Título: RAE-Revista de Administração de Empresas (Journal of Business Management), 2019. V. 59, N. 4

ARTICLES
Conditional pricing model with heteroscedasticity: Evaluation of Brazilian funds
Leandro Santos da Costa | Frances Fischberg Blank | Fernando Luiz Cyrino Oliveira | Cristian Enrique Muñoz Villalobos
Configurations of knowledge-intensive entrepreneurial ecosystems
Andre Cherubini Alves | Bruno Fischer | Nicholas Spyridon Vonortas | Sérgio Robles Reis de Queiroz
Business-to-business marketing: Brazilian scientific production from 2008 to 2018
Renato Calhau Coda | Gustavo Henrique Carvalho de Castro
Voluntary disclosure: Empirical analysis of the tone used in conference calls
Felipe Ramos Ferreira | Diego Campana Fiorot | Fabio Yoshio Suguri Motoki | Nadia Cardoso Moreira
PERSPECTIVES
Current challenges and promising avenues for entrepreneurship research
Rose Mary Almeida Lopes | Edmilson Lima
Evolution and trends of the international innovation research agenda
Bibiana Volkmer Martins | Kadígia Faccin | Gustavo da Silva Motta | Roberto Bernardes | Alsones Balestrin
BOOK REVIEW
Paths, texts, and contexts of creative leadership
Henrique Muzzio
BOOK RECOMMENDATION
Diversities, identities, and inclusion
Eloisio Moulin de Souza
RESEARCH AND
KNOWLEDGE
V. 59, N. 4,
July–August 2019
fgv.br/rae

ISSN 0034-7590; eISSN 2178-938X© RAE | São Paulo | 59(4) | July-August 2019
RAE-Revista de Administração de Empresas (Journal of Business Management)
CONTENTS
EDITORIAL
224 SOMETHING TO CELEBRATE: RAE'S NEW IMPACT FACTOR
Para comemorar: Novo fator de impacto da RAE
Para celebrar: Nuevo factor de impacto de la RAE
Maria José Tonelli
ARTICLES | ARTIGOS | ARTÍCULOS
225 CONDITIONAL PRICING MODEL WITH HETEROSCEDASTICITY: EVALUATION OF BRAZILIAN FUNDS
Modelo de precificação condicional com heteroscedasticidade: Avaliação de fundos brasileiros
Modelo de valoración condicional con heteroscedasticidad: Evaluación de fondos brasileños
242 CONFIGURATIONS OF KNOWLEDGE-INTENSIVE ENTREPRENEURIAL ECOSYSTEMS
Configurações de ecossistemas de empreendedorismo intensivo em conhecimento
Configuraciones de ecosistemas de emprendimiento intensivo en conocimiento
258 BUSINESS-TO-BUSINESS MARKETING: BRAZILIAN SCIENTIFIC PRODUCTION FROM 2008 TO 2018
Marketing business-to-business: Análise da produção científica brasileira de 2008 a 2018
Marketing business-to-business: Análisis de la producción científica brasileña desde 2008 hasta 2018
Renato Calhau Coda | Gustavo Henrique Carvalho de Castro
271 VOLUNTARY DISCLOSURE: EMPIRICAL ANALYSIS OF THE TONE USED IN CONFERENCE CALLS
Evidenciação voluntária: Análise empírica sobre o tom usado em audioconferências
Divulgación voluntaria: Análisis empírico del tono usado en conferencias telefónicas
Felipe Ramos Ferreira | Diego Campana Fiorot | Fabio Yoshio Suguri Motoki | Nadia Cardoso Moreira
PERSPECTIVAS | PERSPECTIVES
284 CURRENT CHALLENGES AND PROMISING AVENUES FOR ENTREPRENEURSHIP RESEARCH
Desafios atuais e caminhos promissores para a pesquisa em empreendedorismo
Desafíos actuales y caminos prometedores para la investigación empresarial
Rose Mary Almeida Lopes | Edmilson Lima
293 EVOLUTION AND TRENDS OF THE INTERNATIONAL INNOVATION RESEARCH AGENDA
Evolução e tendências da agenda de pesquisa internacional em inovação
Evolución y tendencias de la agenda internacional de investigación en innovación
Bibiana Volkmer Martins | Kadígia Faccin | Gustavo da Silva Motta | Roberto Bernardes | Alsones Balestrin
BOOK REVIEW | RESENHA | RESEÑA
308 PATHS, TEXTS, AND CONTEXTS OF CREATIVE LEADERSHIP
Trilhas, textos e contextos da liderança criativa
Pistas, textos y contextos de liderazgo creativo
Henrique Muzzio
BOOK RECOMMENDATION | INDICAÇÃO BIBIOGRÁFICA | RECOMMENDACIÓN BIBLIOGRÁFICA
310 DIVERSITIES, IDENTITIES, AND INCLUSION
Diversidades, identidades e inclusão
Diversidades, identidades e inclusión
Eloisio Moulin de Souza

ISSN 0034-7590; eISSN 2178-938X224 © RAE | São Paulo | 59(4) | July-August 2019 | 224
EDITORIAL
Maria José Tonelli
Editora-chefe
SOMETHINGTO CELEBRATE:RAE’S NEW IMPACT FACTOR
The impact factor (IF) of the RAE-Revista de Administração de Empresas (Journal of Business
Management) has increased in international indexes such as Journal Citation Reports (JCR), Scopus,
SciELO, and Red Iberoamericana de Innovación y Conocimiento Científico (REDIB). According to the
latest ranking published by Clarivate Analytics (June 2019), the increase in JCR was over 70% from
0.404 (2017) to 0.701 (2018). This result ranks RAE among the top 10 journals in Brazil indexed in the
WoS/SSCI database, which tracks social science journals. These numbers are extremely significant
when considering that they were obtained before 2018, when most of the articles were published in
Portuguese. RAE became bilingual in 2018, publishing all its articles in Portuguese and English, or
Spanish and English.
Also indexed in Elsevier’sScopus, the CiteScore of RAE increased from 0.49 (2017) to 0.80 (2018).
CiteScore is calculated as the total number of citations in 2018 divided by the number of documents
published between 2015 and 2017. The IF in SciELO increased from 0.4286 (2017) to 0.6056 (2018),
the highest among Brazilian business administration journals. The RAE’s Global Qualifier in REDIB is
26,637, ranking first in the general ranking of social sciences, which includes publications from Brazil,
Mexico, Chile, Spain, and other countries.
In addition to the invaluable collaboration of the Editorial Policy Committee, the Scientific
Editorial Board (Associate Editors), and all the reviewers who contribute to the development of the
articles, these advances would not be possible without the support of the authors who honor us
with their valuable submissions and by citing our articles. Our deepest thanks to the entire scientific
community, who make RAE’s visibility possible.
We would like to thank Professor Eduardo Diniz, who, during his work as Editor-in-Chief,
witnessed the inclusion of RAE in the WoS/SSCI database and raised RAE’s impact level to 0.408.
We also thank the Assistant Editor, Felipe Zambaldi, and our editorial team, Ilda Fontes, Eldi Soares,
Eduarda Pereira Anastácio, Denise Francisco Cândido, Andréa Cerqueira, Aline Santos, and Rute
Almeida, who are behind the scenes, but without whom this would not be possible.
A JCR of 0.701 allows us to participate in the international scene of academic publications in
business administration, even as we continue to publish articles in Portuguese, which is in line with
the mission of the Getulio Vargas Foundation of contributing to the development of Brazil. We thank
FGV EAESP for its support in this work.
Enjoy reading!
Maria José Tonelli1
ORCID: 0000-0002-6585-1493
1
Fundação Getulio Vargas, Escola de Administração de Empresas de São Paulo, São Paulo, SP, Brazil
Translated version
DOI: http://dx.doi.org/10.1590/S0034-759020190401

225 © RAE | São Paulo | 59(4) | July-August 2019 | 225-241 ISSN 0034-7590; eISSN 2178-938X
LEANDRO SANTOS DA COSTA1
leandrosdcosta@gmail.com
ORCID: 0000-0003-2183-8796
FRANCES FISCHBERG BLANK1
francesblank@puc-rio.br
ORCID: 0000-0002-2022-4226
FERNANDO LUIZ CYRINO
OLIVEIRA1
cyrino@puc-rio.br
ORCID: 0000-0003-1870-9440
CRISTIAN ENRIQUE MUÑOZ
VILLALOBOS2
crisstrink@gmail.com
ORCID: 0000-0002-7563-8469
1
Pontifícia Universidade Católica
do Rio de Janeiro, Departamento
de Engenharia Industrial, Rio de
Janeiro, RJ, Brasil
2
Pontifícia Universidade Católica
do Rio de Janeiro, Departamento
de Engenharia Elétrica, Rio de
Janeiro, RJ, Brasil
ARTICLES
Submitted 05.15.2018. Approved 05.07.2019
Evaluated through a double-blind review process. Scientific Editor: Fernanda Perobelli
Original version
CONDITIONAL PRICING MODEL WITH
HETEROSCEDASTICITY: EVALUATION OF
BRAZILIAN FUNDS
Modelo de precificação condicional com heteroscedasticidade: Avaliação de
fundos brasileiros
Modelo de valoración condicional con heteroscedasticidad: Evaluación de
fondos brasileños
ABSTRACT
Empirical studies have revealed that the conditional Capital Asset Pricing Model (CAPM) has a higher expla-
natory power than its unconditional version, particularly for the model in state-space form where the beta is
estimated using Kalman filter. Most empirical analyses are based on stock portfolios to explain financial ano-
malies, but only a few studies proposed improving investment fund performance. The main contribution of this
study is the assessment of Brazilian investment funds through traditional measures estimated from the CAPM
model in state-space form with heteroscedastic and homoscedastic errors compared to alternative models,
such as the unconditional CAPM and a four-factor model. Using a sample of stock funds from May 2005–April
2015, the results indicate that the conditional CAPM model produces better results than the alternative models,
providing better performance evaluation practices for funds in both stock-picking and market-timing ability.
KEYWORDS | Conditional Capital Asset Pricing Model, Kalman filter, time-varying betas, investment funds, per-
formance analysis.
RESUMO
Os resultados empíricos na literatura demonstram que a versão condicional do Modelo de Precificação de
Ativos Financeiros (CAPM), particularmente no que se refere ao modelo na forma em espaço de estado, no qual
o beta é estimado pelo filtro de Kalman, possui maior poder explicativo do que a sua versão incondicional. A
maioria das análises empíricas na literatura baseia-se em portfólios de ações para explicar anomalias financei-
ras, porém poucos estudos propõem-se a melhorar a avaliação de desempenho de fundos de investimento. A
principal contribuição deste artigo consiste em avaliar fundos de investimento brasileiros por meio de medidas
tradicionais estimadas a partir do CAPM na forma em espaço de estado com erros heteroscedásticos e homos-
cedásticos e comparar seus resultados com modelos alternativos, tais como CAPM incondicional, modelo de
quatro fatores. Utilizando uma amostra de fundos de ações, os resultados indicam que o modelo CAPM con-
dicional produz melhores resultados do que os modelos alternativos, proporcionando melhores práticas de
avaliação de desempenho em relação às habilidades de stock-picking e market-timing.
PALAVRAS-CHAVE | Modelo de precificação de ativos financeiros condicional, filtro de Kalman, betas variantes
no tempo, fundos de investimento, análise de performance.
RESUMEN
Los resultados empíricos en la literatura revelan que la versión condicional del CAPM, particularmente con
respecto al modelo en forma de espacio de estado, en el cual se estima beta mediante el filtro de Kalman,
posee mayor poder explicativo que su versión incondicional. La mayoría de los análisis empíricos se basan
en carteras de valores para explicar anomalías financieras, pero pocos estudios proponen mejorar el rendi-
miento de los fondos de inversión. La principal contribución de este estudio a la literatura es que lleva a cabo
la evaluación de fondos de inversión a través de medidas condicionales generadas a partir del CAPM en forma
espacio-estado con errores heteroscedásticos y homoscedásticos y que compara sus resultados con modelos
alternativos, tales como CAPM incondicional, modelo de cuatro factores. Utilizando una muestra de fondos de
acciones, los resultados indican que el modelo CAPM condicional produce mejores resultados que los modelos
alternativos, proporcionando mejores prácticas de evaluación de desempeño en relación con las habilidades
de stock-picking y market-timing.
PALABRAS CLAVE | Modelo de valoración de activos de capital condicional, filtro de Kalman, betas variantes en
el tiempo, fondos de inversión, análisis de rendimiento.

ARTICLES | CONDITIONAL PRICING MODEL WITH HETEROSCEDASTICITY: EVALUATION OF BRAZILIAN FUNDS
INTRODUCTION
Empirical studies have shown the failures of the capital asset
pricing model (CAPM) through consistently different returns from
those predicted by the model. Since the development of the
Fama and French (1993) and Carhart (1997) factor models, several
studies have used these models in the performance analysis of
investment funds to evaluate to what extent the returns could
be attributed to two main managerial skills: stock-picking and
market-timing. While the former concerns the manager’s ability
to select the best assets for a given level of risk, the latter refers
to managing the funds’ beta in anticipation of future market
movements.
In general, the evidence shows that after considering
all fund expenses, managers do not have stock-picking ability
(Carhart, 1997; Elton, Gruber, & Blake, 2012; Fama & French,
2009). The results are more controversial for market-timing, with
some authors pointing to evidence of positive managerial timing
(Bollen & Busse, 2000; Busse, 1999; Ferson & Schadt, 1996) while
many others pointing to the lack of this ability (Elton et al., 2012;
Treynor & Mazuy, 1966).
In Brazil, the conclusions regarding managers' abilities
vary greatly depending on the sample and the proposed model.
In general, there are more positive results reported than in the
international literature (Borges & Martelanc, 2015). Using the
one-factor model and variations of CAPM, Eid and Rochman
(2006) found evidence of superior performance of stock funds
in relation to the market; Leusin and Brito (2008) observed
positive and significant alphas, as well as weak evidence for
few managers with market timing ability; and Matos and Nave
(2012) verified the persistence among better performing funds.
However, Casaccia, Galli, Macêdo, and Leitao (2011) did not
identify any special abilities of managers in their sample. There
are other studies on Brazilian funds using factor models where,
in most cases, no superior managerial skills are evident. Castro
and Minardi (2009) used the three-factor model (E. F. Fama &
French, 1993) and four-factor model (Carhart, 1997), along with
a fifth factor on market timing. The authors did not observe
stock-picking ability of managers on comparing active and
passive funds. Jordão and Moura (2011) analyzed an extensive
sample from Carhart (1997) model for stock-picking ability and
from Treynor and Mazuy (1966) for market-timing and found that
less than 5% of the funds presented positive and significant
results for such skills. Nerasti and Lucinda (2016) investigated
the persistence in the performance of Brazilian stock funds with
four models: traditional CAPM, the three-factor model proposed
by Fama and French (1993), Carhart's four-factor model (1997),
and an additional model, adding the risk factor associated with
asset liquidity. They did not find persistence in the superior
performance of Brazilian managers. Borges and Martelanc (2015)
estimated the alphas in a sample comparing real funds and
synthetic funds using the four-factor model of Carhart (1997) and
found managers' positive ability to generate abnormal returns,
albeit modest.
Since managers assume different levels of risk depending
on the kind of fund, empirical evidence shows that funds generally
do not maintain constant levels of risk over time (Lee & Rahman,
1990; Mamaysky, Spiegel, & Zhang, 2008), which is different from
the assumption in the unconditional CAPM and traditional factor
models. Therefore, a more accurate modelling of the temporal
variation in the fund’s risk should result in a more realistic
assessment of its performance.
The conditional version of CAPM was developed to address
the limitations of the traditional static version. There are three
main approaches to model the dynamic behavior of the beta: (i)
modelling the conditional distribution function of returns as an
explicit function of lagged conditioning variables (Jagannathan
& Wang, 1996; Lettau & Ludvigson, 2001), (ii) describing the
beta dynamics using conditional or stochastic volatility models
(Bodurtha Jr & Mark, 1991; Bollerslev, Engle, & Wooldridge, 1988;
Yu, 2002), (iii) using state-space models where the beta dynamics
are directly modelled as a stochastic process (Adrian & Franzoni,
2009; Blank, Samanez, Baidya, & Aiube, 2014; Jostova & Philipov,
2005; Mergner & Bulla, 2008).
Several studies have indicated that the conditional
state-space CAPM specification provides more accurate
estimates of beta than the others (Adrian & Franzoni, 2009;
Faff, Hillier, & Hillier, 2000; Mergner & Bulla, 2008). Moreover,
such differences are more pronounced in daily data than in
monthly data (Bollen & Busse, 2000). One problem with financial
returns is the temporal dependence on its conditional variance—
the presence of heteroscedasticity. However, most studies
assumed that the conditional CAPM residuals are homoscedastic,
despite empirical tests finding high heteroscedasticity even
after the conditional treatment of the model (Mergner & Bulla,
2008). Ortas, Salvador, and Moneva (2015) constructed the
heteroscedastic version of the conditional CAPM model using
the Kalman filter algorithm, where the errors are modelled as
a generalized autoregressive conditional heteroscedasticity
(GARCH) process. The results show that the heteroscedastic
model surpasses the homoscedastic model in in-sample and
out-of-sample analysis.
For Brazil, Tambosi Filho, Garcia, Imoniana, and Moreiras
(2010) test a conditional CAPM incorporating macroeconomic

and financial variables and verify a significant increase in the
explanatory power of the model. Using a state-space form
conditional CAPM, Mazzeu, Da Costa Júnior, and Santos (2013)
observed a reduction in pricing errors using the time-varying
beta model in a sample of 13 stocks in the Brazilian market.
Blank et al. (2014) build portfolios of stocks based on book-to-
market and market value characteristics and verify that when the
beta is modeled as a random walk with conditioning variables,
pricing errors are reduced. Caldeira, Moura, and Santos (2013
use a similar approach in combining a dynamic conditional
covariance matrix based on a GARCH model and the risk factors
proposed by Carhart (1997) with time-varying coefficients. The
authors observe satisfactory results compared to benchmark
models.
Some international authors propose incorporating
temporal dynamics in systematic risk. Ferson and Schadt (1996),
based on a conditional model where the funds’ betas depend on
lagged variables, analyze a sample of US funds and find that both
static CAPM and a conditional factor model produce a negative
alpha more often than positive ones. The results show that the
conditional model eliminates evidence of negative timing of
managers found by the unconditional model.
Holmes and Faff (2008) compare conditional CAPM based
on lagged variables and state-space conditional CAPM using
an Australian funds sample. While the first model shows the
presence of stock-picking ability, it is not observed in the state-
space model. Similarly, Mamaysky et al. (2008) find the alpha and
beta estimates of a large sample of US funds. Their predictions
estimated through the Kalman filter are more accurate than those
of ordinary models.
The literature in Brazil does not use conditional models
to analyze managers’ abilities and compare those models with
traditional approaches, such as the unconditional CAPM and
multifactor models. This paper aims to fill this gap, with two
main objectives: (i) to evaluate conditional CAPM models in
the space-state form applied to a sample of Brazilian stock
funds; this is estimated from the Kalman filter with the errors
of the regression equation in homoscedastic (SS-HOM) and
heteroscedastic (SS-HET) forms, and (ii) to analyze how
using traditional measures obtained from the conditional
CAPM model can improve the current practice of evaluating
the performance of investment funds and managers´ abilities
of stock-picking and market timing, compared to alternative
models, such as the unconditional CAPM and the Carhart’s
(1997) four-factor model.
Using selected Brazilian funds, the results show that the
modelling of the heteroscedastic structure of the errors increases
the CAPM conditional capacity to capture the alpha and beta
temporal dynamics of investment funds. Given the superiority of
the proposed conditional models, this study then evaluates the
performance of stocks funds from 02/05/2005 to 30/04/2015.
Evidence suggests that managers’ ability to select the best assets
is directly related to significant appreciation in the stock market
and that managers do not have the ability to anticipate periods
of appreciation and fall in the market.
The rest of this paper is organized as follows. The following
section presents the econometric models of asset pricing. Then,
the performance of the CAPM models in the state-space form with
residuals of the homoscedastic and heteroscedastic observation
equation is analysed. Next, the empirical application of the
models is carried out, and, finally, the conclusions of the research
are presented.
MODEL AND ESTIMATION PROCEDURE
The appeal of CAPM, developed independently by Sharpe (1964),
Lintner (1965), and Mossin (1966), lies in its simplicity, where the
expected return for a given asset is given as:
E Ri
⎡⎣ ⎤⎦ = βi
E Rm
⎡⎣ ⎤⎦( ) (1)
βi
=
cov Ri
,Rm
⎡⎣ ⎤⎦
var Rm
⎡⎣ ⎤⎦
(2)
where Ri and Rm are the excess returns on asset i and the
market portfolio in relation to the risk-free asset, respectively, and
βi is a risk measure that is not eliminated through diversification,
also known as systematic risk or beta.
However, one critical limitation of CAPM is its static nature.
In such a hypothesis, the presence of anomalies could be due
to beta time-varying dynamics that are not captured by CAPM in
its original form. Preserving the structure of one model factor,
different models are used to capture the time-varying systematic
risk, with the conditional models in state-space usually giving
the best results.
The model described by equations (1) and (2) is built on
expected values and is, therefore, non-observable. It is commonly
tested on time series using observable measures of realized
returns. Thus, considering a classic Gaussian univariate linear

regression model, the conditional version with time-varying alpha and beta can be written as the state-space model given by
equations (3) to (5).
Ri ,t
=αi ,t
+ βi ,t
Rm,t
+ ∈i ,t
, ∈i ,t
~N 0,σ∈i
2
( ),t =1,...n (3)
αi ,t +1
=αi ,t
+ϑi ,t
, ϑi ,t
~N 0,σϑi
2
( ) (4)
βi ,t +1
= βi ,t
+ηi ,t
, ηi ,t
~N 0,σηi
2
( ) (5)
Equation (3) is known as observation one and (4) and (5)
are the state equations. It is usually assumed that errors ϵi,t
, ϑi,t
,
and ηi,t
are serially independent and homoscedastic and αi,t
is
known as Jensen’s alpha.
The intercept is statistically assumed to be zero in
CAPM, which means that the market risk premium adjusted
by the assets’ beta would be sufficient to explain the
observed returns. However, if a portfolio manager can better
forecast asset prices, higher returns than the ones implied
in the model could be obtained, and Jensen’s alpha could
represent an average incremental return rate of the portfolio
by the unit of time exclusively due to the manager’s ability.
In an unconditional form, equation (3) would be estimated
by ordinary least squares (OLS) where alpha and beta are
constant over time.
A random walk process describes the alpha and beta
time-varying dynamics. Pizzinga and Fernandes (2006) outlined
three main reasons for justifying such a choice: (i) parsimony, (ii)
simplicity, and (iii) the possibility of fundamental managerial
changes over time due to the non-stationary property.
The specification of the model in equations (3) to (5) allows
the direct application of Kalman filter to estimate time-varying
and constant parameters (Adrian & Franzoni, 2009; Faff et al.,
2000; Mergner & Bulla, 2008). The constant parameters are
estimated, in particular, through prediction error decomposition
and maximization of the log-likelihood function given by:
logL ψi( )= − n
2
log2π − 1
2 t =1
n
log|Ft
ψi( )|+
vt
ψi( )
Ft
ψi( )
⎛
⎝
⎜
⎜
⎞
⎠
⎟
⎟
,∑ (6)
where ψi
= σ∈,i
2
,σϑ ,i
2
,ση,i
2
( ) is the hyper-parameter vector of the model; vt
(ψi
) = Ri
– E[Ri
|Ωt-1,
ψi
] and Ft
(ψi
) = Var (vt
(ψi
))
are calculated through Kalman filter, known as one-step-ahead prediction error and its variance, respectively, and Ωt-1
={R1
,...,Rt-1
}. In
this study, state vectors βt
and αt
follow random walk processes and the Kalman filtering used is in its diffuse exact form.
Nonetheless, the volatility clusters, especially in the daily asset returns series, are responsible for creating a structure of time
dependent conditional variance in this series, which are not captured by the model described in equations (3) to (5). In this context,
Ortas et al. (2015) propose a heteroscedastic version of the conditional CAPM model estimated by Kalman filter algorithm, where
the errors of the regression equation are modelled with conditional variance according to a GARCH (1,1) process:
∈i ,t
|Ωt −1
~N 0,σ∈i ,t
2
( ), Ωt −1
= {R1
,...,Rt −1
} (7)

σ∈i ,t
2
=ωi
+ ρ ∈i ,t −1
2
+γσ∈i ,t −1
2
, ωi
>0,ρi
,γ i
≥0eρi
+γ i
<1 (8)
The estimated model specified in equations (3)–(5) and (7)–(8) follows the iterative procedure proposed by Ortas et al. (2015).
The process maximizes a partial log-likelihood function assuming constant mean parameters and then maximizes a partial log-
likelihood function assuming constant variance parameters. The parameters are estimated based on the Kalman Filter algorithm.
The procedure alternates between those two steps until it achieves convergence.
Two well-established models in asset-pricing literature used in performance analysis of investment funds are also estimated for
comparison purposes. The first is the unconditional CAPM, tested on the time series using observable measures of realized returns,
estimated from equation (9). The estimation is performed using OLS with rolling windows of 120 days, which is a commonly used
alternative in the ad hoc attempt to adjust the time-varying coefficients of the model.
Rit
=αi
+ βi
Rmt( )+εit
(9)
where εit
is the i.i.d. error term, with E(εit
)=0 and Var εit( )=σεi
2
.
The second approach is Carhart's (1997) four-factor model given in equation (10).
Ri ,t
=αi
+ βiM
RM ,t
+ βiS
SMBt
+ βiH
HMLt
+ βiW
WMLt
+εit
(10)
where SMBt
captures the anomaly in the size of the
company; HMLt
captures the anomaly in the book-to-market ratio;
and WMLt
captures the anomaly in the moment of stocks. Again,
OLS is used with rolling windows of 120 days for the estimation.
MODEL PERFORMANCE
This section comparatively analyses the performance of
conditional CAPM models in state-space form for homoscedastic
and heteroscedastic observation equation errors and compares
these results with well-established models.
Data
The data for this study were obtained from the Quantum Axis
online database. We selected the funds classified in the category
‘Stocks’, such as ‘Free Stocks’, which had a complete sample
from February 5, 2005 to April 30, 2015, totaling 2,474 daily
observations for each series. This time span was used because
it coincides with a period of accelerated development of the
Brazilian funds industry.
Itisworthmentioningthattheselectedsamplemightpresent
survival bias since the sample comprises funds that have survived
during the period of time analyzed. However, studies have shown
(Brown, Goetzmann, Ibbotson & Ross, 1992; Carvalho, 2005)
that the inclusion of bias in the end result of performance analysis
measuresisminimal,withapositivebiasonmeanreturnsfrom0.2%
to 0.8% per year. Still, according to Milan and Eid (2014), in the
Brazilian market, the main reason for a majorityofinvestmentfunds
to terminate istheir merger with other funds, notpoor performance.
Both the fund quotas used as the market indexes were
adjusted for the distribution of dividends and are net of the taxes
charged by funds. We used the arithmetic return as a measure
of daily return of the funds, calculated based on excess return in
relation to the risk-free rate.
Table 1 presents the main descriptive statistics for the
daily return series. The funds are organized in descending order
of the total accumulated returns in the analyzed period. A greater
number of selected funds show a mean positive return for the
period. Funds with the worst positions show average negative
returns, which indicates that they did not show the capacity to
overcome the risk-free assets, once the results shown are in terms
of return excess. The Ljung–Box test applied to the squared returns
confirms the heteroscedasticity of the daily returns financial series
once the statistical value of the test for all funds rejects the null
hypothesis that the autocorrelation of the series equals zero at
the significance level of 1%.

Table 1. Descriptive statistics of excess return series of funds
Fund
Average
Return
Standard
Deviation
LB²(6) Net Worth Total Return
Rm,t
0.018% 1.60% 1448.3¹ - 12.4%
1 0.042% 1.31% 1459.5¹ R$ 253,505.195 130.7%
2 0.040% 1.21% 932.0¹ R$ 223,226.073 123.2%
3 0.033% 1.17% 1735.4¹ R$ 96,507.324 89.5%
4 0.037% 1.53% 1789.2¹ R$ 48,425.162 86.0%
5 0.029% 1.17% 1827.4¹ R$ 141,638.916 74.9%
6 0.026% 0.89% 1709.2¹ R$ 12,462.034 74.2%
7 0.028% 1.19% 1072.0¹ R$ 213,115.019 68.7%
8 0.028% 1.27% 1651.3¹ R$ 47,754.743 62.5%
9 0.022% 0.95% 1458.1¹ R$ 16,210.743 54.7%
10 0.024% 1.16% 1503.6¹ R$ 167,376.501 54.8%
11 0.029% 1.57% 1570.5¹ R$ 113,865.041 51.2%
12 0.021% 1.25% 1509.7¹ R$ 43,797.267 38.6%
13 0.022% 1.41% 1748.6¹ R$ 15,557.122 35.6%
14 0.013% 0.86% 38.9¹ R$ 112,662.732 27.0%
15 0.013% 1.27% 1416.7¹ R$ 56,117.911 13.1%
16 0.014% 1.42% 1535.5¹ R$ 5,893.538 11.2%
17 0.013% 1.43% 1895.6¹ R$ 208,973.148 6.7%
18 0.015% 1.58% 2052.1¹ R$ 148,715.381 5.2%
19 0.014% 1.61% 1855.1¹ R$ 19,162.430 2.9%
20 0.009% 1.36% 1339.5¹ R$ 7,459.161 0.3%
21 0.011% 1.82% 1193.8¹ R$ 122,802.719 -13.3%
22 -0.006% 1.07% 515.9¹ R$ 20,835.687 -25.7%
23 0.003% 1.80% 1408.1¹ R$ 6,339.762 -28.8%
24 -0.006% 1.33% 793.2¹ R$ 29,408.209 -31.3%
25 -0.002% 1.68% 1687.2¹ R$ 24,322.706 -32.4%
26 -0.003% 1.80% 986.2¹ R$ 1,075.375 -37.8%
27 -0.006% 1.58% 1487.0¹ R$ 27,290.762 -37.6%
28 -0.005% 1.83% 1061.9¹ R$ 525,664.075 -41.3%
29 -0.012% 1.56% 1227.9¹ R$ 16,259.577 -45.1%
30 -0.013% 1.58% 1597.8¹ R$ 1,585.373 -46.7%
31 -0.013% 1.57% 1325.0¹ R$ 749.245 -47.0%
32 -0.036% 1.64% 1418.4¹ R$ 1,307.690 -70.9%
33 -0.042% 1.42% 907.9¹ R$ 2,354.023 -72.7%
34 -0.012% 2.89% 373.7¹ R$ 171,496.267 -73.9%
35 -0.040% 2.36% 102.4¹ R$ 2,532.896 -81.2%
36 -0.094% 2.24% 427.0¹ R$ 139.201 -94.7%
Notes: 1, 2, 3
Significant at the 1%, 5%, and 10% level, respectively.
This table shows the basic descriptive statistics of the daily fund returns from May 2, 2005 to April 30, 2015. JB is the Jarque–Bera normality test. ADF is the augmented
Dickey–Fuller test for unit root. LB² is the Ljung–Box test applied to the squared returns. The number of ‘lags’ is determined according to Tsay (2010): Ln 2474( )= 7.81 8. The
return in the period is calculated as the accumulated return for the entire sampling period.

Model Estimates
Here, we analyze the constant parameter estimates of the
conditional CAPM in state-space form when the observation
equation residuals, ϵi,t
, are homoscedastic (SS-HOM), in equations
(3) to (5), and heteroscedastic (SS-HET), in equations (3) to (5)
and (7) and (8).
The results for the SS-HOM and SS-HET constant parameter
estimation for each of the sample funds are shown in Table 2.
Like in other studies, we observe that the funds’ betas seem to
follow a time-varying dynamic process once the parameters ση,i
2
are statistically different from zero, at the 1% significance level,
for 34 out of the 36 sample funds. Regarding investment funds,
this result is expected since the forecasting strategies used by
managers and/or the variation in stock beta that are part of their
portfolios generate variations in funds’ betas. It is worth noting
that the constant parameter estimates are lesser for the SS-HET
model than for the SS-HOM model.
Table 2. SS-HOM and SS-HET model estimates
Fund ϵi,t
σ∈,i ,t
2
ση,t
2
σv ,t
2
LogL LB²(6) JB
1
HOM 4.0E-05¹ 6.9E-04¹ 1.9E-10¹ 8941.26 281.53¹ 3323.95¹
HET 4.1E-05 2.1E-04¹ 1.9E-11* 9163.68 5.29 385.77¹
2
HOM 7.2E-05¹ 2.2E-03¹ 6.6E-09¹ 8177.76 105.21¹ 1656.05¹
HET 6.8E-05 1.0E-03¹ 2.2E-10¹ 8388.27 15.09³ 412.54¹
3
HOM 3.0E-05*
4.2E-04*
2.8E-22*
9269.42 227.23¹ 143.37¹
HET 3.2E-05 4.3E-04*
6.4E-23*
9364.25 6.33 31.79¹
4
HOM 3.2E-05¹ 4.3E-04¹ 1.2E-10¹ 9201.67 226.55¹ 5276.06¹
HET 3.4E-05 2.2E-04¹ 9.6E-11¹ 9370.53 29.16¹ 222.78¹
5
HOM 2.4E-05¹ 5.4E-04¹ 1.6E-10¹ 9557.61 346.30¹ 9142.01¹
HET 2.3E-05 1.8E-04¹ 5.7E-11¹ 10226.79 27.89¹ 678.44¹
6
HOM 1.5E-05¹ 5.1E-04¹ 1.5E-09¹ 10093.85 391.11¹ 4828.62¹
HET 1.6E-05 2.5E-04¹ 1.4E-10¹ 10325.09 22.07¹ 413.71¹
7
HOM 6.2E-05¹ 5.2E-04¹ 2.2E-09* 8390.70 42.20¹ 1955.76¹
HET 5.5E-05 4.6E-04¹ 1.6E-10¹ 8558.32 6.37 916.44¹
8
HOM 5.0E-05¹ 9.8E-04¹ 2.7E-10¹ 8633.61 220.74¹ 1600.70¹
HET 5.4E-05 3.2E-04¹ 2.4E-1¹ 8828.70 7.23 548.61¹
9 HOM 3.0E-05¹ 3.2E-04¹ 6.5E-09¹ 9295.25 389.24¹ 2672.03¹
HET 3.2E-05 5.7E-05¹ 1.3E-09¹ 9313.95 19.41² 252.10¹
10 HOM 3.0E-05¹ 5.9E-04¹ 5.7E-09¹ 9269.83 155.45¹ 15029.31¹
HET 3.4E-05 3.5E-04¹ 1.5E-10¹ 9403.12 14.70³ 1566.99¹
11 HOM 6.1E-05¹ 5.3E-04¹ 1.8E-10¹ 8425.89 243.99¹ 1565.72¹
HET 6.5E-05 2.4E-04¹ 1.2E-12* 8364.22 18.19² 66.04¹
12 HOM 2.6E-05¹ 2.1E-04¹ 4.7E-11¹ 9491.40 364.40¹ 14732.96¹
HET 2.5E-05 1.1E-04¹ 2.9E-11¹ 10055.05 23.00¹ 792.44¹
13 HOM 2.8E-05¹ 8.6E-04¹ 3.1E-11¹ 9360.39 468.20¹ 660.44¹
HET 3.1E-05 4.1E-04¹ 1.3E-11¹ 9289.92 21.69¹ 64.22¹
14 HOM 6.7E-05* 2.1E-04¹ 9.5E-09¹ 8313.37 30.02¹ 5783.80¹
HET 7.1E-05 1.6E-04¹ 6.3E-09¹ 8362.87 11.29 10156.94¹
15 HOM 4.5E-05¹ 5.8E-04¹ 1.5E-18* 8795.09 118.35¹ 2725.92¹
HET 4.9E-05 3.0E-04¹ 7.0E-19* 8832.44 6.45 263.13¹
16 HOM 4.7E-05¹ 1.0E-03¹ 1.1E-09¹ 8711.18 71.46¹ 7999.75¹
HET 5.0E-05 9.8E-04¹ 5.1E-10¹ 8804.96 22.47¹ 993.42¹
17 HOM 3.8E-05¹ 6.3E-04¹ 1.0E-18* 8991.46 239.25¹ 6668.44¹
(continue)

Fund ϵi,t
σ∈,i ,t
2
ση,t
2
σv ,t
2
LogL LB²(6) JB
HET 3.9E-05 3.1E-04*
1.7E-20*
9106.42 13.97³ 395.98¹
18 HOM 5.1E-05¹ 4.8E-04¹ 5.1E-09¹ 8643.48 351.34¹ 1132.51¹
HET 5.4E-05 2.4E-04¹ 1.3E-09¹ 8865.84 22.65¹ 220.01¹
19 HOM 3.2E-05¹ 2.5E-04¹ 4.7E-18 9216.70 780.17¹ 5222.83¹
HET 3.1E-05 2.4E-04¹ 1.9E-18¹ 9731.35 32.07¹ 203.54¹
20 HOM 6.6E-05¹ 6.3E-04¹ 8.0E-09¹ 8313.00 239.70¹ 1040.05¹
HET 7.3E-05 6.3E-04¹ 3.4E-09¹ 8472.97 3.22 1038.71¹
21 HOM 6.6E-05¹ 5.5E-04¹ 1.0E-18 8335.80 891.32¹ 1105.52¹
HET 6.8E-05 1.2E-04¹ 1.3E-18* 8836.12 120.66¹ 184.37¹
22 HOM 6.4E-05¹ 8.8E-04¹ 9.1E-08¹ 8316.47 174.52¹ 1706.23¹
HET 7.4E-05 2.2E-04¹ 4.5E-08¹ 8171.99 10.3 502.16¹
23 HOM 3.4E-05¹ 3.3E-04¹ 1.4E-10¹ 9140.05 700.00¹ 4847.10¹
HET 3.5E-05 3.4E-04¹ 7.1E-11¹ 9470.84 54.56¹ 136.66¹
24 HOM 8.5E-05¹ 8.0E-03¹ 9.5E-09¹ 7904.87 408.83¹ 7857.83¹
HET 1.1E-04 3.7E-03¹ 1.9E-09¹ 7832.50 17.66² 1393.25¹
25 HOM 3.4E-05¹ 4.5E-04¹ 7.1E-12* 9143.03 806.43¹ 3548.63¹
HET 3.6E-05 7.3E-04¹ 5.3E-11¹ 9729.34 21.05¹ 119.14¹
26 HOM 7.2E-05¹ 4.0E-03¹ 7.4E-12¹ 8146.67 318.85¹ 427.95¹
HET 8.3E-05 2.4E-03¹ 1.8E-18¹ 8021.39 16.58² 94.87¹
27 HOM 3.5E-05¹ 2.6E-04¹ 1.9E-11 9139.62 358.28¹ 6655.91¹
HET 3.5E-05 1.2E-04¹ 3.8E-19 9356.64 8.39 262.63¹
28 HOM 7.9E-05¹ 4.9E-03¹ 4.8E-18* 8029.70 226.65¹ 186.25¹
HET 9.2E-05 3.1E-03¹ 1.2E-19* 7836.96 13.39³ 75.12¹
29 HOM 3.8E-05¹ 6.9E-04¹ 1.8E-10¹ 9005.35 192.07¹ 15357.91¹
HET 3.8E-05 4.8E-04¹ 2.8E-10¹ 9278.35 33.24¹ 514.22¹
30 HOM 3.7E-05¹ 4.4E-04¹ 4.1E-11 9045.65 702.41¹ 4383.64¹
HET 3.8E-05 3.7E-04¹ 1.1E-18* 9387.07 26.67¹ 117.59¹
31 HOM 6.6E-05¹ 2.2E-04¹ 2.2E-09¹ 8340.41 149.95¹ 1229.98¹
HET 6.9E-05 3.3E-04¹ 2.1E-09¹ 8171.33 7.96 310.94¹
32 HOM 5.8E-05¹ 7.2E-04¹ 1.2E-18² 8477.58 209.68¹ 7414.02¹
HET 5.9E-05 2.0E-04+
6.5E-23*
8750.22 1.76 648.45¹
33 HOM 6.0E-05¹ 5.3E-05¹ 1.3E-10¹ 8465.04 146.18¹ 19756.21¹
HET 6.1E-05 9.9E-05¹ 1.2E-10¹ 8573.00 3.86 35162.41¹
34 HOM 8.0E-04 6.6E-05¹ 1.3E-16¹ 5293.45 372.73¹ 14620.96¹
HET 8.4E-04 2.2E-04¹ 2.7E-09¹ 5452.36 25.90¹ 5426.30¹
35 HOM 4.2E-04¹ 2.0E-02¹ 2.0E-09¹ 5992.38 113.59¹ 1534487.60¹
HET 5.1E-04 2.7E-03¹ 3.9E-09¹ 6546.36 0.16 2104729.98¹
36 HOM 4.5E-04* 6.4E-05¹ 6.5E-09¹ 5992.06 401.48¹ 13132.76¹
HET 4.8E-04 8.9E-06¹ 1.5E-09¹ 6243.57 9.43 8304.16¹
Notes: ¹, ², ³ Significant at the 1%, 5%, and 10% level, respectively; * Significance not measured by the program.
The second column represents the assumption for residue ϵt
of the observation equation (HOM: homoscedastic or HET: heteroscedastic). The next three columns
represent the hyper-parameter variance maximum likelihood estimates, for observation equations and the state variables . The error variance of the observation
equation refers to the unconditional variance of ϵt
. LB² refers to the Ljung–Box test applied to the standardized squared residuals of the models. The number of lags is
determined as in Tsay (2010): . JB refers to the Jarque–Bera normality tests.
Table 2. SS-HOM and SS-HET model estimates (continuation)

With regard to the time-varying dynamics of alpha for the 26
sample funds, it is possible to reject the null hypothesis (σϑ ,i
2
to
be statistically zero). However, it is observed that their estimates
are close to zero. This suggests that the alpha parameter related
to the managers’ stock-picking ability gradually varies over time.
Goodness of fit Measure of Models
The SS-HOM and SS-HET models discussed in the previous
section are comparatively assessed in this subsection. We use
the following measures to test the goodness of fit: Akaike’s
information criterion (AIC), Bayesian information criterion (BIC),
root mean square error (RMSE), and mean absolute error (MAE).
For the residual diagnostic tests, we utilize the Jarque–Bera (JB)
and Ljung–Box (LB) tests.
Table 3 shows the AIC and BIC for the SS-HOM and SS-HET
models for all the sample funds. The majority of funds, 26 out
of the 36, show lower AIC and BIC values for the heteroscedastic
model than for the homoscedastic model, thereby indicating
that the approach of the heteroscedastic structure of errors
ϵt
increases the capacity of the conditional CAPM model to
capture the time-varying dynamics of alpha and beta in the
sample funds.
Table 3. Goodness-of-fit measures
Fund
AIC BIC
HOM HET HOM HET
1 -7.2241 -7.4048 -7.2124 -7.3954
2 -6.6069 -6.7779 -6.5952 -6.7685
3 -7.5046 -7.5822 -7.4928 -7.5728
4 -7.4347 -7.5720 -7.4229 -7.5626
5 -7.7224 -8.2642 -7.7107 -8.2548
6 -8.1559 -8.3436 -8.1442 -8.3342
7 -6.7928 -6.9294 -6.7810 -6.9200
8 -6.9896 -7.1484 -6.9778 -7.1390
9 -7.5103 -7.5262 -7.4986 -7.5168
10 -7.4898 -7.5983 -7.4780 -7.5889
11 -6.8075 -6.7585 -6.7958 -6.7491
12 -7.6689 -8.1253 -7.6571 -8.1159
13 -7.5630 -7.5068 -7.5512 -7.4974
14 -6.7301 -6.7711 -6.7184 -6.7616
15 -7.1060 -7.1370 -7.0942 -7.1276
16 -7.0381 -7.1148 -7.0264 -7.1054
17 -7.2647 -7.3585 -7.2530 -7.3491
18 -6.9834 -7.1640 -6.9717 -7.1546
19 -7.4468 -7.8637 -7.4351 -7.8543
20 -6.7299 -6.8602 -6.7181 -6.8508
21 -6.7347 -7.1400 -6.7229 -7.1306
22 -6.7191 -6.6031 -6.7073 -6.5937
23 -7.3848 -7.6531 -7.3731 -7.6437
(continue)

Fund
AIC BIC
HOM HET HOM HET
24 -6.3863 -6.3286 -6.3746 -6.3192
25 -7.3873 -7.8620 -7.3755 -7.8526
26 -6.5818 -6.4813 -6.5700 -6.4719
27 -7.3845 -7.5607 -7.3727 -7.5513
28 -6.4872 -6.3322 -6.4755 -6.3228
29 -7.2759 -7.4975 -7.2642 -7.4881
30 -7.3085 -7.5853 -7.2968 -7.5759
31 -6.7384 -6.6025 -6.7267 -6.5931
32 -6.8493 -7.0705 -6.8375 -7.0611
25 -7.3873 -7.8620 -7.3755 -7.8526
26 -6.5818 -6.4813 -6.5700 -6.4719
27 -7.3845 -7.5607 -7.3727 -7.5513
28 -6.4872 -6.3322 -6.4755 -6.3228
29 -7.2759 -7.4975 -7.2642 -7.4881
30 -7.3085 -7.5853 -7.2968 -7.5759
31 -6.7384 -6.6025 -6.7267 -6.5931
32 -6.8493 -7.0705 -6.8375 -7.0611
33 -6.8530 -6.9413 -6.8412 -6.9319
34 -4.2752 -4.4045 -4.2635 -4.3951
35 -4.8402 -5.2889 -4.8285 -5.2795
36 -4.8400 -5.0441 -4.8282 -5.0347
Notes: AIC is Akaike information criterion and BIC is Bayesian information criterion. The higher measures are in bold.
Table 3. Goodness-of-fit measures
Other measures used for the explanatory power analysis
are RMSE and MAE. The in-sample forecast returns were estimated
for each time t from January 2, 2006 to April 30, 2015 totaling
2,305 observations.
Apart from the results shown for the SS-HOM and SS-HET
models, we analyze the performance of the unconditional CAPM
and the four-factor model of Carhart (1997) using OLS with rolling
windows of 120 days.
Considering the explanatory, not predictive, objective of
this analysis, for conditional CAPM models in space-state form
estimated by Kalman filter algorithm, the smoothed versions of
the state variables are used. Table 4 shows the sample average
RMSE and MAE for each model.
The resultsinTable 4 show thatthe rolling windowsapproach
could be better when the alpha and beta are time-varying and the
RMSE and MAE tests for theSS-HOM andSS-HET models are inferior
to the CAPM and the factors model. Hence, in the comparison
between conditional models and the factors model, the results
are favorable for conditional models, showing that considering
in-sample explanatory power and time-varying alpha and beta
could bring superior benefits to the introduction of more riskfactors.
Conversely, the comparison between SS-HOM and SS-HET models
suggests a preference for the homoscedastic model, showing a
lower RMSE and MAE on average. From the viewpoint of the model’s
explanatory power, the heteroscedastic model does not seem to
obtain better results than the homoscedastic model, in general.
(continuation)

Table 4. Consolidated result of the in-sample goodness
of fit of the 36 sample funds
Model RMSE MAE
SS-HOM 8.00E-03 5.36E-03
SS-HET 8.15E-03 5.43E-03
CAPM 8.56E-03 5.77E-03
Factors model 8.37E-03 5.68E-03
Notes: Each cell shows the sample average RMSE and MAE for each model.
Residual Diagnostic Tests
The results of the residual diagnostic tests are shown in Table 2,
in columns 7 and 8, for JB and LB tests, respectively.
The normality hypothesis of standardized residual
for both SS-HOM and SS-HET models is rejected for all the
analyzed funds in the JB test. We observe a reduction in the
latter compared to the former even though the t-statistic value
for the SS-HET model is distant from the values of a normal
standard distribution.
The SS-HOM model is not able to capture the
heteroscedasticity in the residuals, once the null hypothesis of
non-autocorrelation of the squared standardized residuals is
rejected for all funds in the LB test. Conversely, this hypothesis is
not rejected for the squared standardized residuals of the SS-HET
model for 14 out of the 36 sample funds. The finding that 22 out
of the 36 analyzed sample funds still feature such dependence
structure in standardized residuals of SS-HET model requires
further investigation.
Therefore, we analyze the correlograms of funds that
persistently show heteroscedasticity signs in the residuals of the
SS-HET model. Figure 1 shows two examples of this analysis. As
with all the other funds not shown here, we observe that despite
the values of residual autocorrelation being statistically different
from zero, they are not relevant. In comparison, Figure 2 shows
the correlograms of the standardized residuals of the SS-HOM
model for the same funds. This comparison shows that the SS-HET
model better captures the time-dependence structure of return
series variance, given the substantial decrease in autocorrelation
among the residuals.
Figure 1. Standardized squared residuals correlograms for SS-HET model
Note: This figure shows the correlograms of the standardized residuals of the SS-HET model for funds 05 (left) and 25 (right).

Figure 2. Standardized squared residuals correlograms for SS-HOM model
Note: This figure shows the correlograms of the standardized residuals of the SS-HOM model for funds 05 (left) and 25 (right).
Once the conditional models show a higher capacity
to adjust to sample data, it is reasonable to suppose that the
evaluation of performance measures obtained from this model
might improve the actual performance evaluation.
CONDITIONAL CAPM IN THE ANALYSIS
OF BRAZILIAN EQUITY FUNDS
This section uses the measures obtained from the conditional
models to carry out the performance analysis of managers’ stock-
picking and market-timing abilities.
Fund Performance Measures: Conditional Alphas
The metric used here to compare fund performance is intercept
αi,t
of equations (3), (9), and (10), also known as Jensen’s alpha.
Since the objective of evaluation is to estimate the
measures of performance evaluation from the impacts of different
models, we built an equally weighted portfolio with all sample
funds, in such a way that the portfolio returns in each period
is given by mean returns of all funds in the same period. The
portfolio’s alpha estimates for each of the models are shown
in Graph 1.
Unlike most of the previous studies, which usually deduce
the existence of stock-picking ability by estimating the alpha in a
given time span, our model allows us to obtain the alpha estimate
at each instant of time, allowing the analysis of managers’ stock-
picking ability over time. In this sense, observing Graph 1, we
can conclude that managers' ability to select the best assets
is directly related to periods of market upswings (2007–2008
and 2009–2010). During periods of market downturn (2008–
2009), managers, in general, deliver negative excess of return
to their investors. Moreover, since the peak in 2010, managers’
stock-picking ability has reduced gradually, having shown overall
negative values in the last years of the sample.
A substantial difference can be noticed among the alpha
estimates of each model, especially between the conditional
models (SS-HOM and SS-HET) and the unconditional ones
(unconditional CAPM and factors model), with the latter
consistently higher than the former in absolute terms. One
possible explanation for this phenomenon is that the lower

explanatory power of the unconditional models, confirmed by the RMSE and MAE measures, tends to overestimate the values of
alpha intercept estimates. Hence, the time-varying beta of funds would not be captured satisfactorily with the rolling windows
estimates, being confused with positive or negative abnormal returns in the CAPM and factors models. This means that a portion of
the alpha values estimated by the commonly used unconditional models do not deal with the superior ability of managers but only
inadequately capture the temporal variation in the fund's beta.
Graph 1. Alpha point estimates for an equally weighted portfolio for all sample funds
Note: This figure shows the portfolio’s alpha estimates for the SS-HOM, SS-HET, CAPM, and factors models.
Thus, these results indicate four issues regarding managers’
stock-picking ability: (1) the managers’ ability to select the best
assets might be directly related to periods of stock market
upswings; (2) during periods of market downturns, the managers’
search for assets with potential appreciation greater than their
risk level incurred abnormal or negative returns; (3) managers
have consistently showed negative alphas in the last years of the
sample; and (4) in general, a portion of the alpha values estimated
by the unconditional models are not due to the managers’ stock
picking ability, but merely the model’s inability to adequately
capture the temporal variation in the beta.
Fund Performance Measures: Conditional Betas
Since managers’ market-timing ability is directly related to fund
beta variation analysis over time, one must understand the
evolution of estimated beta for the different models. Graph
2 shows the series of betas for some of the sample funds for
SS-HOM and SS-HET models. A less noisy series of the SS-HET
estimates than of the SS-HOM is observed in the first years of
the sampling, especially between 2006 and 2008, a time of
considerable market volatility caused by the global financial
crisis of 2008. Further investigation shows that the daily
standard deviation of market factor returns of 2006–2008
(2.04%) is higher than for 2009–2015 (1.29%). According to
Ortas et al. (2015), a less noisy estimate from the heteroscedastic
model outperformance occurs because the leptokurtosis of the
unconditional distribution of ϵi,t
reduces the influence of outliers
during the beta estimation process. In other words, periods of
greater volatility would be marked by greater differences in the
beta estimates—as is observed.

Graph 2. Beta point estimates for sample funds
Regarding conditional and unconditional models, it can
be noted that periods of rising beta for the former, in general,
are periods of falling beta for the latter, and vice versa. Since
the estimate for conditional beta in a given period is carried out
conditionally on the information for the following periods, the
smoothed estimates usually anticipate the future changes in
the estimated variable.
To analyze managers’ market-timing ability, we used the
approach of Holmes and Faff (2008), where a daily series of beta
estimated by conditional and unconditional models is used as a
dependent variable in market factor regression, as presented in
the following equation:
βi ,t
k
= constant +γ i
k
Rm,t
+εi ,t
(11)
where subscript i is the analysed sample fund and
superscript k the analyzed model; γ i
k
is the regression coefficient,
estimated by OLS; and Rm,t
represents the excess returns of the
market factor. The manager shows market forecasting ability
when γ i
k
>0.If γ i
k
<0 the manager increases beta during market
downswing and lowers beta in market upswing.
The results of the estimates of γ i
k
and their p-values
are detailed in Table 5. The use of SS-HOM and SS-HET
conditional models reveals more funds for which managers’
market-timing ability is negative than for the CAPM and
factors models. Moreover, we verify a decrease in the mean
coefficients of the sample funds γ i
k
compared to the CAPM and
factors models, which begin to show substantially negative
values for conditional models. These results show that the
conditional models alter the market-timing ability analysis
compared to the CAPM and factors models, thereby indicating
that the managers of the analyzed sample cannot forecast
periods of market downturns or upswings, even when doing the
opposite by systematically increasing beta funds during market
downswings and decreasing beta funds during upswings.
However, with regard to the comparison between the results
of SS-HOM and SS-HET models, no significant differences are
observed.

Table 5. Comparison of market-timing coefficients γ i
k
for the conditional CAPM and factors model
SS-HOM SS-HET CAPM Factors model
Panel A: summary statistics
No. of positive cases γ y
k
>0( ) 3 2 21 13
No. of negative cases γ y
k
<0( ) 33 (8) 34 (6) 15 23
Panel B: correlation coefficients between estimated beta series
SS-HOM 0.9715 0.7560 0.6728
SS-HET 0.7812 0.6896
CAPM 0.8931
Notes: The numbers in brackets represent results that are statistically different from zero at the significance level of 10%. Panel B shows the means correlation coefficients
of the beta series for the 36 sample funds.
A negative coefficient related to market-timing ability
warrants explanations. According to Ferson and Schadt (1996),
a negative coefficient could arise when a manager has perverse
ability to forecast market movement in the opposite direction.
The negative correlation between the funds’ beta and the market
factor returns could also be caused by the flow of investment
funds: since significant investments in funds tend to decrease
its beta, they tend to increase during periods of market upswings.
The explanation for managers’ negative market-timing ability
would be the significant flow of investments during market
upswings.
Further, Panel B of Table 5 shows that, in general, there is
a strong correlation between the beta estimates of conditional
models. That could be one explanation for the market-timing
results not displaying substantial differences. This correlation
decreases when the estimates of conditional models are analyzed
against other models, the conditional and factors models. This
could explain the market-timing results displaying more significant
differences among the different models.
CONCLUSIONS
The results obtained in this study show that the measures of the
heteroscedastic Kalman Filter model provide better performance
evaluation for funds regarding managers’ ability in Brazilian
stock funds, for stock-picking and market-timing, than the
traditional models.
Unlike many previous papers, especially in the Brazilian
context, a heteroscedastic version of conditional CAPM is
compared with the results of the homoscedastic version of the
model, and alternative models, such as the four-factor model of
Carhart (1997). Keeping the one-factor model structure, the results
show that the modelling of the heteroscedastic structure of errors
increases the capacity of the conditional CAPM model to capture
the funds’ alpha and beta timing dynamics.
The state-space models were also compared for goodness
of fit to the unconditional CAPM and four-factor model of Carhart
(1997), both estimated with rolling-windows. The results indicate
the superiority of the conditional models for all pricing measures
used. These results suggest that the alpha and beta timing
variation bring superior benefits than the introduction of more
risk factors.
After the higher quality of goodness of fit to the conditional
models is determined, the conditional measures of investment
fund performance evaluation were estimated.
We can draw four conclusions regarding managers’ stock-
picking ability: (1) the managers’ ability to better select assets
can be directly related to periods of market upswings; (2)
during periods of market downturns, the managers’ search for
better assets with an appreciation potential higher than their
risk level, leads to abnormal negative returns; (3) managers
have consistently shown negative alphas in the last years of
the sample; and (4) in general, a portion of the alpha values
estimated by the unconditional models do not show the presence
of managers’ stockpicking ability but only an inability of the model

to adequately capture the temporal variation in the beta. As for
the market-timing ability, the results indicate that managers of
the analyzed sample not only do not have forecasting ability
for market downturns or upswings, rather end up doing just the
opposite— systematically increasing (decreasing) fund betas in
market downturns (upswings).
Further research in this area can be highlighted. Hybrid
models combining the Kalman Filter approach and lagged
macroeconomic variables as conditioning variables can bring
more information on the managers’ performance based on their
strategies. Heteroscedastic multifactor models in the context
of conditional pricing models with time-varying coefficients
can bring even better results both in terms of goodness of fit
and forecasting. That would be a step ahead from this paper.
Particularly considering the methodology displayed here and
given the potential different applications of Kalman filter based-
methods, equivalent procedures could be reproduced and applied
to other areas of study in Finance and Economics. Finally, more
sophisticated models based on machine learning techniques can
also be an avenue of research in this area.
REFERENCES
Adrian, T., & Franzoni, F. (2009). Learning about beta: Time-varying
factor loadings, expected returns, and the conditional CAPM.
Journal of Empirical Finance, 16(4), 537-556. doi:10.1016/j.
jempfin.2009.02.003
Agudo, L. F., Magallón, M. V., & Sarto, J. L. (2006). Evaluation of
performance and conditional information: The case of Spanish
mutual funds. Applied Financial Economics, 16(11), 803-817.
doi:10.1080/09603100500397245
Blank, F. F., Samanez, C. P., Baidya, T. K. N., & Aiube, F. A. L. (2014).
CAPM condicional: Betas variantes no tempo no mercado brasileiro.
Revista Brasileira de Finanças, 12(2), 163-199.
BodurthaJ.N.,&Mark,N.(1991).TestingCAPMwithtime-varyingrisksand
returns. Journal of Finance, 46(4), 1485-1505. doi:10.2307/2328868
Bollen, N., & Busse, J. (2001). On the timing ability of mutual fund
managers. TheJournalofFinance,56(3), 1075-1094. doi:10.1111/0022-
1082.00356
Bollerslev, T., Engle, R.F., & Wooldridge, J. M. .(1988.). A capital asset
pricing model with time-varying covariances. Journal of Political
Economy, 96(1), 116-131.
Borges, E. C., & Martelanc. (2015). Sorte ou habilidade: Uma avaliação
dos fundos de investimento no Brasil. Revista de Administração,
50(2), 196-207. doi:10.5700/rausp1194
Brown, S. J., Goetzmann, W., Ibbotson, R. G. & Ross, S. A. (1992).
Survivorship bias in performance studies. The Review of Financial
Studies, 5(4), 553-580.
Busse, J. A. (1999). Volatility timing in mutual funds: Evidence from daily
returns. Review of Financial Studies, 12(5), 1009–1041. doi:10.1093/
rfs/12.5.1009
Caldeira, J. F., Moura, G. V. & Santos, A. P. (2013). Seleção de carteiras
utilizando o modelo Fama-French-Carhart. Revista Brasileira de
Economia, 67(1), 45-65. doi:10.1590/S0034-71402013000100003
Carhart, M. M. (1997). On persistence in mutual fund performance. The
Journal of Finance, 52(1), 57-82. doi:10.2307/2329556
Carvalho, M. R. A. (2005). Avaliação de desempenho de fundos
multimercado: Resultados passados podem ser utilizados para
definir uma estratégia de investimento? Revista de Economia e
Administração, 4(3), 367-387.
Casaccia, M. C., Galli, O. C., Macêdo, G. R., & Leitao, C. (2011). Análise
do desempenho de fundos de investimentos: Um estudo em ações
brasileiras no período de janeiro de 2004 a agosto de 2009. Revista
Organizações em Contexto, 7(13), 1-30. doi:10.15603/1982-8756/ roc.
v7n13p1-30
Castro, B., & Minardi, A. (2009). Comparação do desempenho dos
fundos de ações ativos e passivos. Revista Brasileira de Finanças,
7(2), 143-161.
Cuthbertson, K., Nitzsche, D., & O’Sullivan, N. (2008). UK mutual fund
performance: Skill or luck? Journal of Empirical Finance, 15(4), 613–
634. doi:10.1016/j.jempfin.2007.09.005
Eid, W., & Rochman, R. (2006). Fundos de investimento ativos e passivos
no Brasil: Comparando e determinando seus desempenhos. Article
presented in the thirtieth Encontro da ANPAD, Salvador, BA.
Elton, E. J., Gruber, M. J., & Blake, C. R. (2012). An examination of mutual
fund timing ability using monthly holdings data. Review of Finance,
16(3), 619-645. doi:10.1093/rof/rfr007
Faff, R. W., Hillier, D., & Hillier, J. (2000). Time varying beta risk: An
analysis of alternative modelling techniques. Journal of Business
Finance & Accounting, 27(5–6), 523-554. doi:10.1111/1468-
5957.00324
Fama, E. F., & French, K. R. (1993). Common risk factors in the returns
on stocks and bonds. Journal of Financial Economics, 33(1), 3-56.
doi:10.1016/0304-405X(93)90023-5
Fama, E., & French, K. (2009). Luck versus skill in the cross-section of
mutual fund returns. Journal of Empirical Finance, 65(June), 1915-
1947.
Ferson, W. E., & Schadt, R. W. (1996). Measuring fund strategy and
performance in changing economic conditions. Journal of Finance,
51(2), 425-461. doi:10.2307/2329367
Holmes, K. A., & Faff, R. (2008). Estimating the performance attributes
of Australian multi-sector managed funds within a dynamic Kalman
filter framework. International Review of Financial Analysis, 17(5),
998-1011. doi:10.1016/j.irfa.2008.05.001
Jagannathan, R., & Wang, Z. (1996). The conditional CAPM and the
cross-section of expected returns. The Journal of Finance, 51(1), 3-53.
doi:10.2307/2329301
Jordão, G. A., & de Moura, M. L. (2011). Performance analysis of Brazilian
hedge funds. Journal of Multinational Financial Management, 21(3),
165-176. doi:10.1016/j.mulfin.2011.02.002
Jostova, G., & Philipov, A. (2005). Bayesian analysis of stochastic betas.
The Journal of Financial and Quantitative Analysis, 40(4), 747-778.

Laes, M. A., & da Silva, M. E. (2014). Performance of mutual equity
funds in Brazil–A bootstrap analysis. Economia, 15(3), 294-306.
doi:10.1016/j.econ.2014.08.002
Lee, C. F.., & Rahman, S. (1990). Market timing, selectivity, and mutual
fund performance: An empirical investigation. The Journal of
Business, 63(2), 261-278.
Lettau, M., & Ludvigson, S. (2001). Resurrecting the (C)CAPM: A cross-
sectional test when risk premia are time-varying. Journal of Political
Economy, 109(6), 1238-1287. doi:10.1086/323282
Leusin, L. de M.., & Brito, R. D. (2008). Market timing e avaliação de
desempenhodosfundosbrasileiros.RAE-RevistadeAdministraçãode
Empresas, 48(2), 22-36. doi:10.1590/S0034-75902008000200003
Lintner, J. (1965). The valuation of risk assets and the selection of risky
investments in stock portfolios and capital budgets. The Review of
Economics and Statistics, 47(1), 13-37. doi:10.2307/1924119
Mamaysky, H., Spiegel, M., & Zhang, H. (2008). Estimating the dynamics
of mutual fund alphas and betas. Review of Financial Studies, 21(1),
233-264.
Matos, P., & Nave, A. (2012). Fundos de investimento em ações no
Brasil: Performance e expertise de gestão. Brazilian Business Review,
9(Especial), 1-38.
Mazzeu, J. H. G., Da Costa Júnior, N. C. A. Da. & Santos, A. A. P. (2013).
CAPM condicional com aprendizagem aplicado ao mercado brasileiro
de ações. RAM-Revista de Administração Mackenzie, 14(1), 143-175.
doi:10.1590/S1678-69712013000100007
Mergner, S., & Bulla, J. (2008). Time-varying beta risk of Pan-European
industry portfolios : A comparison of alternative modeling
techniques. The European Journal of Finance, 14(8), 37-41.
doi:10.1080/13518470802173396
Milan, P. L. A. B., & Eid, W. (2014). Elevada rotatividade de carteiras
e o desempenho dos fundos de investimento em ações. Revista
Brasileira de Finanças, 12(4), 469-497.
Mossin, J. (1966). Equilibrium in a capital asset market. Econometrica,
34(4), 768-783. doi:10.2307/1910098
Nerasti, J. N., & Lucinda, C. R. (2016). Persistência de desempenho em
fundos de ações no Brasil. Revista Brasileira de Finanças, 14(2), 269-
–297.
Ortas, E., Salvador, M., & Moneva, J. M. (2015). Improved beta
modeling and forecasting: An unobserved component approach with
conditional heteroscedastic disturbances. North American Journal of
Economics and Finance, 31, 27-51. doi:10.1016/j.najef.2014.10.006
Pizzinga, A., & Fernandes C. (2006). State space models for dynamic
style analysis of portfolios. Revista de Econometria, 26(1), 31-66.
doi:10.12660/bre.v26n12006.2497
Sharpe, W. F. (1964). Capital asset prices: A theory of market equilibrium
under conditions of risk. The Journal of Finance, 19(3), 425-442.
doi:10.2307/2977928
Tambosi Filho, E., Garcia, F. G., Imoniana, J. O., & Moreiras, L. M. F. (2010).
Teste do CAPM condicional dos retornos de carteiras dos mercados
brasileiro, argentino e chileno, comparando-os com o mercado
norte-americano. RAE-Revista de Administração de Empresas, 50(1),
60-74. doi:10.1590/S0034-75902010000100006
Treynor, J. L.., & Mazuy, K. K. (1966)). Can mutual funds outguess the
market? Harvard Business Review, 131-136.
Tsay, R. S. (2010). Analysis of Financial Time Series. John Wiley.
Yu, J. (2002). Forecasting volatility in the New Zealand stock market. Applied
Financial Economics, 12(3), 193-202. doi:10.1080/09603100110090118

ANDRE CHERUBINI ALVES1
andre.alves82@gmail.com
ORCID: 0000-0002-4222-5334
BRUNO FISCHER2
bfischer@unicamp.br
ORCID: 0000-0003-3878-9097
NICHOLAS SPYRIDON VONORTAS3
vonortas@gwu.edu
ORCID: 0000-0002-6745-4926
SÉRGIO ROBLES REIS DE QUEIROZ1
squeiroz@ige.unicamp.br
ORCID: 0000-0002-6534-9022
1
Universidade Estadual de
Campinas, Departamento de
Política Científica e Tecnológica,
Campinas, SP, Brazil
2
Universidade Estadual de
Campinas, Faculdade de Ciências
Aplicadas, Limeira, SP, Brazil
³The George Washington
University, Institute for
International Science and
Technology Policy, Department
of Economics, Washington, DC,
United States of America
ARTICLES
Submitted 07.31.2018. Approved 06.04.2019
Evaluated through a double-blind review process. Scientific Editor: Yeda Swirski de Souza
Original version
CONFIGURATIONS OF KNOWLEDGE-INTENSIVE
ENTREPRENEURIAL ECOSYSTEMS
Configurações de ecossistemas de empreendedorismo intensivo em conhecimento
Configuraciones de ecosistemas de emprendimiento intensivo en conocimiento
ABSTRACT
The dominant discourse on Entrepreneurial Ecosystems (EE) remains focused on the profile of a han-
dful of successful locations. This has hindered a deeper comprehension of the economic mechanisms
that shape evolutionary trends in entrepreneurial activity and how they operate in distinct places. We
propose that EE have regularities, but they can also assume different configurations, i.e., varying com-
binations of influential dimensions. Through fuzzy set qualitative comparative analysis, we address this
issue with data from the State of São Paulo, Brazil. This research focuses on five EE dimensions: Science
& Technology, Human Capital, Market Dynamics, Business Dynamics, and Infrastructure. Findings point
at the heterogeneous nature of EE distributed in three different paths. While configurations’ vary in
terms of causal conditions, research universities, knowledge-intensive jobs and wider credit operations
are core-causal conditions. Proximity to the main economic hub appears as a key differentiator among
ecosystems.
KEYWORDS | Entrepreneurial ecosystems, knowledge-intensive entrepreneurship, qualitative compara-
tive analysis, configurations, geography of innovation.
RESUMO
O discurso dominante sobre os ecossistemas de empreendedorismo (EE) enfatiza o perfil de algumas
localidades com reconhecido histórico de sucesso. Isso tem dificultado uma compreensão mais pro-
funda dos mecanismos econômicos que moldam as tendências evolutivas na atividade empreendedora
e como elas operam em lugares distintos. Nós propomos que esses ecossistemas possuem regulari-
dades, mas elas também podem assumir diferentes configurações. Por meio de técnicas de fuzzy-set
Qualitative Comparative Analysis (QCA), abordamos essa questão com dados do estado de São Paulo.
Esta pesquisa concentra-se em cinco dimensões dos ecossistemas de empreendedorismo: ciência e tec-
nologia, capital humano, dinâmica de mercado, dinâmica dos negócios e infraestrutura. Os resultados
apontam para a natureza relativamente heterogênea dos ecossistemas. Não obstante, as universidades
de pesquisa, a intensidade de empregos intensivos em conhecimento e a disponibilidade de crédito são
condições fundamentais. A proximidade do principal centro econômico aparece como um diferencial
importante entre os ecossistemas.
PALAVRAS-CHAVE | Ecossistemas de empreendedorismo, empreendedorismo intensivo em conheci-
mento, qualitative comparative analysis, configurações, geografia da inovação.
RESUMEN
El discurso dominante respecto a los Ecosistemas de emprendimiento (EE) pone énfasis en el perfil de
algunas localidades con reconocido historial de éxito. Esto ha dificultado una comprensión más profunda
de los mecanismos económicos que moldean las tendencias evolutivas en la actividad emprendedora y
cómo ellas operan en lugares distintos. Partimos de la proposición de que estos ecosistemas tienen
regularidades, pero también pueden asumir diferentes configuraciones. A través de técnicas de fuzzy-set
Qualitative Comparative Analysis, abordamos el caso del Estado de São Paulo, Brasil, enfocando cinco
dimensiones de los Ecosistemas de emprendimiento: Ciencia y Tecnología, Capital Humano, Dinámica
de Mercado, Dinámica de Negocios e Infraestructura. Los resultados apuntan a la naturaleza relativa-
mente heterogénea de los ecosistemas. No obstante, las universidades de investigación, la intensidad
de empleos intensivos en conocimiento y la disponibilidad de crédito son condiciones fundamentales.
La proximidad del principal centro económico representa un aspecto diferencial importante entre los
ecosistemas.
PALABRAS CLAVE | Ecosistemas de emprendimiento, emprendimientos intensivos en conocimiento, qua-
litative comparative analysis, configuraciones, geografía de la innovación.

ARTICLES | CONFIGURATIONS OF KNOWLEDGE-INTENSIVE ENTREPRENEURIAL ECOSYSTEMS
INTRODUCTION
Knowledge-intensive entrepreneurship (KIE) refers to a
phenomenon that drives economic competitiveness and innovation
capabilities (Ács, Autio, & Szerb, 2014). It has, however, received
relatively little attention in studies related to the approaches of
innovation systems.The influence of context upon entrepreneurial
activity is still often ignored in favor of a focus on individuals and
firms (Borissenko & Boschma, 2016; Stam, 2015). As a result, we
are still far from developing a thorough understanding of issues
related to the emergence of new technology-based companies and
its systemic determinants (Audretsch, 2012).
It is clear that KIE is unevenly distributed across territories,
a function of heterogeneous endowments in terms of knowledge,
institutions, resources, and demand (Isaksen & Trippl, 2017). As a
consequence, we observe a spiky geography of economic activities
(Florida, 2005). In addition, evidence suggests that the impacts of
entrepreneurial activity are mainly felt at the regional level (Ács &
Armington, 2004), placing Entrepreneurial Ecosystems (EEs) as a
key subject of interest for researchers and policymakers (Audretsch
& Belitski, 2017; Borissenko & Boschma, 2016).
The fact that KIE is deeply embedded in local contexts
(Alvedalen & Boschma, 2017) poses challenges for analysts and
policymakers, as one-size-fits-all initiatives and analytical models
may be deemed inappropriate for most locations – although the
dominant discourse remains focused on the profile of a handful
of successful EEs (Nicotra, Romano, Giudice, & Schillaci, 2018;
Stam, 2015). Accordingly, the economic mechanisms that shape
evolutionary trends in entrepreneurship are not linear, and they
operate differently in distinct locations (Ács, Stam, Audretsch &
O’Connor, 2017; Brown & Mason, 2017; Boschma & Martin, 2010).
Thus, a mechanistic approach to entrepreneurial
ecosystems (EEs) – based on an input-output logic – is likely to
ignore the context-specific traits of regions and their respective
interactions (Feldman, 2001). In this article, we propose that
entrepreneurial ecosystems can have regularities, but they can
also assume different configurations, i.e., varying combinations
of influential dimensions that may generate similar outcomes
in terms of entrepreneurial intensity. This is a function of
the distinct evolutionary path of each location. Our guiding
research questions can be stated as follows: Are there different
configurations of vectors of interest that shape successful
entrepreneurial ecosystems? What are the key “ingredients” of
these combinations? Drawing from different strands of literature
addressing the dynamics of business locations, we examine a
broad set of variables in order to identify the different foundational
patterns behind ecosystems of entrepreneurship.
As a concrete case, we assess the state of São Paulo, Brazil.
We addressKIEthrough data from Pesquisa Inovativa em Pequenas
Empresas(Innovative Research inSmallBusiness[PIPE]) projects, a
program funded by Fundação de Amparo à Pesquisa do Estado de
São Paulo (São Paulo Research Foundation [FAPESP]) thatsupports
innovative initiatives in small enterprises. Our goal is to develop
further knowledge on the evolutionary geography of innovation
and entrepreneurship in the context of a developing country,
acknowledging the substantial discrepancies that these countries
present relative to developed economies when it comes to the
geographyofentrepreneurship (Crescenzi & Rodríguez-Pose, 2012).
The empirical method applied is Qualitative Comparative
Analysis (QCA) with fuzzy sets. We assessed data from a sample of
299 cities in the state ofSão Paulo. The analytical exercise focused
on five dimensions as causal conditions for the emergence of
entrepreneurialecosystems,namely:scienceandtechnology,human
capital, market dynamics, business dynamics, and infrastructure.
Our findings present distinct general paths across different
configurationalsolutions,suggestingtheexistenceofheterogeneous
patterns in entrepreneurial ecosystems. Nonetheless, a common
coreisperceivedacrossthedifferentconfigurations,mainlyinvolving
the science and technology environment, the availability of human
capital, and local market dynamics.
The article is structured as follows: we begin with the
conceptual background of our assessment, as well as our
proposal of a workable analytical model for the case of the state
of São Paulo. After, we discuss the state of the literature and the
need for a more flexible comprehension of the configurations of
entrepreneurial ecosystems. We then outline the method and data
used in our approach and empirical findings. Lastly we offer some
concluding remarks, implications, and avenues for future research.
Knowledge-intensive entrepreneurial
ecosystems: Conceptual background
Regions differ in their propensity to establish knowledge-inten-
sive entrepreneurial activity (Florida, 2005). The resulting patterns
reinforce themselves over time, as geographic proximity functions
as a fundamental vector for knowledge exchange (Alvedalen &
Boschma, 2017). These conditions put significant emphasis on
local-level context as a determinant for KIE emergence, moving
the analytical target beyond the mere understanding of firm-level
capabilities (Mason & Brown, 2014; Ács et al., 2014; Audretsch
& Belitski, 2017).
It is by acknowledging these features of the socioeconomic
environment that the Entrepreneurial Ecosystems concept has

gained ground in recent years. EEs are said to represent a “set of
interconnected entrepreneurial actors, entrepreneurial organiza-
tions, institutions and entrepreneurial processes which formally
and informally coalesce to connect, mediate and govern the per-
formance within the local entrepreneurial environment” (Mason
& Brown, 2014; p. 5). The underlying rationale of entrepreneurial
ecosystems is focused on interactions between agents, tangi-
ble and intangible factors of production, and how these vectors
translate locally into entrepreneurship (Ács et al., 2017; Nicotra
et al., 2018).
Because of its (eco)systemic character, the productivity of
these geographical units is affected by the performance of any of
their components (Ács et al., 2014). By productivity we understand
what Radosevic and Yoruk (2013) call “entrepreneurial propensity,”
i.e., local capacity to generate and exploit innovation-oriented
opportunities through the actions of entrepreneurs.
Figure 1. Knowledge-Intensive Entrepreneurial Ecosystem
SCIENCE AND TECHNOLOGY
HUMANCAPITAL
MARKET
DYNAMICS
BUSIN
ESSDYNAMICSINFRASTRUCTURE
PATENTS
HUMAN
DEVELOPMENT
RESEARCH
UNIVERSITIES
KNOWLEDGE
INTENSIVE LABORKNOWLEDGE
INTENSIVE
ENTREPRENEURSHIP
BUSINESS
CONCENTRATION
MULTINATIONAL
INVESTIMENT
DISTANCE TO
ECONOMIC
HUB
POPULATION
ENERGY
URBANIZATION
GDP
CREDIT
Source: Adapted from Isenberg (2010), Mason and Brown (2014), and Stam (2015).
Based on this literature, and with particular emphasis
on the works of Isenberg (2010), Mason and Brown (2014), and
Stam (2015), we offer a workable model of the entrepreneurial
ecosystem concept (Figure 1). Our adaptation follows the basic
principles contained in existing frameworks, thus contemplating
issues associated to infrastructure, human capital, science
and technology (comprising education, universities, and the
technological support system), markets, and business dynamics.
While the element (local-level) policy is not explicitly addressed
in our model, it is intrinsically associated with features related to
infrastructure, business dynamics, and human capital. The only
dimension missing from our proposed model is entrepreneurial
culture. However, as recent research underscores (Fritsch et al.,
2019), such cultural traits are related to long-standing traditions
and need adequate historical proxies, which are often unavailable
for empirical exercises. Hence, our operational model offers
a comprehensive perspective of EEs according to previous
theoretical proposals, while remaining parsimonious as to
address the entrepreneurial dynamics of ecosystems.
Because of the local nature of EEs, the literature has
recognized cities as the most adequate loci for empirical
assessments (Audretsch & Belitski, 2017). Next, we address
each of the five dimensions set out in the model in further detail.
Science and technology
Amongst the factors related to locational dynamics, access to a
relevant knowledge base can be considered fundamental – more
so for knowledge-intensive entrepreneurship (Nicotra et al., 2018;
Boschma & Martin, 2010). Rich technological environments tend
to facilitate entrepreneurial activity at the local level (Nicotra et
al., 2018). In the state of São Paulo, Brazil, city-level patenting
activity is strongly related to the emergence of KIE activity (Fischer,
Queiroz, & Vonortas, 2018).
Universities and research institutes are strategic agents
in this respect (Spigel, 2017; Stam, 2015). These institutions not
only contribute to knowledge generation, they also shape local
conditions related to the population’s educational attainment
(Isaksen & Trippl, 2017; Dorfman, 1983). Therefore, geographical
proximity to research-oriented universities and research
centers can be a valuable source of knowledge for high-tech
entrepreneurial activity (Stam, 2015).
Moreover, science-based entrepreneurship is significantly
related to academic spin-offs (Di Gregorio & Shane, 2003), a
situation that attributes key relevance to the local presence of
preeminent universities. Guerrero et al. (2016) proposed that
universities and academic researchers are fundamental agents
of innovation systems through their involvement with knowledge
transfer and entrepreneurial activities. This helps explain why
high-tech clusters are often attached to university campus towns
(Isaksen & Trippl, 2017). Fitjar and Rodríguez-Pose (2011) make
such a case for regions that lack agglomeration economies while
presenting high relative levels of innovative activity. However,
universities’ impact on strong entrepreneurial ecosystems is

argued to be contingent upon the existence of a socioeconomic
environment conducive to firm entry (Patton & Kenney, 2010).
Human capital
Chatterji, Glaeser, and Kerr (2013) state that the key pillars of
entrepreneurship are essentially related to human capital. This
is associated with the general education of the workforce and
the local supply of individuals with entrepreneurial inclinations.
This feature is widely recognized in the literature, be it under
the concepts of human capital (Isenberg, 2011; WEF, 2014),
worker talent (Spigel, 2017), human resources (Nicotra et al.,
2018), or talent pool (Stam, 2015). Notably, the ratio of high-tech
employment often predicts entrepreneurial activity in regions
(Motoyama & Danley, 2012).
The availability of highly skilled labor is a structural
precondition for the creation of innovative entrepreneurial
systems (Bresnahan, Gambardella, & Saxenian, 2001). Also,
although labor is theoretically considered a mobile resource,
entrepreneurial clusters are more likely to arise wherever
professional talent is located or where it can be easily attracted
(Dorfman, 1983). Accordingly, Audretsch and Feldman (1996)
found that the propensity to cluster is stronger in industries
that rely heavily on new economic knowledge, depending to
a large degree upon skilled labor. This proximity allows KIE
companies to more easily access available human capital
(Storper, 1995).
In addition, opportunity-driven entrepreneurial activity is
strongly related to local income levels (Radosevic & Yoruk, 2013).
Income approximates the level of education and capabilities
within a pool of individuals and the quality of life and overall level
of development of regions, key drivers of the location choices of
knowledge-intensive entrepreneurs (Florida & Mellander, 2014).
Market dynamics
The third dimension of interest in our EE model concerns market
dynamics. Despite the likelihood that knowledge-intensive
new ventures will be oriented towards global markets, local-
level conditions shape such ventures’ access to necessary
complementary resources (Florida & Mellander, 2014). Larger
markets offer increased levels of opportunities (Armington & Ács,
2002) and demand diversity (Bosma, Schott, Terjesen, & Penny,
2015). Isenberg (2010) follows this rationale, including local
markets as significant vectors in the formation of entrepreneurial
ecosystems. In addition, evidence from developed countries
suggests that entrepreneurship tends to be more concentrated
in large urban areas (Bosma & Sternberg, 2014). Market size also
feeds entrepreneurial endeavors from a different direction: the
size of the population is a representation of the pool of potential
entrepreneurs in a given location (Stuart & Sorenson, 2003).
Our category of market dynamics also includes the available
funding for knowledge-intensive entrepreneurs (Isenberg, 2010).
Geographic proximity is a critical feature in this discussion, as
larger distances between capitalists and entrepreneurs increase
the monitoring costs involved in financing operations (Dorfman,
1983). This reflects the strategic relevance attributed to credit
as a platform for start-ups to operate (Lerner, 2002; Feldman,
2001). Besides private flows of capital, governments can engage
in financing these incipient ventures (Lerner, 2002). Considering
the Brazilian context, this vector is particularly critical, as small
and medium-sized companies find it excessively difficult to access
funding mechanisms due to scarce credit lines (Neto, Farias Filho,
& Quelhas, 2014).
Business dynamics
Complementary to the idea of market dynamics, we introduce a
perspective on the business environment of cities as an influential
component of entrepreneurial ecosystems. Included here are
the aspects of the level of development of regional economic
structures, as well as its attractiveness for investments by
incumbents.
The literature recognizes the importance of adjacent firms
as a driver of local competitiveness (Stuart & Sorenson, 2003).
Positive externalities arise from the local clustering of firms,
generating a critical mass of support for new companies (Isenberg,
2010; Isaksen & Trippl, 2017; Storper, 1995). Incumbent firms can
also leverage the growth conditions for new firms (WEF, 2014)
providing incentives for new business creation and contributing
to start-up survival (Delgado et al., 2010). Established firms, and
particularly multinational corporations, also play an important
role in setting the stage for the emergence of new, knowledge-
intensive ventures (Brown & Mason, 2017), as they can leverage
overall capabilities in clusters (Bresnahan et al, 2001).
An additional item of interest concerns the geographical
reach of large (“core”) markets and how the proximity to these
hubs can exert effects in neighboring areas.This happens because
large cities do not just represent agglomeration of people, they are
arguably also associated with the generation of innovative, science-
based entrepreneurship (Duranton & Puga, 2002). In thisregard, an

“efficiency gap” exists when comparing peripheral regions to those
located close to large markets (Crescenzi & Rodríguez-Pose, 2012).
This is why high-tech clusters of entrepreneurship are frequently
located in core regions (Isaksen & Trippl, 2017). This situation
can be magnified in regions such as the state of São Paulo, which
includes a megahub that functions as a center for business services,
venture capital, and corporate demand (WEF, 2018).
Infrastructure
The fifth dimension in our model deals with the quality of
infrastructure, a platform upon which economic activity strongly
relies. Infrastructure facilitates urban linkages, labor mobility,
and knowledge flows (Audretsch et al., 2015). Thus, its impacts
as a determinant of the formation of entrepreneurial ecosystems
must be acknowledged (Spigel, 2017; Nicotra et al., 2018; Stuart
& Sorenson, 2003). The quality of physical infrastructure can also
mitigate the detrimental effects associated with large market
agglomeration diseconomies (Audretsch & Belitski, 2017).
Knowledge-intensive entrepreneurial ecosystems:
Making the case for heterogeneous configurations
Notwithstanding agreement on the very basic definitions of
entrepreneurial ecosystems, there are persistent controversies
on causal links within its intrinsic dynamics (Borissenko &
Boschma, 2016). What dimensions really matter for EEs? One of
the main issues here concerns the poor generalizability of highly
eminent cases to other contexts. These expositions – although
informative – are developed around an idea of relative stability in
the configuration of influential attributes in EEs. A first step in our
approach deals with assessing the validity of such expectations.
Hence, our first proposition can be stated as follows:
Proposition 1. Successful Entrepreneurial Ecosys-
tems rely on a set of critical dimensions that shape
their respective capacity of sustaining the genera-
tion of knowledge-intensive entrepreneurship.
On the other hand, researchers have increasingly criticized
such formulations, as they lead to one-size-fits-all implications.
Even if the dimensions included in these models are inclusive,
some of its attributes may be more dominant in some cases
than in others (Spigel, 2017). The central argument here is that
economic mechanisms operate differently in distinct locations
as a function of their historical backgrounds (Boschma & Martin,
2010). This is why top-down policies that aim to organize clusters
of entrepreneurship are often deemed ineffective (Bresnahan et
al, 2001; Chatterji et al, 2013; Feldman, 2001).
This is the pillar of Evolutionary Economic Geography (EEG).
In dealing with the dynamics of entrepreneurial ecosystems,
the evolutionary view pinpoints the relevance of pre-existing
conditions and assets for ecosystems to emerge (Isaksen, 2016).
Accordingly, new science-based firms can be understood as
functions of the prior existence of scientific research undertaken
by universities and research institutes in a given location (Feldman
& Lendel, 2011). In this debate, we hope to add the perspective
that, because of distinct evolutionary paths, entrepreneurial
ecosystems can achieve efficiency through different configurations.
In other words, the relevance of EEs’ dimensions is bounded
by location-specific trajectories, thus altering the dynamics of
interconnection among actors, institutions, and organizations.
Accordingly, our second proposition is presented:
Proposition 2. Because of idiosyncratic evolu-
tionary paths, Entrepreneurial Ecosystems can
present heterogeneous configurations in terms
of relevant drivers without compromising their
respective level of entrepreneurial propensity.
Recent literature has indicated some efforts in a similar
direction. Brown and Mason (2017) develop a simplified taxonomy
of entrepreneurial ecosystems based on illustrative cases of
“embryonic” and “scale-up” locations. By assessing this argument,
our goal is to empirically refine these introductory propositions
and provide a more nuanced view on the variegated combinations
of characteristics that can form a functional EE.
METHODOLOGICAL APPROACH
To address our research questions, we use Fuzzy-Set Qualitative
Comparative Analysis (QCA). QCA is a method that is used
to identify configurations or “recipes” of causal conditions
associated with different outcomes, following the “equifinality
principle,” meaning that multiple paths or solutions can lead
to the same outcome (Ragin, 2008). Different from regression
analysis, QCA also follows the principle of causal complexity,
accounting for the combination of causal measures to a specific
outcome within a property space. Analytical benefits associated
with the QCA approach in comparison with standard econometric
techniques applied to address the dynamics of entrepreneurial
ecosystems concern its capacity for developing robust evaluations

of configurational issues (Fiss et al., 2013). While this is an aspect
of interest in this field, traditional regression models fall short
in offering the necessary knowledge for such research questions.
Moreover, there is an increasing interest in the use of QCA in
entrepreneurship studies (Kraus et al., 2018).
The basic locus of empirical information in this study is
knowledge-intensive entrepreneurship in the state of São Paulo,
Brazil. The grants of the PIPE program are used as a proxy for KIE
activity. This program is funded by the Fundação de Amparo à
Pesquisa do Estado de São Paulo (Research Foundation of the
state of São Paulo [FAPESP]) to support innovative initiatives
in small enterprises. The program has a similar structure and
objectives to the Small Business Innovation Research (SBIR)
program in the United States. While SBIR data has been used in
the context of entrepreneurial ecosystems by an extensive body of
research (e.g. Wallsten, 2001; Qian & Haynes, 2014), information
from PIPE has only recently been used for these purposes (e.g.
Fischer et al., 2018). Such sources provide robust evidence on KIE
activity, but also introduce sample bias in the analysis, since they
deal with pre-selected and funded R&D projects that are often
attached to academic spin-offs. Hence, conclusions drawn from
this group should be taken cautiously, as they do not necessarily
represent the broader context of overall KIE firms.
The full dataset includes 1130 grants allocated across 114
cities in the state during the period 1998-2014. We collected
data for a total of 299 cities in the state of São Paulo in order
to account for differences in configurations between cities that
have received project grants and those that have not. City-level
analysis was chosen due to the local nature of entrepreneurial
ecosystems, in which cities seem to be the most appropriate units
of analysis (Audretsch & Belitski, 2017). While QCA has typically
been used as a research methodology for Small-N, comparative
work in the case-oriented tradition (Ragin, 1987), the method has
increasingly been applied to examine large-N phenomena (Fiss,
et al., 2013; Greckhamer et al., 2013; Emmenegger et al., 2014).
Small-N samples in QCA range from 12-50 cases, whereas, Large-N
samples involve 50+ cases (Greckhamer et al., 2013).
Model and unit of analysis
Following the literature review and the graphical representation
of our analytical model (Figure 1), the socioeconomic dimensions
included in our assessment are described in Exhibit 1. Most
variables represent averages of city-level characteristics as proxies
for economic conditions. This procedure avoids problems related
to year-to-year variations, while also controlling for the time span
during which projects have started (1998-2014).
Calibration procedures
Qualitative Comparative Analysis works using the principle of
set-theoretic “membership.” We used quartiles to calibrate the
fuzzy-sets. The only exception to this rule was the causal condition
RESUNI, which is a binary variable indicating the presence
(absence) of a major research university in the city. For the variable
PROXCAP, quartiles were obtained by using only the cities that
have displayed the outcome. This was done with the intent of
creating a better visual geographical distribution and separation
of cases. For the outcome variable, a rather flexible threshold
was used to generate enough diversity in the sample and avoid
skewness. We used the following set-membership threshold
for our outcome variable: above 10 projects (full membership),
between 4 to 10 projects (more-in-than-out), between 2 to 4 (more-
out-than-in), and below 2 (full non-membership).
Necessary conditions
Conditions are considered necessary when they represent the
superset of the outcome, that is, when the set membership
values for the outcome Y are lower than that for a given causal
condition (Ragin, 2006). Necessary conditions are verified using
the following formulas for Consistency and Coverage.
Consistency: ∑{min(Xi
, Yi
)]/∑(Yi
) (1)
Covergage: (Yi
≤ Xi
): ∑{min(Xi
, Yi
)]/∑(Xi
) (2)
According to Ragin (2008, p. 44), “consistency addresses
the degree to which instances of the outcome agree in displaying
the causal condition to be necessary, while coverage assesses
the degree to which instances of the condition are paired with
instances of the outcome.” Coverage indicates the relevance or
importance of a set-theoretic connection. However, necessary
conditions may be considered trivial if a condition is present in
most cases whether the outcome is present or not (Schneider &
Wagemann, 2012).
Sufficient conditions
The second test intends to identify the combinations of conditions
that are in accordance with the presence of KIE activity, thus
forming knowledge-intensive entrepreneurial ecosystems.
A sufficient condition is one that, if satisfied, results in the
achievement of the outcome (Schneider & Wagemann, 2012,

Título: RAE-Revista de Administração de Empresas (Journal of Business Management), 2019. V. 59, N. 4

Título: RAE-Revista de Administração de Empresas (Journal of Business Management), 2019. V. 59, N. 4

Recomendados

Recomendados

Más contenido relacionado

Similar a Título: RAE-Revista de Administração de Empresas (Journal of Business Management), 2019. V. 59, N. 4

Similar a Título: RAE-Revista de Administração de Empresas (Journal of Business Management), 2019. V. 59, N. 4 (20)

Más de FGV | Fundação Getulio Vargas

Más de FGV | Fundação Getulio Vargas (20)

Último

Último (20)

Título: RAE-Revista de Administração de Empresas (Journal of Business Management), 2019. V. 59, N. 4