THE X FACTOR: GROUPING SECURITIES, DEFINING "SIMILAR" AND FORMING ESTIMATES - Nick Wade Northfield Information Services 2011

Page created by Lester Burns
 
CONTINUE READING
THE X FACTOR: GROUPING SECURITIES,
 DEFINING “SIMILAR” AND FORMING
            ESTIMATES

   Nick Wade
   Northfield Information Services
   2011

                                     1
WHAT’S THE MAIN POINT?
   Most of the models we use feature1:
         a fixed factor structure that does not allow for change or evolution in the way that markets
          convert information into price change
         and are estimated using one of many techniques that assume all the data in the sample
          follow some nice, clean, simple rules that bear no resemblance to real life
   We contend that risk models:
         Need some kind of adaptive structure that allows them to change as new or transient
          effects appear
         Need adjustment for “regimes” in the data – this impacts asset allocation and “insurance”
         Should harness current or forward-looking information as a conditioning input
         Should be estimated using a flexible technique that allows for evolution in the underlying
          data
   We also offer some thoughts about emergent techniques, new candidates for
    factors, and directions for future research

1 There are a lot more problems, but the focus for today is JUST the idea that things change over time

                                                                                                         2
ROADMAP
 Introduction to Factor Models
 Choices of Factors
 Pros and Cons
 A Hybrid Approach
 Better Factors
 Choices of Estimation Methodology
 Data Regimes
 Better Estimation Methodology
 Conclusions

                                      3
WHAT IS A FACTOR MODEL?
   The main purpose of a factor model is to find a set of
    common themes that explain the variability in security
    prices that is shared across securities.
   Having defined a set of factors, we then look to estimate
    the return associated with those factors and the
    individual security exposures/sensitivities/betas to
    those factors
   The end result is a model of how the portfolio will
    behave – how much the securities will move together
    and how much they will behave uniquely
   That can lead us to various useful risk characteristics

                                                            4
THE LINEAR MODEL

                   N              M               
          Rst = ∑ Eit Fit + S st  = ∑ G jt H jt 
                i =1              j =1            
      Relationship between R and F is linear ∀F
      There are N common factor sources of return
      Relationship between R and H is linear ∀H
      There is no correlation between F and H ∀ F,H
      The distribution of F is stationary, Normal, i.i.d. ∀F
      There are M stock-specific sources of return
      There is no correlation between H across stocks
      The distribution of H is stationary, Normal, i.i.d. ∀H
      (Implicitly also the volatility of R and F is stationary)

                                                                   5
ESTIMATING A RISK MODEL

                                 N
                          Rst = ∑ Eit Fit + S st
                                i =1
              N    N                         M
      V p = ∑∑ Ei E jσ iσ j ρ i , j + ∑ Wkσ 2 (S k )
              i =1 j =1                      k =1

   The Variance of a portfolio is given by the double sum over the
   factors contributing systematic or common factor risk, plus a
   weighted sum of the stock-specific or residual risks.

                                                                     6
COMMON FACTOR CHOICES

                        7
HOW MANY FACTORS…?

   The academic consensus seems to be that
    there is not much difference going from 5 to 10
    to 15 factors. In other words, 5 do the job.
     Lehmann   & Modest (1988)
     Connor & Korajczyk (1988)

     Roll & Ross (1980)

   If somebody is suggesting a 90 factor model to
    you, tell them to try harder

                                                     8
ENDOGENOUS MODEL
 The Endogenous or Fundamental Model seeks
  to estimate Fit assuming Eit by regression.
 Typical factors include E/P, D/E, Industry
  membership, Country membership…
     King(1966)
     Rosenberg and Guy (1975) etc.
 Model is pre-specified
 These models are hugely popular, partly
  because we’ve been building them for so long
  we’ve become habituated

                                                 9
EXOGENOUS MODEL

 The Exogenous or Macro model seeks to
  estimate Ei from Fit.
 Typical factors include Market, Sector, Oil,
  Interest-Rates…
     Ross(1976)
     Chen (1986)

   Model is pre-specified

                                                 10
STATISTICAL MODEL

 Assume N
 Use Factor Analysis or Principle Components to
  estimate Ei
 Use Regression to estimate Fit

 Errors in Variables

                                               11
PROS AND CONS OF EACH APPROACH
     Approach                                  Pros                                               Cons

Fundamental (micro)   • Suitable for concentrated portfolios             • Number of factors is fixed thus unchanging
model                                                                    • Dependent on accounting statement accuracy
                                                                         • Dependent on accounting standards
                                                                         comparability
                                                                         •Membership factors for industry/country/sector
                                                                         • Errors will be in factor returns, hence in
                                                                         covariance matrix, and hence not diversifiable

Macro-economic        • No dependence on accounting data                 • Factor number fixed and unchanging
model                                                                    • Exposure to factors is stationary over time
                      • The response of each security to changes in
                      market/sector/industry/ whatever to be different
                      across securities
                      • Errors will be in loadings (exposures), thus
                      diversifiable

Statistical model     • All correlation is information                   • Attribution of risk is difficult
                      • Captures new, or transient effects               • Issues with noise in data
                      • Adaptive to the market                           • Errors in variables
                      • Great for a short-term model                     • Number of factors is either pre-specified or
                                                                         sample-dependent

                                                                                                                           12
OTHER APPROACHES
   Combined Models:
       Northfield Hybrid Model
       Stroyny (2001)
   Simultaneous Estimation
       Black et al (1972)
       Heston and Rouwenhorst (1994, 1995)
       Satchell and Scowcroft (2001)
       GMM Hansen (1982)
       McElroy and Burmeister (1988) using NLSUR (which is assymptotically
        equivalent to ML)
   Bayesian Approach:
       Pohlson and Tew (2000)
       Ericsson and Karlsson (2002)

                                                                              13
SIMULTANEOUS ESTIMATION
   Removing the limitation of binary or membership variables (such as
    industry, country, sector, region etc).
        Marsh and Pfleiderer (1997)
        Scowcroft and Satchell (2001)
   Start with an estimate of the exposures (e.g. 1.00 for all companies)
    use that estimate to solve for the factor return, then use that factor
    return in turn to re-solve for a revised set of exposures, thus converging
    iteratively on a better solution for both Eit and Fit.
        Black et al (1972)
        Heston and Rouwenhorst (1994, 1995)
        Scowcroft and Satchell (2001)
   Given various limiting restrictions we can ensure that the model
    converges and that it is unique.
   This is the idea behind the UBS range of models and it’s good. But it’s
    not adaptive.

                                                                                 14
HYBRID MODEL (NORTHFIELD 1998)
   Combine macro, micro, and statistical factors
   An observable factor core, and statistical factors
    trawling the residual return to find new factors
   Gain the advantages of each, whilst mitigating the
    limitations of each
       Intuitive, explainable, justifiable observable factors
       Minimal dependence on accounting information
       Rapid inclusion of new or transient factors

                                                                 15
THOUGHTS
   Notice we are already trying to allow for an adaptive set of
    factors – we will pick this theme up later
   In a minute we’ll be allowing for time-dependent volatility
    and correlations as well
   But are all price movements based on fundamentals?
       Apple?
   What about news?
   What about liquidity impacts?
   What about index membership or common ownership?
       Exposed to “have to” sales as index weight changes for example
       Some securities are held in many funds, high “centrality”

                                                                         16
BETTER FACTOR CHOICES

                    17
A FEW SUGGESTIONS FOR BETTER FACTORS

   A factor can be any shared behavior
       HISTORICAL: Semantic clustering (text mining)
            Dig into everything published on a universe of companies and look
             for similarities by phrase comparison etc
       PREDICTIVE: News flows
            Look at instances of occurrence in news, sentiment
       Inference from other asset classes
          What does a bond spread change tell us about equity vol?
          What about a change in option implied volatility/implied correlation?
       Social network analysis
          Apply emergent techniques to look at influence within groups,
           measures of asset centrality, flow of information, diversification?
          Influence: types of network shape

                                                                                 18
NEWS AS A FACTOR

   Mitra, Mitra, diBartolomeo (2008)

   Since Dan’s going into depth later I will restrain
    myself – just note that you can use news
    incidence/sentiment to condition risk forecasts

                                                     19
OWNERSHIP DATA

   Not a well-explored source of information
       Starmine:[see Dirk Renick “Research on sentiment
         based smart money”]
            Use        ownership data to reverse out factor preferences
                     Funds are attracted to companies that are “like” ones they
                      already own
                     Funds exhibit biases toward companies with certain
                      fundamental characteristics, and these biases change over time
1999                2001            2003                2004                2005              2007
StarMine PriceMo    EPS_CAGR3       ROE                 ROE                 ROE               ROE
LTG                 ROE             Profit Margin       Interest Coverage   F12m E/P          F12m E/P
G5 EPS              Profit Margin   Interest Coverage   F12m E/P            Interest Coverage Interest Coverage
Debt/Assets         Debt/Assets     LTG                 Profit Margin       Profit Margin     Profit Margin
Interest Coverage   LTG             F12m E/P            LTG                 StarMine EQ       StarMine PriceMo

                                                                                                              20
COUNTRY, INDUSTRY, SECTOR, REGION…
   A useful (I hope) digression into the world of factor selection.
   It is pretty much standard practice to take note of membership
    in, or exposure to, one or more countries or regions, and one or
    more industries or sectors
   Problems: multinational firms, globalization, index domination
        Heston and Rouwenhorst (1994, 1995)
        Scowcroft and Sefton (2001)
        Diermeier and Solnik (2000)
        MacQueen and Satchell (2001)
   Suggestions:
        Estimate a different kind of index FTSE (1999), Bacon and Woodrow (1999)
        Split into “global” market and “domestic” market either by some cut off on a variable like foreign sales (Diermeier and
         Solnik 2000) or by some statistical process (MacQueen and Satchell 2001)
        Solve Model iteratively using Heston and Rouwenhorst (1994, 1995) approach
        Or extensions to that: Scowcroft and Sefton (2001).

   The real problem is that something as bland as “industry” is one-dimensional and
    does not pick up enough nuance about company relationships

                                                                                                                                   21
SEMANTIC CLUSTERING

   Let’s get rid of Industry Classifications
     Why?   Are all banks the same? Nope.
   Semantic clustering, or text mining can be used
    to:
     Help predict ratings changes [Starmine]
     Help update risk/return estimates [Ravenpack etc]

     Measure distance between companies based on
      published information [Quid.com]

                                                          22
SOCIAL NETWORK ANALYSIS
   A million followers, and still not winning… [thanks Charlie]
   Recent developments in SNA look at the different kinds of
    groups and influence between/within groups
       Solis (2009) stocks form a “small world” network
       “Six degrees of separation” etc.
       Measures of “centrality”
            Degree centrality
            Reach centrality
            Flow centrality
            Betweeness centrality
            Kritzman: Asset Centrality
            Google: page rank algorithm
   Intuitive groups: index membership, ownership in mutual
    funds

                                                                   23
IN CONTEXT
   For our purposes, how is a group connected?
     Think of localized version of CAPM – which asset
      best represents “the market”
     Hubs? What degree of connectedness?
     Ownership commonality across mutual funds

   Why do we care?
     Investment strategy, asset allocation
     Co-movement (risk), diversification, crowding
     Information flow

                                                         24
ASSET CENTRALITY

   We can take this idea from Social Network
    Analysis and apply it to a variety of contexts:
     Was  Lehman too big to fail? Can we quantify it’s
      centrality?
     Is BHP more influential than Rio, or less with the
      Resources sector?
     Which sectors are the most/least “democratic”?

     Note that this links again with James’ work on
      diversification

                                                           25
ABSORPTION RATIO (KRITZMAN 2011)
   On a related note – how tightly connected is the market, or a
    particular sector?
       Lo (2008)
       Yenilmez and Saltoglu (2011)
   Absorption ratio quantifies this by looking at the proportion
    of variance explained by common themes.
       As this number rises, the level of “systemic” risk rises, since
        assets are more tightly connected.
       This is one requirement for a crash – just add panic
       A signal for when to apply costly insurance – e.g. zero-cost collar
       You could use Dispersion and get a similar result (but without
        allowing for idiosyncratic vol to move indep of syst vol)
       You could use Implied Correlation and get a similar forward-
        looking result

                                                                              26
ARTIFICIAL IMMUNE SYSTEMS
   At a high level, our immune system consists of two
    pieces:
     Innate immunity
     Learned immunity
   In our context
     The factors we believe to be useful at t=0
     Plus the factors the model learns along the way
   Tune the model
     Criteria for accepting a new factor
     Criteria for archiving / forgetting factors
     Memory length for previously useful factors

                                                        27
SINGLE-PASS CLUSTERING AND RELATED
METHODS
 Concept: high frequency data in high volumes
  presents a storage problem, so need
  techniques that can analyze data as it arrives
  rather than data-mine.
 Issues: it’s a goldfish. No memory.

                                                   28
ISSUES WITH ESTIMATION

                     29
SOME PROBLEMS
   We are dealing with an evolving data set, not a static one
       Explore how this impacts our common techniques
       Look at more advanced / better techniques to fit evolving data
        sets
   We are (potentially) dealing with different regimes in the
    data, not one uniform set
       Look at models that explicitly allow for regime change (not in a
        George Bush sense)
   We are dealing with complex behavior within groups
       For example, some groups play follow the leader
       Some groups herd. There is no leader
       THOUGHTS: sefton: beta compression, social network analysis,
        influence

                                                                           30
THE WORLD IS VERY OBVIOUSLY TIME-VARYING

   Non-stationary volatility (ARCH, GARCH, etc)
       We spend an heroic amount of time trying to forecast non-
        stationary volatility
       But we often just ignore it when we calculate correlation, or
        perform regression analysis, or run factor analysis (or PCA)
   Non-stationary mean (Trend)
       We often build models to capture the alpha in momentum,
        reversals, and other manifestations of a non-stationary
        mean
       But we often ignore those when we calculate correlation, or
        perform regression analysis, or run factor analysis
   Read the fine print…

                                                                    31
SIMPLE ADJUSTMENTS (1)

   Non-stationary factor
    return series will lead
                                                                   2
    to the model                                    xi − x 
                                            V = ∑            
                                                               
    underestimating
                                                    n (n − 1) 
    portfolio risk
   Adjust by changing
    variance calculation to                                        2
                                                    xi 
    include trend                           V = ∑           
    component of return                             n(n − 1) 
        Adjust Model for the influence of non-stationary factor returns
                                                                          32
SIMPLE ADJUSTMENTS (2)
 What about security volatility?
 We observe:
     Serialcorrelation (not i.i.d.)
     Bid-ask bounce

     Non-Normal distributions

   Parkinson volatility

      Adjust Model for the influence of non-stationary security returns

                                                                          33
CONTEMPORANEOUS OR FORWARD-LOOKING
SIGNALS
   You could make the argument that these are “factors” rather than an
    estimation approach, but we use them to condition existing factors.
   Take a model that has been estimated on purely historical data
   Find true forward-looking signals
        E.g. option-implied volatility
   Find other contemporaneous signals
        E.g. dispersion measures, range measures, volume
   Adjust the parameters of the “historical” model so that the forecasts of the
    model match the signals from “now” and the “future”
   Update it daily so that it stays “current”

   The Advantage: we have kept the same factor structure but removed the
    sole dependency on the past.

                                                                                   34
GENERALIZE THE IDEA:
NORTHFIELD “ADAPTIVE NEAR-HORIZON RISK MODELS”
(ANISH SHAH, 2008)
The richest source of information about the future is not the past – increasingly it is consensus
estimates about the future from e.g. option markets, prediction markets…

   Take any risk model. e.g. one of our models estimated monthly

   Add “flexibility points” and fit to information about current conditions

   Adjust for statistical differences between short and long term returns

   Many benefits of this approach
         Avoid statistical complexities of high frequency data
         Keep familiar factor structure
         Common factor structure for long and short horizons permits interpolating any
          horizon in between
         Works with any factor model

                                                                                                    35
Northfield Asia ex. Japan Risk Models
     Tracking Error LH   Tracking Error SH   Linear (Tracking Error SH)
68

64

60

56

52

48

44

40

36

32

28

24

20

16

12

8

4

0

                                                                          36
DATA WITH REGIMES

                37
KEY QUESTIONS

   Before we get carried away…
     What  evidence of detectable regimes?
     How many regimes?

     What kind of model can fit multiple regimes?

     Can any of these fit multiple regimes on evolving
      data?
        i.e.   learn new regimes as they appear

                                                          38
EVIDENCE FOR REGIMES:
TIME-VARYING CORRELATIONS
   Increasing attention is being paid to the issue of correlations varying over time:
        (stocks) De Santis, G. and B. Gerard (1997), International asset pricing and portfolio
         diversification with time-varying risk, Journal of Finance, 52, 1881-1912.
        (stocks) Longin, F. and B. Solnik (2001), Extreme correlation of international equity markets,
         Journal of Finance, LVI(2), 646-676.
        (bonds) Hunter, D.M. and D.P. Simon (2005), A conditional assessment of the relationships
         between the major world bond markets, European Financial Management, 11(4), 463-482.
        (bonds) Solnik, B., C. Boucrelle and Y.L. Fur (1996), International market correlation and
         volatility, Financial Analysts Journal, 52(5), 17-34.
   Markov switching model: Chesnay, F and Jondeau, E “Does Correlation Between
    Stock Returns really increase during turbulent periods?” Bank of France research
    paper.

   To date little explored – however, Implied Correlation also seems useful, and more
    powerful than historical correlation in forecasting (we saw the same result with
    volatility):
        Campa, J.M. and P.H.K. Chang (1998), The forecasting ability of correlations implied in foreign
         exchange options, Journal of International Money and Finance, 17, 855-880.

                                                                                                           39
CORRELATION STABILITY
   One of the first… Kaplanis (1988): STABLE
   Tang (1995), Ratner (1992), Sheedy (1997): STABLE –
    although crash of 1987 regarded as an “anomaly”
   Bertero and Mayer (1989), King and Wadwhani (1990)
    and Lee and Kim (1993): correlation has increased, but
    STABLE
   Not quite so stable? Erb et al (1994) – increases in bear
    markets
   Longin and Solnick (1995) – increases in periods of
    high volatility
   Longin and Solnick (2001) – increases in bear markets

                                                            40
DETECTING REGIMES

 Use Viterbi’s algorithm (Viterbi 1967) to detect
  states.
 Use Jennrich tests (Jennrich 1970) to decide
  whether correlation differences between states
  are significant
 Mahalanobis Distance (Mahalanobis 1939,
  Kritzman 2009)

                                                 41
MAHALANOBIS DISTANCE
   MD is one example of a “Bregman divergence” , a group of distance measures.

   Clustering: classifies the test point as belonging to that class for which the Mahalanobis distance is
    minimal. This is equivalent to selecting the class with the maximum likelihood.

   Regression: Mahalanobis distance and leverage are often used to detect outliers, especially in the
    development of linear regression models. A point that has a greater Mahalanobis distance from the
    rest of the sample population of points is said to have higher leverage since it has a greater influence
    on the slope or coefficients of the regression equation. Specifically, Mahalanobis distance is also
    used to determine multivariate outliers. A point can be an multivariate outlier even if it is not a
    univariate outlier on any variable.

   Factor Analysis: recent research couples Mahalanobis distance with Factor Analysis and use MD to
    determine whether a new observation is an outlier or a member of the existing factor set. [Zhang
    2003]
   MD depends on covariance (S^-1 is the inverse of the covariance matrix), so is exposed to the same
    stationarity issues that affect correlation, however as described above it can help us reduce
    correlation’s outlier dependence.

                                                                                                             42
MAHALANOBIS IN ACTION

 Borrowed from Kritzman: Skulls, financial turbulence, and the
 implications for risk management. July 2009                     43
TURBULENCE IN THE MARKET

   Kritzman (2009):
     Correlation of US and foreign stocks when both
      markets’ returns are one standard deviation above
      their mean: -17%
     Correlation of US and foreign stocks when both
      markets’ returns are one standard deviation below
      their mean: +76%
       “Conditional correlations are essential for constructing
        properly diversified portfolios”

                                                                   44
CORRELATION REGIMES
   If it’s not stable, how about Markov switching
    models?
     Ramchand and Susmel (1998), Chesnay and Jondreau
      (2001) – correlation, conditioned on market regime,
      increases in periods with high volatility
     Ang and Bekaert (1999) – evidence for two regimes; a
      high vol/high corr, and a low vol/low corr.
   Don’t forget this needs to adapt as our world
    changes:
       Hidden Markov Experts on evolving data

                                                         45
FLEXIBLE ESTIMATION TECHNIQUES

                             46
TECHNIQUES FOR EVOLVING DATA
   Most of our favorite tools are designed to fit static data
    sets where behaviors are mostly unchanged
       Neural network, Kalman filter, OLS/GLS regression, PCA,
        ICA, factor analysis, variance, correlation… just about all of
        them
   Recent developments in cluster analysis are
    encouraging
       Artificial Immune Systems
       Single-pass clustering
       Regime-switching models e.g. HME etc
       [recent] EPCIA
       [very recent] HME on evolving data

                                                                         47
REGRESSION WITH NONSTATIONARY DATA

   Techniques have been developed specifically to
    allow time-varying sensitivities
     FLS (flexible least-squares)
     FLS is primarily a descriptive tool that allows us to
      gauge the potential for time-evolution of exposures

              T                      T −1
                                                      ′
             ∑ ( yt − xt β t )   + λ ∑ (β t +1 − β t ) (β t +1 − β t )
                             2

              t =1                    t =1

     Minimze both sum of squared errors and sum of squared dynamic errors
     (coefficient estimates)
                                                                            48
FLS EXAMPLE

   An example from Clayton and MacKinnon (2001)
   The coefficient apparently exhibits structural shift in 1992

                                                                   49
CLUSTER ANALYSIS WITH NONSTATIONARY DATA

   Guedalia, London, Werman; “An on-line agglomerative clustering method for
    nonstationary data” Neural Computation, February 15, 1999, Vol. 11, No. 2,
    Pages 521-54
   C. Aggarwal, J. Han, J. Wang, and P. S. Yu, On Demand Classification of Data
    Streams, Proc. 2004 Int. Conf. on Knowledge Discovery and Data Mining (KDD'04),
    Seattle, WA, Aug. 2004.
   G. Widmer and M. Kubat, “Learning in the Presence of Concept Drift and Hidden
    Contexts”, Machine Learning, Vol. 23, No. 1, pp. 69-101, 1996.

   Again, there are techniques available to
    conquer the problem

                                                                                      50
FACTOR ANALYSIS WITH NONSTATIONARY DATA
   Dahlhaus, R. (1997). Fitting Time Series Models to Nonstationary
    Processes. Annals of Statistics, Vol. 25, 1-37.
   Del Negro and Otrok (2008): Dynamic Factor Models with Time-
    Varying Parameters: Measuring Changes in International Business
    Cycles (Federal Reserve Bank New York)
   Eichler, M., Motta, G., and von Sachs, R. (2008). Fitting dynamic
    factor models to non-stationary time series. METEOR research
    memoranda RM/09/002, Maastricht University.
   Stock and Watson (2007): Forecasting in dynamic factor models
    subject to structural instability (Harvard).
   There are techniques available, and they are
    being applied to financial series.

                                                                        51
EVOLVING PRINCIPAL COMPONENT INNOVATION
ANALYSIS (EPCIA)
   You want PCA but your factor structure is
    changing
     EFA (evolving factor analysis): keep adding factors
     EFWFA (evolving fixed-window factor analysis): keep
      adding new factors, but forget old ones to make
      room!
     EPCIA allows for new factors to emerge

                                                        52
LAYER-EMBEDDED NETWORKS
   Within our networks, information and interactions
    may flow at multiple levels
     E.g. Pasquel and de Weck (2011 working paper)
     E.g. Luttrell (2010 working paper)
   Enter multi-layer modeling
     At a high level, a layer-embedded network captures the
      effects of multiple interacting processes at different
      frequencies or across different groups.
     E.g. combining short and long-term alpha signals
     Combining global and local factors
     Combining pervasive and short-term factors

                                                           53
SUMMARY THOUGHTS

               54
CONCLUSIONS

   Our world changes
       This requires an adaptive risk model factor structure
       This requires the ability to accommodate regimes in our risk models,
        our portfolio construction, hedging, and asset allocation by
        harnessing contemporaneous and forward-looking signals
   The market is not driven solely by fundamentals
       We need to leverage news/perception
       We need to explore nuanced relationships beyond bland
        membership
   Techniques exist to address all of these issues, and are
    being applied today.

                                                                               55
TAKE HOME
   Northfield:
       risk models that utilize implied volatility since 1997
       adaptive hybrid risk models since 1998
       risk models utilizing cross-sectional dispersion since 2003
       using implied volatility and dispersion in our entire range of
        short-horizon adaptive models since 2009
   If you’re doing some kind of time-series analysis on
    financial data you need to keep time-dependence,
    regimes, and evolving data in mind
   There are techniques to conquer all of these
    challenges, but they’re not the easy ones that come as
    part of Excel!
                                                                         56
REFERENCES
   Black F., Jensen M., Scholes M. “The Capital Asset Pricing Model: some empirical tests” In Jensen M.C., editor, “Studies
    in the Theory of Capital Markets” Praeger, New York, 1972.
   Bulsing M., Scowcroft A., and Sefton J., “Understanding Forecasting: A Unified framework for combining both analyst
    and strategy forecasts” UBS Working Paper, 2003.
   Chen N.F. Roll R. Ross S.A. “Economic Forces and the Stock Market” Journal of Business 59, 1986.
   Connor G and Korajczyck R.A. “Risk and Return in an equilibrium APT: application of a new test methodology” Journal of
    Financial Economics 21, 1988.
   diBartolomeo D. “Why Factor Risk Models Often Fail Active Quantitative Managers. The Completeness Conflict.”
    Northfield, 1998.
   Diermeier J. and Solnik B. “Global Pricing of Equity”, FAJ Vol. 57(4).
   Ericsson and Karlsson (2002)
   Fama E. and MacBeth J. “Risk, Return, and Equilibrium: empirical tests” Journal of Political Economy 71, 1973.
   GARP “Managing Tracking Errors in a Dynamic Environment” GARP Risk Review Jan/Feb 2004
   Hansen L. “Large Sample Properties of Generalized Method of Moments Estimators” Econometrica 50, 1982
   Heston S. and Rouwenhorst K. G. “Industry and Country Effects in International Stock Returns” Journal of Portfolio
    Management, Vol 21(3), 1995
   Hwang S. and Satchell S. “Tracking Error: ex ante versus ex post measures”. Journal of Asset Management, vol 2,
    number 3, 2001.
   King B.F. “Market and Industry Factors in Stock Price Behavior” Journal of Business, Vol. 39, January 1966.
   Lawton-Browne, C.L. Journal of Asset Management, 2001.

                                                                                                                               57
REFERENCES II
   Lintzenberger R. and Ramaswamy K. “The effects of dividends on common stock prices:
    theory and empirical evidence” Journal of Financial Economics 7, 1979.
   MacQueen J. “Alpha: the most abused term in Finance” Northfield Conference,
    Montebello, 2005
   MacQueen J. and Satchell S. “An Enquiry into Globalisation and Size in World Equity
    Markets”, Quantec, Thomson Financial, 2001.
   Mandelbrot B. “The variation of certain speculative prices” Journal of Business, 36. 1963.
   Markowitz, H.M. “Portfolio Selection” 1st edition, John Wiley, NY, 1959.
   Marsh T. and Pfleiderer P. “The Role of Country and Industry Effects in Explaining Global
    Stock Returns”, UC Berkley, Walter A. Haas School of Business, 1997.
   McElroy M.B., Burmeister E. “Arbitrage Pricing Theory as a restricted non-linear
    multivariate regression model” Journal of Business and Economic Statistics 6, 1988.
   Northfield Short Term Equity Risk Model
   Northfield Single-Market Risk Model (Hybrid Risk Model)
   Pfleiderer, Paul “Alternative Equity Risk Models: The Impact on Portfolio Decisions” The
    15th Annual Investment Seminar UBS/Quantal, Cambridge UK 2002.

                                                                                                 58
REFERENCES III
   Pohlson N.G. and Tew B.V. “Bayesian Portfolio Selection: An empirical analysis of the S&P 500 index
    1970-1996” Journal of Business and Economic Statistics 18, 2000.
   Pope Y and Yadav P.K. “Discovering Errors in Tracking Error”. Journal of Portfolio Management, Winter
    1994.
   Rosenberg B. and Guy J. “The Prediction of Systematic Risk” Berkeley Research Program in Finance,
    Working Paper 33, February 1975.
   Ross S.A. “The Arbitrage Theory of Capital Asset Pricing” Journal of Economic Theory, 13, 1976.
   Satchell and Scowcroft “A demystification of the Black-Litterman model: managing quantitative and
    traditional portfolio construction” Journal of Asset Management 1, 2000.
   Scowcroft A. and Sefton J. “Risk Attribution in a global country-sector model” in Knight and Satchell
    2005 (“Linear Factor Models in Finance”)
   Scowcroft A. and Sefton J. “Do tracking errors reliably estimate portfolio risk?”. Journal of Asset
    Management Vol 2, 2001.
   Shanken J. “The Arbitrage Pricing Theory: Is it testable?” Journal of Finance, 37, 1982.
   Sharpe W. “Capital Asset Prices: a theory of market equilibrium under conditions of risk” Journal of
    Finance, 19, 1964.
   Stroyny A.L. “Estimating a combined linear model” in Knight and Satchell 2005 (“Linear Factor Models
    in Finance”)
   Willcox J. “Better Risk Management” Journal of Portfolio Management, Summer 2000.

                                                                                                            59
REFERENCES IV
   Ang, A. and Bekaert, G. (1999) ‘International Asset Allocation with time-varying Correlations’, working
    paper, Graduate School of Business, Stanford University and NBER.
   Banerjee, Arindam; Merugu, Srujana; Dhillon, Inderjit S.; Ghosh, Joydeep (2005). "Clustering with
    Bregman divergences". Journal of Machine Learning Research 6: 1705–1749.
    http://jmlr.csail.mit.edu/papers/v6/banerjee05b.html.
   Bertero, E. and Mayer, C. (1989) ‘Structure and Performance:Global Interdependence of Stock
    Markets around the Crash of October 1987’, London, Centre for Economic Policy Research.
   Chesnay, F. and Jondeau, E. (2001) ‘Does Correlation between Stock Returns really increase during
    turbulent Periods?’, Economic Notes by Banca Monte dei Paschi di Siena SpA, Vol. 30,No. 1, pp.53–
    80.
   Jim Clayton and Greg MacKinnon (2001), "The Time-Varying Nature of the Link Between REIT, Real
    Estate and Financial Asset Returns" (pdf,6.3M), Journal of Real Estate Portfolio Management,
    January-March Issue
   Erb, C.B., Harvey, C.R. and Viskanta, T.E. (1994) ‘Forecasting international Equity Correlations’,
    Financial Analysts Journal,pp.32–45.
   Jakulin A & Bratko I (2003a). Analyzing Attribute Dependencies, in N Lavra\quad{c}, D Gamberger, L
    Todorovski & H Blockeel, eds, Proceedings of the 7th European Conference on Principles and
    Practice of Knowledge Discovery in Databases, Springer, Cavtat-Dubrovnik, Croatia, pp. 229-240
   Jennrich R. (1970) ‘An Asymptotic χ2 Test for the Equality of Two Correlation Matrices’, Journal of the
    American Statistical Association,Vol. 65, No. 330.

                                                                                                         60
REFERENCES V
   R. Kalaba, L. Tesfatsion. Time-varying linear regression via flexible least squares. International Journal
    on Computers and Mathematics with Applications, 1989, Vol. 17, pp. 1215-1245.
   Kaplanis, E. (1988) ‘Stability and Forecasting of the Comovement Measures of International Stock
    Market Returns’, Journal of International Money and Finance, Vol. 7, pp.63–75.
   Lee, S.B. and Kim, K.J. (1993) ‘Does the October 1987 Crash strengthen the co-Movements among
    national Stock Markets?’,Review of Financial Economics, Vol. 3, No. 1, pp.89–102.
   Longin, F. and Solnik, B. (1995) ‘Is the Correlation in International Equity Returns constant: 1960–
    1990?’, Journal of InternationalMoney and Finance, Vol. 14, No. 1, pp.3–26.
   Longin, F. and Solnik, B. (2001) ‘Extreme Correlation of International Equity Markets’, The Journal of
    Finance, Vol. 56, No.2.
   Mahalanobis, P C (1936). "On the generalised distance in statistics". Proceedings of the National
    Institute of Sciences of India 2 (1): 49–55. http://ir.isical.ac.in/dspace/handle/1/1268. Retrieved
    2008-11-05
   Nemenman I (2004). Information theory, multivariate dependence, and genetic network inference

                                                                                                            61
REFERENCES VI
   Osborne, Jason W. (2003). Effect sizes and the disattenuation of correlation and regression
    coefficients: lessons from educational psychology. Practical Assessment, Research & Evaluation,
    8(11).
   Qian, Edward and Ronald Hua. “Active Risk and the Information Ratio”, Journal of Investment
    Management, Third Quarter 2004.
   Ramchand, L. and Susmel, R. (1998) ‘Volatility and Cross Correlation across major Stock Markets’,
    Journal of Empirical Finance, Vol. 5, No. 4, pp.397–416.
   Ratner, M. (1992) ‘Portfolio Diversification and the inter-temporal Stability of International Indices’,
    Global Finance Journal, Vol. 3, pp.67–78.
   Sheedy, E. (1997) ‘Is Correlation constant after all? (A Study of multivariate Risk Estimation for
    International Equities)’, working paper.
   Sneath PHA & Sokal RR (1973) Numerical Taxonomy. Freeman, San Francisco.
   Tang, G.: ‘Intertemporal Stability in International Stock Market Relationships: A Revisit’, The Quarterly
    Review of Economics and Finance, Vol. 35 (Special), pp.579–593.Sharpe W. F., "Morningstar’s Risk-
    adjusted Ratings", Financial Analysts Journal, July/August 1998, p. 21-33.
   Viterbi, A. (1967) ‘Error Bounds for convolutional Codes and an asymptotically Optimum Decoding
    Algorithm’, IEEE Transactions on Information Theory, Vol. 13, No. 2, pp.260–269.Watanabe S (1960).
    Information theoretical analysis of multivariate correlation, IBM Journal of Research and
    Development 4, 66-82.
   Yule, 1926. G.U. Yule, Why do we sometimes get nonsense-correlations between time series?. Journal
    of the Royal Statistical Society 89 (1926), pp. 1–69.

                                                                                                           62
REFERENCES VII
   Solis, Rafael “Visualizing Stock Mutual Fund Relationships
    through Social Network Analysis”, Global Journal of Finance
    and Banking Issues Vol. 3 No. 3 2009
   Pascal, Michael C., de Weck, Olivier “Multilayer Network
    Model for Analysis and Management of Change Propagation”
    (working paper 2011)
   Luttrell, S.P. “Adaptive Cluster Expansion: A Multilayer
    Network for Estimating Probability Density Functions”
    (working paper 2010)
   Yenilmez, T, Saltoglu, B, “Analyzing Systemic Risk with
    Financial Networks during a Market Crash” (presentation
    March 10th 2011)

                                                             63
You can also read