds, use available data extended by a series
  “representative of the way in which the fund would have
  behaved in the past”
 For funds with a markedly asymmetrical distribution, use
  a VaR approach
                      Risk buckets
 A numerical scale of 1-6
 Prescribed bucket boundaries - two options:
   - Split fund universe along risk continuum (possible bunching of
     funds)
   - Split fund universe into equal-sized sub-sets (some buckets v
     large or v small in risk terms)

 Migration rules to minimise bucket hopping – three
  options:
   - No rule
   - “Observation period” (e.g. change has to have held for last three
     months)
   - Different entry/exit thresholds around the bucket boundaries

The BIG ISSUE is frequency of bucket hopping
              ABI/IMA Research
The aim of the project:
• To identify whether there exists a simple and reliable
  measure of investment risk
• To suggest how measure could then be used to
  categorise (bucket) funds
• To suggest what structure would be needed to
  implement the process
• To consider whether the same measure could/should
  be applied to structured products and absolute return
  funds
      What do Managers currently do?
   Fund managers that risk rate their funds do not “return rate” them
   Managers estimate risk inherent in broad asset classes rather than for
    individual funds. Funds are then fitted into the asset class risk spectrum
   The way in which managers assess asset class risk varies greatly, but the
    methodologies are generally based upon measures of historic volatility.
    The results in terms of risk ordering seem to vary less
   Managers also use their discretion based upon their professional expertise
    to adjust the risk order of asset classes that seem to be ‘out of line’
   There is no clear consensus on how to group asset classes into risk
    buckets, or on how many buckets to use. It is perhaps here that there
    seems to be the greatest diversity of approach to risk rating
   Structured products and absolute return funds seen to pose particular
    issues
      The risk measures tested
The research considers:
    Maximum drawdown
    Range
    Standard deviation
    VaR
    Sharpe ratio
    Downside deviation
    Sortino ratio
    Beta
         How reliable are the measures?
                                                          Correlation coefficients using 36 month pre-assessment and
                                                                12 month post assessment periods respectively

                                                         100.0%


                                                         80.0%


                      Spearman's rank correlation test   60.0%


                                                         40.0%


                                                         20.0%


                                                          0.0%


                                                         -20.0%


                                                         -40.0%


                                                         -60.0%
                                                                  1990   1992     1994     1996      1998      2000    2002      2004      2006   2008
                                                                     Stdev      Max draw     Range          Down dev    Sharpe          Sortino   Beta


•   Standard deviation and Range appear to be the most “reliable”, Sharpe and Sortino the worst
•   Under the assumption of normality, the rankings using VaR and Standard deviation will be
    observationally the same. However, standard deviation operationally easier to compute and a more
    commonly used measure. VaR measures very sensitive to assumptions.
                                                                   How to bucket funds?
                                                    40.0

                                                               Stdev           Range
                                                               Max Draw        Down SD
     Risk level, relative to "cash" (full sample)



                                                    30.0




                                                    20.0




                                                    10.0




                                                     0.0
                                                           1   2   3   4   5    6   7   8   9   10 11 12 13 14 15 16 17 18 19 20 21 22 23
                                                                                                   Asset class

The risk continuum is not smooth. Assets in the top sextile are not six
  times more risky than those in the bottom.
                            Risk bucket boundaries
       Standard deviation                        Range                      Max drawdown                  Downside deviation
    Asset class             Bucket   Asset class             Bucket   Asset class             Bucket   Asset class             Bucket
    Money market              1      Money market              1      Money market              1      Money market              1
    UK direct property        1      UK direct property        1      UK Gilts                  1      UK direct property        1
    UK Gilts                  2      UK Gilts                  1      Defensive                 1      UK Gilts                  1
    Defensive                 2      Defensive                 2      UK direct property        1      Defensive                 2
    £ fixed interest          2      Sterling long bond        2      Global fixed interest     2      £ fixed interest          2
    Global high yield         2      Cautious                  2      Sterling long bond        2      Cautious                  2
    Cautious                  2      £ fixed interest          2      Cautious                  2      Global fixed interest     2
    Global fixed interest     3      Global fixed interest     3      £ fixed interest          3      Sterling long bond        2
    Sterling long bond        3      Global high yield         3      Balanced                  3      Global high yield         3
    Balanced                  3      Balanced                  3      Global high yield         3      Balanced                  3
    Flexible                  4      Global equities           4      Global equities           4      Global equities           5
    North America             4      North America             4      Japan                     4      Flexible                  5
    Global equities           4      Europe excl UK            4      North America             4      North America             5
    UK all companies          4      Flexible                  4      Europe excl UK            5      UK all companies          5
    UK equity income          5      Japan                     4      Flexible                  5      UK equity income          5
    Europe excl UK            5      UK equity income          5      UK smaller comps          5      Europe excl UK            5
    UK smaller comps          5      UK smaller comps          5      UK all companies          5      Japan                     6
    Japan                     5      UK all companies          5      UK equity income          5      Commodities/energy        6
    Private Equity            6      Property securities       6      Property securities       5      UK smaller comps          6
    Commodities/energy        6      Commodities/energy        6      Commodities/energy        5      Property securities       6
    Property securities       6      Private Equity            6      Private Equity            6      Private Equity            6
    Asia Pacific excl Jap     6      Emerging equities         6      Asia Pacific excl Jap     6      Emerging equities         6
    Emerging equities         6      Asia Pacific excl Jap     6      Emerging equities         6      Asia Pacific excl Jap     6

The bands are the sextiles
The ordering is similar but differs, and in one case there are no 4s, so
  the choice of metric matters
                                                                                                                                     # of switches
                                                                                                De
                                                                                                     fe




                                                                                                                   0.0
                                                                                                                         2.0
                                                                                                                               4.0
                                                                                                                                       6.0
                                                                                                                                             8.0
                                                                                                                                                     10.0
                                                                                                                                                            12.0
                                                                                                                                                                                                        14.0

                                                                                                         n
                                                                                                    C a s iv
                                                                                                        u e
                                                                                                    B a t io u
                                                                                    UK                 la s
                                                                                                         nc
                                                                                  U K a ll F le e d
                                                                                                c         x
                                                                                        s m o m ib l
                                                                                   UK a pa e
                                                                                                  l
                                                                                         e q le r n ie s
                                                                                               u co
                                                                                       G ity mp
                                                                                           l o in s
                                                                                      E u ba l co m
                                                                                            ro e q e
                                                                                                 p u
                                                                               A s N e e it i e
                                                                                  i a o rt x c s
                                                                                      Pa h A l U
                                                                                           cif m K
                                                                                                 ic er
                                                                                  Em                ex ica
                                                                                                       cl
                                                                                       er                 Ja
                                                                                            gi
                                                                                                n g J ap p
                                                                                                      eq a n
                                                                                                        ui
                                                                                        £                  t
                                                                                  S t f i x U K ie s
                                                                                     e e                  G
                                                                               G r l i n d i n il t s
                                                                                  lo g
                                                                                     b a lo t e r
                                                                                         l           n es
                                                                                     G f ix e g b t
                                                                                       lo d o
                                                                                  U K b a i nt n d
                                                                                                 l h er
                          eighteen year period is high                           P r d i r e ig h e s t
                                                                                     op ct y i
                                                                                                                                                              Number of switches




                                                                                         e r p r e ld
                                                                                             ty o
                                                                               C o M se p er
                                                                                   m o n c u ty
                                                                                     m e y r it
                                                                                        od                  ie
                                                                                                                                                                                                               Risk bucket boundaries




                                                                                              i t ie m a s
                                                                                          P r s / e rk e
                                                                                               iv a n t
                                                                                                    t e e rg
                                                                                                       Eq y
                                                                                                           ui
                                                                                                              ty
                                                                                                                                                              Number of switches greater than 2 categories




Using 3-year calculation period, the number of one category switches over an
          Risk bucket boundaries
   16.0
          3yr   5yr   10yr
   14.0

   12.0
 s 10.0
 e
 h
 c
 t
 i
 w 8.0
 s
 f
 o
 r 6.0
 e
 b
 m
 u 4.0
 N
    2.0

    0.0




        Increasing the calculation period increases stability
Using 5-year or 10-year data, no switches of two or more categories
             Research recommendations
 The risk standards should focus on the risks inherent in asset classes
 Slight preference for downside deviation as the standard measure of risk, since it
  focuses on the part of the return distribution that preoccupies investors in
  practice. Of the other measures, standard deviation does best.
 The risk metric would be used to bucket funds based on their underlying asset
  classes
 This calculated bucket would represent the minimum acceptable rating for a
  product focusing on that asset class. Managers can use intuition to increase risk
  rankings but not reduce them.
 A long data span should be used to calculate risk metrics – ideally at least 10-
  years. This will help improve the stability of the risk bucketing and ensure the
  data accounts for at least one business cycle and facilitate independent
  verification.
 The bucket boundaries should be set using as long a data span as possible.
            Research recommendations
   Multi-asset class funds can be “slotted in” to the risk continuum by analysing
    the risk characteristics of their asset mix relative to the single asset classes.
   Capital guarantee funds are effectively a combination of a zero-coupon bond
    and a call option. They can be ranked using a combination of the guarantee
    and the asset the investor is exposed to. Intuitively its ranking will be above the
    underlying zero coupon bond, but below the underlying asset.
   Income-enhanced structured products generally involve the selling of insurance,
    which is complex to value and the value is not independent of the model used.
    They should be included in a separate “complex” risk category, where it is
    made clear that the investor is effectively selling an insurance product, rather
    than on the main scale.
   Ranking absolute return funds should probably be based on the asset mix. An
    adjustment to the risk calculations should be made to reflect the leverage
    allowance in the fund.
                Timetable

 Responses submitted to CESR 10 September
 ABI/IMA Research to be published
 CESR to report to Commission by 30 October
 Implementation July 2011?

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:414
posted:2/19/2010
language:English
pages:17