Leopold Sögner

Current Research (updated Febuary 2023)

E-Mail: soegner@ihs.ac.at

ORCID-ID: https://orcid.org/0000-0001-5388-0601


“Stock-Oil Comovement: Fundamentals or Financialization?”

With Alessandro Melone, Otto Randl, and Josef Zechner

Abstract

The return correlation between U.S. stocks and oil has shifted from negative to positive since 2008. We use a return decomposition framework to demonstrate that the underlying reason for this structural change is a shift in the correlation between cash flow news for the two assets. Intuitively, as the U.S. turned from an oil importer to a net exporter, the correlation between the cash flow news associated with oil and the U.S. stock market turned positive. Our findings help to understand the set of potential determinants of equity-commodity correlations and the diversification benefits of investing in commodities.


Download from SSRN:
https://ssrn.com/abstract=4205724


“Extending the Demand System Approach to Asset Pricing”

With Thomas Gehrig and Arne Westerkamp

Abstract

We extend the demand systems approach of Koijen and Yogo (2019) to more general classes of preferences. Specifically, we analyse constant absolute and constant relative risk aversion, provide conditions for the existence of equilibrium, and evaluate equilibrium prices at US-data. We find that constant absolute risk aversion works particularly well at moderate levels of risk aversion. In the case of relative risk aversion, optimal interior portfolio solutions may not even exist. In both preference classes especially out-of-sample predictions are rather volatile. In order to improve out-of-sample performance we augment the optimal strategies by a shrinkage device. As a side product we establish that the characteristics-based parametric portfolio approach of Brandt, Santa Clara and Valkanov (RFS 2009) can only be justified as optimal investments under exceedingly strong assumptions. In empirical data the shrinkage approach outperforms the parametric approach and the naive 1/N -strategy over quite a wide range of levels of absolute and relative risk aversion.


Download from SSRN:
https://ssrn.com/abstract=4292475



 "Fully Modified Estimation of Spatially Correlated Cointegrated Systems"

 

With Martin Wagner

Abstract

We consider a system of spatially correlated cointegrating relationships. Deterministic trend terms are also allowed for and it is shown that the convergence rate of the spatial correlation parameter is determined by the order of the trend polynomial. In addition to the correlation induced by the spatial autoregressive formulation we also allow for cross-unit correlation of the integrated regressors as well as the error terms. Cointegration amongst the regressors is not allowed, as is standard in the cointegrating regression literature. In order to obtain limiting distributions that allow for standard asymptotic inference, the fully modified OLS estimation principle is extended to the current situation. Finally, the methodology is applied to a credit risk data set.



“Global-VAR Modeling with Mixed Frequency Data”
 
Abstract
This article considers a global vector auto-regressive model (G-VAR) developed in Di Mauro et al. (2007) as well as Chudik and Pesaran (2016). In this set-up we allow for  common variables as well as foreign variables to influence the country specific variables. The variables are considered to be integrated of order one. We allow for $r$ cointegrating relationships. Since empirical data used in G-VAR modeling are available at different frequencies we extend the approach described in Chudik and Pesaran (2016) to the mixed frequency case. We follow Seong et al. (2013) and construct an expectation-maximization (EM) based parameter estimation procedure. For the global model forecasting as well as generalized impulse response analysis is performed. 


“Optimal High-Risk Investment

With Martin Meier and Gregor Kastner

Abstract
This article augments a Bruss (2002)-type investment model, where  n investor observes a finite sequence of investment alternatives endowed with a quality characteristic. The information available is the current and all prior quality characteristics.  The investor has to decide whether to invest in the same period the project shows up. Finally, after the last investment alternative had shown up, only the projects with the best and the second best quality characteristic generate positive returns, while the payoffs of all other projects are zero. These returns are either a-priori known or stochastic. Under these assumptions this article derives the value functions and optimal investment rules for risk-neutral or risk averse investors. A simulation study demonstrates how optimal investment decisions are affected by the time horizon and by the attitudes towards risk.



“Hunting for Superstars

With Martin Meier

Abstract
An investor faces a sequence of high-risk investment opportunities. The investor ranks the corresponding projects seen so far and must immediately decide whether and how much to invest into the currently observed opportunity. Returns are realized at the end of the investment horizon, where only a small number of superstars, that is top ranked projects, generate positive returns. The payoffs of all other opportunities are zero. Under these assumptions we derive the value functions and optimal investment rules for risk-neutral or risk averse investors. As well, we obtain a few results on the comparative statics of these investment problems. A simulation study is performed for risk-neutral and risk-averse investors.

Download from SSRN: https://ssrn.com/abstract=4148313


"Bayesian Reconciliation of Return Predictability”

With Borys Koval and Sylvia Frühwirth-Schnatter

Abstract
This article considers a stable vector autoregressive (VAR) model and investigates return predictability in a Bayesian context. The VAR system comprises asset returns and the dividend-price ratio as proposed in Cochrane (2008), and allows pinning down the question of return predictability to the value of one particular model parameter. We develop a new shrinkage type prior for this parameter and compare our Bayesian approach to ordinary least squares estimation and to the reduced-bias estimator proposed in Amihud and Hurvich (2004). A simulation study shows that the Bayesian approach dominates the reduced-bias estimator in terms of observed size (false positive) and power (false negative). We apply our methodology to annual CRSP value-weighted returns running, respectively, from 1926 to 2004 and from 1953 to 2021. For the first sample, the Bayesian approach supports the hypothesis of no return predictability, while for the second data set weak evidence for predictability is observed.

Download from SSRN: https://ssrn.com/abstract=4288973


 

“REMUS: Generic Identifiability for the Cointegrated Unit Root VAR

With Philipp Gersing and Manfred Deistler

Abstract
The “REtrieval from MUltiperiod Systems” (REMUS) approach based on blocking developed in Anderson et al. (2016) is concerned with retrieving an underlying high frequency model from mixed frequency observations. In this paper we investigate parameter-identifiability in the Johansen (1995) vector error correction model for mixed frequency data. We prove that from the second moments of the blocked process after
taking differences at lag N (N is the slow sampling rate), the parameters of the high frequency system are generically identified. We treat the stock and the flow case as well as deterministic terms.

Download from SSRN: https://ssrn.com/abstract=4068560


Financial and Economic Uncertainty and Their Effects on the Economy”

With Ines Fortin and Jaroslava Hlouskova

Abstract
We estimate new indices measuring financial and economic uncertainty in the euro area, Germany, France, the United Kingdom and Austria, following the approach of Jurado, Ludvigson and Ng (2015), which measures uncertainty by the degree of predictability. We perform an impulse response analysis in a vector error correction framework, where we focus on the impact of both local and global uncertainty shocks on industrial production, employment and the stock market. We find that global financial and economic uncertainty have significant negative effects on local industrial production, employment, and the stock market, while we find hardly any influence of local uncertainty on these variables. In addition we perform a forecasting analysis, where we assess the merits of uncertainty indicators for forecasting industrial production, employment and the stock market, using different performance measures. The results suggest that financial uncertainty significantly improves the forecasts of the stock market in terms of profit-based measures, while economic uncertainty gives, in general, more insight when forecasting macroeconomic variables.

Download from SSRN: https://ssrn.com/abstract=4349929


 

 


Recent Publications  2015-

 

 

“Deviations from Triangular Arbitrage Parity in Foreign Exchange and Bitcoin Markets” with J. Reynolds and M. Wagner.

 

Central European Journal of Economic Modelling and Econometrics, Vol. 13(2), pp. 105-146. DOI: 10.24425/cejeme.2021.137358

 

Abstract

This paper applies recently developed procedures to monitor and date so-called “financial market dislocations”, defined as periods in which substantial deviations from arbitrage parities take place. In particular, we use a cointegration perspective to focus on deviations from the triangular arbitrage parity for exchange rate triplets. Due to increasing attention on and importance of mispricing in the market for cryptocurrencies, we include the cryptocurrency Bitcoin in addition to fiat currencies in our analysis. We do not find evidence for substantial deviations from the triangular arbitrage parity when only traditional fiat currencies are considered, but document significant deviations from triangular arbitrage parities in the newer market for Bitcoin. We tentatively confirm the importance of our results for portfolio strategies by showing that a currency portfolio that trades based on our detected break-points outperforms a simple buy-and-hold strategy.

 



"Generalized Method of Moment based Parameter Estimation of Affine Term Structure Models", with Jaroslava Hlouskova

With J. Hlouskova, Econometrics and Statistics, 2020, Vol. 13, pp. 2-15. https://doi.org/10.1016/j.ecosta.2019.10.001



Abstract

This article studies dynamic panel data models in which the long run outcome for a particular cross-section is affected by a weighted average of the outcomes in the other cross-sections. We show that imposing such a structure implies a model with several cointegrating relationships that, unlike in the standard case, are nonlinear in the coecients to be estimated. Assuming that the weights are exogenously given, we extend the dynamic ordinary least squares methodology and provide a dynamic two-stage least squares estimator. We derive the large sample properties of our proposed estimator under a set of low-level assumptions. Then our methodology is applied to US financial market data, which consist of credit default swap spreads, as well as firm-specific and industry data. We construct the economic space using a “closeness” measure for firms based on input–output matrices. Our estimates show that this particular form of spatial correlation of credit default swap spreads is substantial and highly significant.



"An Exploratory Analysis on the Risk to be Offended on the Internet", with Susanne Kirchner

Archives of Data Science, Series A, 2018, Vol.. 3(1), pages 1-26,

https://doi.org/10.5445/KSP/1000083488/02

Abstract

Questionnaire data is used to identify socio-demographic as well as the risk-awareness characteristics of users offended on the Internet. The data comprises a representative sample of 3,000 individuals, containing information on employment, education, age, the frequency of Internet usage and security measures taken by the users. By means of a cluster analysis, within the sub-sample of offended users, we identify a female group, where employment and education are high, a male cluster with similar characteristics,
a group of urban users with low security awareness and a group of young users. Regressions show that the frequency of using the Internet increases, while to communicate only to people known in real life reduces the risk to be offended on the Internet.



" Parameter Estimation and Inference with Spatial Lags and Cointegration", with Jan Mutl


Econometric Reviews, 2019, Vol. 38}(6), pp.597-635,
https://doi.org/10.1080/07474938.2017.1382803


Abstract

We study dynamic panel data models where the long run outcome for a particular cross-section is affected by a weighted average of the outcomes in the other cross-sections. We show that imposing such a structure implies several cointegrating relationships that are nonlinear in the coefficients to be estimated. Assuming that the weights are exogenously given, we extend the dynamic ordinary least squares methodology and provide a dynamic two-stage least squares estimator. We derive the large sample properties of our proposed estimator and investigate its small sample distribution in a simulation study. Then our methodology is applied to US financial market data, which consist of credit default swap spreads, firm specific and industry data. A "closeness" measure for firms is based on input-output matrices. Our estimates show that this particular form of spatial correlation of credit default spreads is substantial and highly significant

 



"A new strategy for Robbins’ problem of optimal stopping", with Martin Meier

 

Journal of Applied Probability, 2017, Vol. 54, pp. 331 - 336.
https://doi.org/10.1017/jpr.2016.103


Abstract

In this article we study the expected rank problem under full information. Our approach uses the planar Poisson approach from Gnedin (2007) to derive the expected rank of a stopping rule that is one of the simplest non-trivial examples combining rank dependent rules with threshold rules.  This rule attains an expected rank lower than the best upper bounds obtained in the literature so far, in particular we obtain an expected rank of 2.32614..



"Bayesian Learning, Shutdown and Convergence"

 

Mathematical Social Sciences, 2015, 75,  pp. 27 - 43.
https://doi.org/10.1016/j.mathsocsci.2015.01.004


Abstract

This article investigates a partial equilibrium production model with dynamic information aggregation. Firms use observed prices to estimate the unknown model parameter by applying Bayesian learning.  In the baseline setting, the demand structure is linear and the noise term is Gaussian. Then, prices and quantities are supported by the real line and convergence of the limited information to rational expectations quantities is obtained.  Since a production economy is considered, the economic constraint of non-negative quantities is imposed.  This non-negativity constraint and the assumption that signals about demand are only received in periods where production takes place destroy the ``optimistic'' convergence result observed in the baseline model. With this constraint firms learning an unknown demand intercept parameter exit with strictly positive probability, even when the true value of this parameter would induce production in the full information setting. In a second step, the linear demand structure is replaced by piece-wise linear demand, such that prices become non-negative. Also in this stetting the convergence result of the baseline model does not hold. 




"Weather and SAD Related Mood Effects on the Financial Market", with Manfred Frühwirth

The Quarterly Review of Economics and Finance
, 57, pp. 11-31.
https://doi.org/10.1016/j.qref.2015.02.003


Abstract

We investigate the relationship between weather or seasonal affective disorder and the financial market, using a wide variety of financial market data as well as several weather variables and a seasonal affective disorder proxy. We distinguish between a model with a direct effect of the weather and seasonal affective disorder on the financial market and one with an indirect effect via a latent variable mood. Whereas only the latter model is justified by psychological literature, the former model is often used as an approximation. One major innovation of this paper is a consistent econometric implementation of the indirect effects model. We demonstrate that the approximation by direct effects yields inconsistent estimates.  Our study supports some weather related, but no seasonal affective disorder related effects on the financial market. We show that, instead of focusing on single market segments, an analysis of various financial market segments is required. We also show that the analysis of individual stock returns or bond spreads reveals additional information, compared to the analysis of aggregate stock or bond indexes.


Download Supplementary Materials