Marcelo Fernandes
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Marcelo Fernandes has a total of 43 co-authors. Here are the top co-authors by number of shared publications:
- Adrian Pizzinga
Institute of Mathematics and Statistics, Fluminense Federal University (UFF), Niteroi, Brazil
2 shared publications Fausto Vieira
2 shared publicationsJoachim Grammig
Department of Economics, University of Tübingen, Mohlstr. 36, D-72074 Tübingen, Germany
2 shared publications- Olivier Scaillet
Swiss Finance Institute - University of Geneva
2 shared publications - Walter Distaso
Imperial College Business School
2 shared publications Alvaro Veiga
1 shared publication
Latest Publications
- Open Access
- 3 August 2026
- Published by Elsevier in SSRN Electronic Journal
Upside risk and return timing in Bitcoin
This paper studies the benefits of timing Bitcoin returns by upside and downside volatilities. Standard volatility management implicitly treats volatility spikes as signals of adverse states, reducing exposure when total volatility increases. However, in Bitcoin, volatility spikes are frequently due to price rallies, which typically indicate subsequent positive returns. We show that semivolatility timing rules that account for both downside and upside risk concerns yield substantially stronger risk-adjusted performance than buy-and-hold and volatility-managed strategies. This stems from the fact that high upside-driven volatility states in Bitcoin are disproportionately associated with positive returns in the next period.
- 3 August 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessUpside risk and return timing in Bitcoin
This paper studies the benefits of timing Bitcoin returns by upside and downside volatilities. Standard volatility management implicitly treats volatility spikes as signals of adverse states, reducing exposure when total volatility increases. However, in Bitcoin, volatility spikes are frequently due to price rallies, which typically indicate subsequent positive returns. We show that semivolatility timing rules that account for both downside and upside risk concerns yield substantially stronger risk-adjusted performance than buy-and-hold and volatility-managed strategies. This stems from the fact that high upside-driven volatility states in Bitcoin are disproportionately associated with positive returns in the next period.
- Open Access
- 19 April 2026
- Published by Fundacao Getulio Vargas in Brazilian Review of Finance
- Vol. 24, e202606
Portfolio allocation under sovereign risk
This paper examines a portfolio allocation strategy in which investment decisions depend on whether sovereign risk is high or low. We gauge sovereign risk by means of the spread on the 5-year credit default swap (CDS). The central premise is that the CDS spread provides a timely, market-based proxy for institutional risk, enabling portfolio weights to adjust dynamically across low- and high-risk environments. We show that the mean-variance portfolio conditional on the CDS spread regime entails a better risk-adjusted performance in the Brazilian market than the unconditional mean-variance portfolio and equal-weighted portfolio. This is particularly true for periods of high sovereign risk.
- 19 April 2026
- Published by Fundacao Getulio Vargas in Brazilian Review of Finance
- Vol. 24, e202606
Open AccessPortfolio allocation under sovereign risk
This paper examines a portfolio allocation strategy in which investment decisions depend on whether sovereign risk is high or low. We gauge sovereign risk by means of the spread on the 5-year credit default swap (CDS). The central premise is that the CDS spread provides a timely, market-based proxy for institutional risk, enabling portfolio weights to adjust dynamically across low- and high-risk environments. We show that the mean-variance portfolio conditional on the CDS spread regime entails a better risk-adjusted performance in the Brazilian market than the unconditional mean-variance portfolio and equal-weighted portfolio. This is particularly true for periods of high sovereign risk.
- Open Access
- 22 January 2026
- Published by Fundacao Getulio Vargas in Brazilian Review of Finance
- Vol. 24, e202601
Identificando sinais esparsos para o Ibovespa
Neste trabalho, prevemos o Ibovespa um passo à frente usando os retornos passados de todas as suas ações constituintes. Consideramos retornos nas frequências de 1, 5 e 10 minutos em janelas móveis de 30 observações. Adotamos uma regularização via LASSO para reduzir a dimensionalidade do problema e identificar sinais esparsos. A nossa estratégia de previsão traz ganhos estatisticamente significativos de poder preditivo em relação a modelos autorregressivos, embora os ganhos líquidos sejam limitados pelos custos de transação.
- 22 January 2026
- Published by Fundacao Getulio Vargas in Brazilian Review of Finance
- Vol. 24, e202601
Open AccessIdentificando sinais esparsos para o Ibovespa
Neste trabalho, prevemos o Ibovespa um passo à frente usando os retornos passados de todas as suas ações constituintes. Consideramos retornos nas frequências de 1, 5 e 10 minutos em janelas móveis de 30 observações. Adotamos uma regularização via LASSO para reduzir a dimensionalidade do problema e identificar sinais esparsos. A nossa estratégia de previsão traz ganhos estatisticamente significativos de poder preditivo em relação a modelos autorregressivos, embora os ganhos líquidos sejam limitados pelos custos de transação.
- 23 June 2025
- Published by Elsevier in The Quarterly Review of Economics and Finance
- Vol. 103
The equity premium and the disconnect between uncertainty and volatility: A global perspective
- 23 June 2025
- Published by Elsevier in The Quarterly Review of Economics and Finance
- Vol. 103
The equity premium and the disconnect between uncertainty and volatility: A global perspective
- 10 June 2022
- Published by Taylor & Francis in Communications in Statistics - Theory and Methods
- Vol. 52 (24) , 8884-8893
Diffuse Kalman filtering with linear constraints on the state parameters
We give a general proof, based on Piet de Jong’s diffuse Kalman filter, that imposing linear constraints on state smoothing is still feasible under diffuse initialization. We also offer simple derivations of existing identities related to the diffuse Kalman filter and smoother.
- 10 June 2022
- Published by Taylor & Francis in Communications in Statistics - Theory and Methods
- Vol. 52 (24) , 8884-8893
Diffuse Kalman filtering with linear constraints on the state parameters
We give a general proof, based on Piet de Jong’s diffuse Kalman filter, that imposing linear constraints on state smoothing is still feasible under diffuse initialization. We also offer simple derivations of existing identities related to the diffuse Kalman filter and smoother.
- 19 November 2020
- Published by Wiley in Journal of Time Series Analysis
- Vol. 42 (3) , 355-371
Extensions to the invariance property of maximum likelihood estimation for affine‐transformed state‐space models
Replacing the state vector of a linear state‐space model by any one‐to‐one linear transformation does not alter maximum likelihood estimation. We extend this invariance property to more general settings, with possibly diffuse initialization of the Kalman filter and injective affine transformations of the state vector. Our results hold for both direct maximization of the likelihood function and the EM algorithm. We offer two real examples that illustrate how one may employ our results to handle a variety of affine‐transformed state‐space models in the literature.
- 19 November 2020
- Published by Wiley in Journal of Time Series Analysis
- Vol. 42 (3) , 355-371
Extensions to the invariance property of maximum likelihood estimation for affine‐transformed state‐space models
Replacing the state vector of a linear state‐space model by any one‐to‐one linear transformation does not alter maximum likelihood estimation. We extend this invariance property to more general settings, with possibly diffuse initialization of the Kalman filter and injective affine transformations of the state vector. Our results hold for both direct maximization of the likelihood function and the EM algorithm. We offer two real examples that illustrate how one may employ our results to handle a variety of affine‐transformed state‐space models in the literature.
- Open Access
- 25 March 2020
- Published by Fundacao Getulio Vargas in Brazilian Review of Econometrics
- Vol. 39 (2) , 303-326
The efficiency of risk sharing between UK and US: Robust estimation and calibration under market incompleteness
We propose a novel estimator for the amount of international risk sharing that depends exclusively on asset returns data. In particular, our estimator has a nonparametric flavor in that it makes no parametric assumption on preferences and on the stochastic process that governs the dynamics of asset returns. This is in contrast with the existing estimators in the literature that either assume a specific utility function or that asset returns follow a geometric Brownian motion (GBM). Our estimates reveal there is less risk sharing between UK and US than one would find under the GBM assumption, though much more than what consumption data might suggest. Moreover, a simple calibration analysis shows that market incompleteness alone is enough to explain the difference between the consumption-based estimate of the risk-sharing index and ours.
- 25 March 2020
- Published by Fundacao Getulio Vargas in Brazilian Review of Econometrics
- Vol. 39 (2) , 303-326
Open AccessThe efficiency of risk sharing between UK and US: Robust estimation and calibration under market incompleteness
We propose a novel estimator for the amount of international risk sharing that depends exclusively on asset returns data. In particular, our estimator has a nonparametric flavor in that it makes no parametric assumption on preferences and on the stochastic process that governs the dynamics of asset returns. This is in contrast with the existing estimators in the literature that either assume a specific utility function or that asset returns follow a geometric Brownian motion (GBM). Our estimates reveal there is less risk sharing between UK and US than one would find under the GBM assumption, though much more than what consumption data might suggest. Moreover, a simple calibration analysis shows that market incompleteness alone is enough to explain the difference between the consumption-based estimate of the risk-sharing index and ours.
- 4 October 2019
- Published by Taylor & Francis in Journal of Business & Economic Statistics
- Vol. 39 (1) , 338-357
Smoothing Quantile Regressions
We propose to smooth the objective function, rather than only the indicator on the check function, in a linear quantile regression context. Not only does the resulting smoothed quantile regression estimator yield a lower mean squared error and a more accurate Bahadur-Kiefer representation than the standard estimator, but it is also asymptotically differentiable. We exploit the latter to propose a quantile density estimator that does not suffer from the curse of dimensionality. This means estimating the conditional density function without worrying about the dimension of the covariate vector. It also allows for two-stage efficient quantile regression estimation. Our asymptotic theory holds uniformly with respect to the bandwidth and quantile level. Finally, we propose a rule of thumb for choosing the smoothing bandwidth that should approximate well the optimal bandwidth. Simulations confirm that our smoothed quantile regression estimator indeed performs very well in finite samples.
- 4 October 2019
- Published by Taylor & Francis in Journal of Business & Economic Statistics
- Vol. 39 (1) , 338-357
Smoothing Quantile Regressions
We propose to smooth the objective function, rather than only the indicator on the check function, in a linear quantile regression context. Not only does the resulting smoothed quantile regression estimator yield a lower mean squared error and a more accurate Bahadur-Kiefer representation than the standard estimator, but it is also asymptotically differentiable. We exploit the latter to propose a quantile density estimator that does not suffer from the curse of dimensionality. This means estimating the conditional density function without worrying about the dimension of the covariate vector. It also allows for two-stage efficient quantile regression estimation. Our asymptotic theory holds uniformly with respect to the bandwidth and quantile level. Finally, we propose a rule of thumb for choosing the smoothing bandwidth that should approximate well the optimal bandwidth. Simulations confirm that our smoothed quantile regression estimator indeed performs very well in finite samples.
- 1 September 2019
- Published by Elsevier in Journal of Economic Dynamics and Control
- Vol. 106
A dynamic Nelson–Siegel model with forward-looking macroeconomic factors for the yield curve in the US
- 1 September 2019
- Published by Elsevier in Journal of Economic Dynamics and Control
- Vol. 106
A dynamic Nelson–Siegel model with forward-looking macroeconomic factors for the yield curve in the US
- Open Access
Smoothing quantile regressions
We propose to smooth the entire objective function, rather than only the check function, in a linear quantile regression context. Not only does the resulting smoothed quantile regression estimator yield a lower mean squared error and a more accurate Bahadur-Kiefer representation than the standard estimator, but it is also asymptotically differentiable. We exploit the latter to propose a quantile density estimator that does not suffer from the curse of dimensionality. This means estimating the conditional density function without worrying about the dimension of the covariate vector. It also allows for two-stage efficient quantile regression estimation. Our asymptotic theory holds uniformly with respect to the bandwidth and quantile level. Finally, we propose a rule of thumb for choosing the smoothing bandwidth that should approximate well the optimal bandwidth. Simulations confirm that our smoothed quantile regression estimator indeed performs very well in finite samples.
Open AccessSmoothing quantile regressions
We propose to smooth the entire objective function, rather than only the check function, in a linear quantile regression context. Not only does the resulting smoothed quantile regression estimator yield a lower mean squared error and a more accurate Bahadur-Kiefer representation than the standard estimator, but it is also asymptotically differentiable. We exploit the latter to propose a quantile density estimator that does not suffer from the curse of dimensionality. This means estimating the conditional density function without worrying about the dimension of the covariate vector. It also allows for two-stage efficient quantile regression estimation. Our asymptotic theory holds uniformly with respect to the bandwidth and quantile level. Finally, we propose a rule of thumb for choosing the smoothing bandwidth that should approximate well the optimal bandwidth. Simulations confirm that our smoothed quantile regression estimator indeed performs very well in finite samples.