Evaluating the performance of merger simulation using different demand systems

Competition and Market Regulation master project by Leandro Benítez and Ádám Torda ’19

Photo credit: Diego3336 on Flickr

Evaluating the performance of merger simulation using different demand systems: Evidence from the Argentinian beer market

Editor’s note: This post is part of a series showcasing Barcelona GSE master projects. The project is a required component of all Master’s programs at the Barcelona GSE.

Abstract

This research arises in a context of strong debate on the effectiveness of merger control and how competition authorities assess the potential anticompetitive effects of mergers. In order to contribute to the discussion, we apply merger simulation –the most sophisticated and often used tool to assess unilateral effects– to predict the post-merger prices of the AB InBev / SAB-Miller merger in Argentina.

The basic idea of merger simulation is to simulate post-merger equilibrium from estimated structural parameters of the demand and supply equations. Assuming that firms compete a la Bertrand, we use different discrete choice demand systems –Logit, Nested Logit and Random Coefficients Logit models– in order to test how sensible the predictions are to changes in demand specification. Then, to get a measure of the precision of the method we compare these predictions with actual post-merger prices.

Finally, to conclude, we point out the importance of post-merger evaluation of merger simulation methods applied in complex cases, as well as the advantages and limitations of using these type of demand models.

Conclusion

Merger simulations yield mixed conclusions on the use of different demand models. The Logit model is ex-ante considered inappropriate because of its restrictive pattern of substitution, however it performed better than expected. Its predictions on average were close to the predictions of the Random Coefficients Logit model, which should yield the most realistic and precise estimates. Conversely, the Nested Logit model largely overestimated the post-merger prices. However, the poor performance is mainly motivated by the nests configuration: the swap of brands generates almost two close to monopoly positions in the standard and low-end segment for AB InBev and CCU, respectively. This issue, added to the high correlation of preferences for products in the same nest, generates enhanced price effects.

table_1_estimation_results

Regarding the substitution patterns, the Logit, Nested Logit and Random Coefficients Logit models yielded different results. The own-price elasticities are similar for the Logit and Nested Logit model, however for the Random Coefficients Logit model they are more almost tripled. This is likely driven by the estimated larger price coefficient as well as the standard deviations of the product characteristics. As expected, by construction the Random Coefficients Logit model yielded the most realistic cross-price elasticities.

table_2_elasticities

Our question on how does the different discrete choice demand models affects merger simulation –and, by extension, their policy implications– is hard to be answered. For the AB InBev / SAB-Miller merger the Logit and Random Coefficients Logit model predict almost zero changes in prices. Conversely, according to the Nested Logit, both scenarios were equally harmful to consumers in terms of their unilateral effects. However, as mentioned above, given the particular post-merger nests configuration, evaluating this model solely by the precision of its predictions might be misleading. We cannot discard to have better predictions under different conditions.

table_3_evaluation

As a concluding remark, we must acknowledge the virtues and limitations of merger simulation. Merger simulation is a useful tool for competition policy as it gives us the possibility to analyze different types of hypothetical scenarios –like approving the merger, or imposing conditions or directly blocking the operation–. However, we must take into account that it is still a static analysis framework. By focusing only on the current pre-merger market information, merger simulation does not consider dynamic factors such as product repositioning, entry and exit, or other external shocks.

Authors: Leandro Benítez and Ádám Torda

About the Barcelona GSE Master’s Program in Competition and Market Regulation

Forecasting Currency Crises

Macroeconomic master project by Ivana Ganeva and Rana Mohie ’19

Editor’s note: This post is part of a series showcasing Barcelona GSE master projects. The project is a required component of all Master’s programs at the Barcelona GSE.

Introduction

The question of whether a currency crisis can be predicted beforehand has been discussed in the literature for decades. Economists and econometricians have been trying to develop different prediction models that can work as an Early Warning System (EWS) for a currency crisis. The significance of such systems is that they provide policy makers with a valuable tool to aid them in tackling economic issues or speculation pressure, and in taking decisions that would prevent that from turning into a crisis. This topic is especially relevant to Emerging Markets Economies due to the presence of a greater number of fluctuations in their exchange rate translating to a bigger currency crisis risk.

In this paper, we propose an Early Warning System for predicting currency crises that is based on an Artificial Neural Networks (ANN) algorithm. The performance of this EWS is then evaluated both in-sample and out-of-sample using a data set of 17 developed and developing countries over the period of 1980-2019. The performance of this Neural-Network-based EWS is then compared to two other models that are widely used in the literature. The first one is the Probit model dependent variable which is considered the standard model in predicting currency crises, and is based on Berg and Patillo, 1999. The second model under consideration is a regime switching prediction model based on that proposed by Abiad, 2006.

Artificial Neural Networks

Artificial Neural Networks (ANN-s) is a Machine Learning technique which drives its inspiration from biological nervous systems and the (human) brain structure. With recent advancement in the computing technologies, computer scientists were able to mimic the brain functionality using artificial algorithms. This has motivated researchers to use the same brain functionality to design algorithms that can solve complex and non-linear problems. As a result, ANN-s have become a source of inspiration for a large number of techniques across a vast variety of fields. The main financial areas where ANN-s are utilized include credit authorisation and screening, financial and economic forecasting, fixed income investments, and prediction of default and bankruptcy and credit card manipulations (Öztemel, 2003).

Main Contributions

1. Machine Learning Techniques:

(a) Using an Artificial Neural Network predictive model based on the multi-layered feed forward neural network (MLFN), also known as the “Back-propagation Network” which is one of the most widely used architectures in the financial series neural network literature (Aydin and Savdar 2015). To the best of our knowledge, this is the first study that used a purely neural network model in forecasting currency crises.

(b) Improving the forecast performance of the Neural Network model by allowing the model to be trained (learn) from the data of other countries in the cluster; i.e countries with similar traits and nominal exchange rate depreciation properties. The idea behind this model extension is mainly adopted from the “Transferred Learning” technique that is used in image recognition applications.

2. The Data Set: Comparing models across a large data set of 17 countries in 5 continents, and including both developing and developed economies.

3. Crisis Definition: Adding an extra step to the Early Warning System design by clustering the set of countries into 6 clusters based on their economy’s traits, and the behavior of their nominal exchange rate depreciation fluctuations. This allows for having a crisis definition that is uniquely based on each set of countries properties – we call it the ’middle-ground’ definition. Moreover, this allowed to test for the potential of improving the forecasting performance of the neural network by training the model on data sets of other countries within the same cluster. 4. Reproducible Research: Downloading and Cleaning Data has been automated, so that the results can be easily updated or extended.

Conclusions

We compare between models based on two main measures. The Good Signals measure captures the percentage of currency crises predicted out of the total crises that actually occurred in the data set. The second measure used for comparing across models is the False Alarms measure. That is, the percentage of false signals that the EWS gives out of the total number of crises it predicts. In other words, that is the percentage of times when the EWS predicts a crisis that never happens.

The tables presented below show our findings and how the models perform against each other on our data set of 17 countries. We also provide the relevant findings from literature as a benchmark for our research.

The results in Table 1 show that Berg & Patillo’s clustering of all countries together generally works worse than our way of clustering data. Therefore, we can confirm that the choice of a ’middle-ground’ crisis definition has indeed helped us preserve any potential important country- or cluster-specific traits. In brief, we get comparable results to the ones found in the literature when using conventional methods, as highlighted by the table to follow.

After introducing the ANN model and its extension, we observe their Out-of-Sample models performance and obtain some of the key results to our research.

Summary of the key results

  • The proposed Artificial Neural Network model crisis predictability is shown to be comparable to the standard currency crisis forecasting model across both measures of Good Signals and less False Alarms. However, the modified Neural Network model on the special clustering data set has shown superior performance to the standard forecasting model.
  • The performance of the Artificial Neural Network model observed a tangible improvement when introducing our method of clustering the data. That is, data from similar countries as part of the training set of the network could indeed serve as an advantage rather than a distortion. To the contrary, using the standard Probit model with the panels of clustered data resulted in lower performance as compared to the respective country-by-country measures.

Authors: Ivana Ganeva and Rana Mohie

About the Barcelona GSE Master’s Program in Macroeconomic Policy and Financial Markets

On the evolution of altruistic and cooperative behaviour due to schooling system in Spain

Economics master project by Shaily Bahuguna, Diego Loras Gimeno, Davina Heer, Manuel E. Lago, and Chiara Toietta ’19

Editor’s note: This post is part of a series showcasing Barcelona GSE master projects. The project is a required component of all Master’s programs at the Barcelona GSE.

Abstract

This paper aims to find a pattern in the evolution of altruistic and cooperative behaviour whilst distinguishing across different types of schools in Spain. In specific, we design a controlled laboratory experiment by running the standard dictator game and a public goods game in a public and private (“concertada”) high school. Using a sample of 180 students, we compare 12 and 16 year old children to distinguish the evolutionary pattern and test if there is a significant change by the type of schooling system. Alongside, we control for variants such as parental wealth status, religious views and ethical opinions. Interestingly, evidence from our data highlights that altruism levels rise throughout public school education whilst it falls in private schools. On the contrary, cooperation levels are relatively stable in public schools but rise in private schools. The results from this paper can be exploited to understand how education may influence selfish and individualistic behaviour in our society. 

Key results

Diff-in-Diff (Altruism (L) & Cooperation (R))

Our results show that at the initial stage, i.e. for the first year students, the level of altruism is higher in public schools and this prevails throughout the students’ education in a public school. On the other hand, we observe an opposite trend for students attending private school, as over the four years of education, the average level of altruism declines. In regards to cooperation, we find some surprising results. Although students attending a public school initially show higher levels of cooperation than private schools, over the course of their education, this gap is not only reduced but it is also surpassed by the private school. Our results are in line with previous research which state that females are more likely to donate and cooperate than males but contradict the popular view in literature that income has a positive correlation with both dependent variables.

Authors: Shaily Bahuguna, Diego Loras Gimeno, Davina Heer, Manuel E. Lago, and Chiara Toietta

About the Barcelona GSE Master’s Program in Economics

Women’s Status in Rural Bangladesh: Exploitation and Empowerment

Economics of Public Policy master project by Agrima Sahore, Ah Young Jang, and Marjorie Pang ’19

Editor’s note: This post is part of a series showcasing Barcelona GSE master projects. The project is a required component of all Master’s programs at the Barcelona GSE.

Abstract

Using household survey data from rural Bangladesh, we explore determinants of domestic violence. We propose two hypotheses: first, women suffer more domestic abuse as a result of marrying young; and second, women who are empowered suffer less gender-based violence. We isolate the causal effect of marriage timing using age at first menstruation and extreme weather as instruments; and the effect of empowerment using the number of types of informal credit sources as instrument. We find robust evidence contrary to our hypotheses. Our findings highlight that mere empowerment or increasing age at first marriage are insufficient mediums to combat gender-based violence and can in fact be counterproductive to reducing domestic violence against women, if the socio-economic context is not carefully considered.

Conclusion

Interestingly, we find a positive relationship between age at first marriage and domestic violence; and empowerment and domestic violence. This highlights the complexity of the nature of domestic violence against women in a highly conservative setting like rural Bangladesh.

Violence against women continues to be a social and economic problem Bangladesh struggles with. Although the government had aimed to eliminate gender based violence in the country by 2015, their efforts have not achieved the desired results. However, if the empowerment of women (an improvement in their economic and social status) and violence against them follows an inverted U-shaped curve, it is possible that Bangladesh is still adjusting to egalitarian gender norms and expectations and is stationed somewhere on the positive slope of the curve, wherein increase in empowerment initially would increase violence against women, before reducing it.

In order to design successful policies to combat violence against women, our study highlights the importance of understanding traditional cultural norms – especially prevailing gender norms – economic conditions, and how the interplay of various socio-economic factors contribute to domestic violence against women. Ultimately, actions and practices aimed at improving women’s condition in societies should work towards confronting existing circumstances and environments that underlie women’s risk of experiencing domestic violence.

Authors: Agrima Sahore, Ah Young Jang , and Marjorie Pang

About the Barcelona GSE Master’s Program in Economics of Public Policy

An Empirical Framework: Financial Globalization and Threshold Effects

Economics master project by Eimear Flynn, Florencia Saravia, Josefina Cenzon, Nimisha Gupta, and Selena Tezel ’19

Editor’s note: This post is part of a series showcasing Barcelona GSE master projects. The project is a required component of all Master’s programs at the Barcelona GSE.

Abstract

Is financial globalization beneficial to economies at all levels of development? Or are there certain “threshold” levels of financial, institutional and economic development a country must first attain in order to realize the growth benefits of globalization? Kose, Prasad and Taylor (2009) develop a unified empirical framework to answer this question. The debate on the literature is ongoing. Yet few studies have explored these questions in a post-crisis context. In this paper, we replicate and extend their work, paying close attention to the period 2005-2014. Our analysis yields three key results. First, the financial depth threshold above which countries can benefit from financial globalization increases from 66% to 81% when we consider the extended period. Second, the proportion of countries with depth levels above this threshold declines over time. Finally, the coefficients are smaller in absolute value over the period 1975-2014. Taken together, these results imply a breakdown in the relationship between financial depth, openness and growth since the Great Recession. Financial deepening on its own can no longer ensure positive growth effects of financial integration.

Conclusion

In the paper, we examine two periods 1975-2004 and 1975-2014 and test for threshold effects in three variables, financial depth, institutional quality and trade openness. This paper is unique in its inclusion of the years immediately before and following the Great Recession. Following a surge in international financial integration between 2005 and 2007, financial openness plummeted with the onset of the crisis in 2007. This effect was most pronounced in advanced economies. As economic growth rates declined, countries turned their backs on financial globalization. Financial flows have since rebounded, albeit not to their pre-crisis levels. The effect of this volatility on the relationship between financial openness and economic growth however is not well understood.

thresh_FD_Graph
Overall Financial Openness Coefficient and Financial Depth in 1975-2014 vs 1975-2004

We analyze changes in the financial depth threshold over time as well as changes in the proportion of countries with depth levels above this threshold over time. We present three key findings. We first document an increase in the threshold level of financial depth from 66% to 81% when we extend the period to 2014. It follows that the proportion of countries with levels of depth above this threshold decreases over time. Our estimate of 66% for the period 1975-2004 is remarkably close to that of Kose et al. Secondly, the coefficient estimates are smaller and less significant which points to a breakdown in the relationship between financial openness and growth in the post-crisis period. Finally, we identify significant threshold effects of institutional quality.

Together these results point to a weaker relationship between financial openness and growth in the post-crisis period. Our estimates suggest that once a country reaches the threshold level of financial depth, further improvements in depth stop being important quite rapidly. It is now more difficult for countries to attain the benefits of financial integration, not just because the threshold of financial depth is higher but because financial depth alone may no longer be sufficient to ensure growth. The trade-off that further financial deepening can generate between higher growth and a higher risk of crisis needs to be addressed. The Great Recession was a reminder that financial depth and financial stability need not go hand in hand. The risks of financial deepening are more evident than before. Focusing only on the long run growth view overlooks this trade-off. In order to conduct policy relevant research, a new approach that realistically accounts for both the growth and crisis effects of financial deepening is required.

Authors: Eimear Flynn, Florencia Saravia, Josefina Cenzon, Nimisha Gupta, and Selena Tezel

About the Barcelona GSE Master’s Program in Economics

Bank Interconnectedness and Monetary Policy Transmission: Evidence from the Euro Area

ITFD master project by Sofia Alvarez, Michael Barczay, Guadalupe Galambos, Ina Sandler, and Rasmus Herløw Schmidt ’19

Editor’s note: This post is part of a series showcasing Barcelona GSE master projects. The project is a required component of all Master’s programs at the Barcelona GSE.

Abstract

Whereas monetary policy in the euro area is conducted by one single authority, the European Central Bank (ECB), the real effects of these decisions provoke different reactions among the set of targeted countries. One common explanation for this finding is structural heterogeneity across the members of the eurozone. The paper assesses how commercial bank interconnectedness – as a specific source of structural heterogeneity – affects the propagation of common monetary policy shocks across the euro countries between 2003 and 2018. While recent research has found that bank interconnectedness can counteract the contraction in loans resulting from a monetary policy tightening in the US, evidence on the effects in the euro area is scarce. The paper tests this hypothesis empirically by constructing a panel data set for the original euro area countries, creating a measure for country-specific bank interconnectedness, and by identifying an exogenous monetary policy shocks based on high-frequency data. Employing local projections we find evidence that – in accordance with the theoretical model we discuss – a contractionary monetary policy surprise leads to a reduction in the supply of corporate and household loans in countries with a low degree of interconnectedness. Conversely, when the banking sector is highly interconnected, the impact of monetary policy is reduced or even counteracted. This implies that a common monetary policy shock can have heterogeneous effects among countries of the euro area, depending on the degree of interconnectedness of their banking industries. These results are robust to the inclusion of a wide set of controls and alternative shock specifications.

Conclusions and Key Results

Standard theory, assuming that the balance sheet transmission channel is at play, suggests that an increase in the interest rate reduces the value of pledgeable assets held by firms and households. This reduction in the borrowers’ creditworthiness consequently induces banks to constrain their amount of lending. A higher degree of interconnectedness of the banking sector, however, makes banks less sensitive to changes in the value of the collateral posted by borrowers since loan portfolios can be traded among banks with different risk exposures. Thus, when banks are highly interconnected, the reduction of the loans supply after contractionary monetary policy is expected to be smaller and the effects of monetary policy are likely to be less pronounced.

The paper tests this hypothesis for the eurozone using local projections and a panel dataset of 12 euro-area countries in the period between 2003 and 2018. The eurozone constitutes an excellent subject to study: It exhibits significant variation of bank interconnectedness – both across time and countries – which makes the hypothesis of bank interconnectedness as a determinant of heterogeneous reactions to monetary policy particularly relevant.

The key findings of the paper are that interconnectedness is in fact an important driver of heterogeneity in the transmission of monetary policy in the eurozone. More precisely, the analysis shows that when bank interconnectedness is low, contractionary monetary policy leads to a reduction in lending. In countries with highly interconnected banking sectors, however, the paper documents that the impact of monetary policy on loans may be offset. Even more so, our findings suggest that the effect of monetary policy may even be reversed at certain points in time after a monetary policy shock (see Figure 1). More generally, these effects seem to persist for approximately 15 months and hold for both household and corporate loans. A potential conjecture for the observed increase in loans, when bank interconnectedness is high, could be that contractionary shocks induce lending from highly interconnected core countries to countries with low bank interconnectedness in the European periphery.

Additionally, higher interest rates may induce banks to increase the loan supply due to higher potential returns in a context of efficient risk-sharing. The analysis also suggests that cross-country variation of bank interconnectedness, in particular, plays an important role in explaining heterogeneous responses to monetary policy. Muting the cross-country variation, by contrast, delivers effects that are  still in line with the theory but which turn out to be statistically insignificant. In other words, accounting for cross-country heterogeneity is essential. Moreover, employing alternative specifications, the results are found to be robust to using alternative shock measures, including additional controls, and excluding outliers.

Figure_1
Figure 1: Impulse response of total loans to a one basis point increase in the monetary policy shock variable (Red: Low bank interconnectedness, blue: high bank interconnectedness)

Using our estimates to predict the country specific responses to a monetary policy shock, we show that countries are expected to react very heterogeneously. While Greece, for instance, is predicted to experience a contraction of loans after an interest rate hike at its 10th, 50th, and 75th percentile of bank interconnectedness, the effect of  monetary policy is almost completely offset in Germany or Austria (see Figure 2).

Figure 2
Figure 2: Implied heterogeneous responses after 6 months to a common monetary policy shock of 1 basis point by each country’s degrees of interconnectedness

In sum, the analysis contributes to the understanding of how monetary policy is transmitted across countries in the euro area. The paper shows how heterogeneous responses can be explained by the variation in countries’ individual levels of bank interconnectedness. Furthermore, it provides a potential explanation for why recent research has found rather modest responses of the loan supply to monetary policy. Such observations appear inevitable given that many countries display levels of interconnectedness at which the reaction of loans to monetary policy is predicted to be only small or even nonexistent. Only by incorporating interbank lending into the analysis, sizable effects of monetary policy become apparent again. 

Authors: Sofia Alvarez, Michael Barczay, Guadalupe Galambos, Ina Sandler, and Rasmus Herløw Schmidt

About the Barcelona GSE Master’s Program in International Trade, Finance, and Development

Estimating Time-Varying Network Effects with Application to Portfolio Allocation

Finance master project by Daniel A. Landau and Gabriel L. Ramos ’19

Editor’s note: This post is part of a series showcasing Barcelona GSE master projects. The project is a required component of all Master’s programs at the Barcelona GSE.

Abstract

In this paper, we characterize a variety of international financial markets as partially correlated networks of stock returns via the implementation of the joint sparse regression estimation techniques of Peng et al. (2009). We explore a number of mean-variance portfolios, with the aim of enhancing out-of-sample portfolio performance by uncovering the hidden network dynamics of optimal portfolio allocation. We find that Markowitz portfolios generally dissuade the inclusion of central stocks in the network, yet the interaction of a stock’s individual and systemic performance is more complex. This motivates us to explore the time-varying correlation of these topological features, which we find are highly market dependent. Building on the work of Peralta & Zareei (2016), we implement a number of investment strategies aimed at simplifying the portfolio selection process by allocating wealth to a targeted subset of stocks, contingent on the time-varying network dynamics. We find that applying mean-variance allocation to a restricted sample of stocks with daily portfolio re-balancing can statistically significantly enhance out-of-sample portfolio performance in comparison to a market benchmark. We also find evidence that such portfolios are more resilient during periods of major macroeconomic instability, with the results applicable to both developed and emerging markets.

Conclusion and Future Research

In our work, we represent 4 international exchanges as individual networks of partially correlated stock returns. To do so, we build a Graph, comprised of a set of Vertices and Edges, via the implementation of the joint sparse regression estimation techniques of Peng et. al (2009). This approach allows us to uncover some of the hidden topological features of a series of Markowitz tangency portfolios. We generally find that investing according to MPT dissuades the inclusion of highly central stocks in an optimally designed portfolio, hence keeping portfolio variances under control. We find that this result is market-dependent and more prevalent for certain countries than for others. From this cross-sectional network analysis, we learn that the interaction between a stock’s individual performance (Sharpe ratio) and systemic performance (eigenvector centrality) can be complex. This motivates us to explore the time-varying correlation ρ between Sharpe ratio and eigencentrality.

Optimal_Weights_for_Tangency_Portfolio_Strategy
Optimal Weights for Tangency Portfolio Strategy.

Overall, we show that in considering the time-varying nature of partially correlated networks, we can enhance out-of-sample performance by simplifying the portfolio selection process and investing in a targeted subset of stocks. We also find that our work proposes a number of future research questions. Although we implement short-sale constraints, it would also be wise to introduce limits on the amount of wealth that can go into purchasing stocks, as this would help to avoid large portfolio variances. Furthermore, our work paves the way for future research into the ability of ρ-dependent investment strategies to enhance portfolio performance in times of macroeconomic distress and major financial crises.

Authors: Daniel A. Landau and Gabriel L. Ramos

About the Barcelona GSE Master’s Program in Finance

Certain uncertainty? The response of Venezuelan banks to a political dilemma

Economics of Public Policy master project by Mary Armijos and Guillem Cuberta ’19

Editor’s note: This post is part of a series showcasing Barcelona GSE master projects. The project is a required component of all Master’s programs at the Barcelona GSE.

Introduction

Our work focuses on the analysis of the Venezuelan banks, which have become more vulnerable. The leading causes of this vulnerability have been oil price shocks, political instability, pro-cyclical monetary conditions (i.e., interest rates), low level of financial intermediation, changing structure (e.g., consolidation, closure or nationalization), non-traditional bank transactions, high exposure to the public sector, and government intervention in their operations.  (Blavy R., 2014) Hence, studying the response of banks to policy uncertainty becomes relevant. However, even more critical, in Venezuela’s context, would be to ask: do banks respond differently to uncertainty if they are politically connected?

To address this question, we construct two indices: the policy uncertainty index and political connection index. Our uncertainty index is in a monthly basis and bank-invariant; it is developed following the work previously done by Baker, Bloom, and David, on the Economic Policy Uncertainty Index (EPU Index) and by Ahir, Bloom, and Furceri, on the World Uncertainty Index (an extension of the EPU).  Similar to Xu and Zhou (2008), for the case of the political connection index, we define the dummy variable connected bank equal one based on whether it is a state-owned bank. Alternatively, in case it is a private one if at least one of the board members has any connection with someone from the government. We adapt these criteria to the information there is available from Venezuela. We also look if one of the members is part of the ‘bolibourgeoisie’ ( a combination of the words Bolivarian and bourgeoisie, a term used to classify the businessmen and public officials linked to the government). For our dependent and micro-control variables (i.e., bank characteristics), we use monthly data from 2006 to 2018 obtained from SUDEBAN (Venezuela’s Superintendence of Banks). Moreover, for the macro-control variables, we obtain monthly data from the International Financial Statistics of the International Monetary Fund (IMF).  The variables we choose for our specifications are based on the works done by Vera et al. (2019) and Bordo et al. (2016).

Our central hypothesis is that when there is high policy uncertainty, political connections may allow connected banks to smooth the effects of uncertainty. We believe that connected banks might respond differently to uncertainty because they have privileged information or preferential treatment from the government, which grants them a competitive advantage over non-connected banks. To test the response of banks, we look at their behavior respect to credit supply and provisions, and we also investigate banks’ profitability in periods of uncertainty through the ROA. Our identification strategy to investigate the causal effect of policy uncertainty and political connection considers the fact that the uncertainty index is a high-frequency time series, which makes it be almost an exogenous variable. Also, the political connection index is not endogenous as it does not vary over time.

In addition to that, we control for bank-invariant and time-specific factors that affect both the right-hand side and left-hand-side variables by adding bank and time fixed effects. Furthermore, to separate the impact of our primary explanatory variable (interaction) from other confounding factors, we control for a block of bank-specific covariates. We also consider different specifications with and without lags of these controls to mitigate the potential reverse-causality concern. Even though we know that all of these adjustments might not entirely correct for omitted variable bias, we consider it adjusts well enough to investigate this relationship. We consider that one of the significant sources of potential bias comes from monthly macro changes (i.e., exogenous shocks like oil prices or U.S. sanctions, and government decisions) that are accounted by including time fixed-effects.

Figure 1: Monthly Economic Policy Uncertainty Index for Venezuela (EPUV)

Results

In our main results, we find that politically connected banks acquire more risks when there is higher uncertainty as an increase of 10 percent in our uncertainty measure leads them to give on average, 0.0262 percent more credits than non-politically connected banks. This result corroborates similar results from the literature that establishes a positive value from being politically connected (Kostovetsky 2015).  Also, we observe that an increase in the uncertainty index induces politically connected banks to hold more loss provisions in their portfolio than non-politically connected banks. A 10 percent increase in our uncertainty index prompts politically connected banks to hold 0.0192 percent more loss provisions than non-politically connected banks. Lastly, the effect of economic policy uncertainty for politically connected banks on ROA has a positive sign. A 10 percent increase in uncertainty increases the average returns on assets of politically connected banks by almost 11 percent compared to non-politically connected banks.

Summing up, connected banks can give more credit to the public in periods of higher uncertainty, at the same time that they hold more loss provisions. The first result is consistent with the ones shown by Cheng et al. (2017), where they find that banks supply much more credit when there is high uncertainty. However, our result of provisions does not coincide with theirs. Contrarily, they find that under higher uncertainty, connected banks reserve lower provisions than unconnected banks. In the case of the profitability analysis, connected banks have lower profits when there is high uncertainty. These results go along with the ones found by Dicko (2016).

We consider that this study presents new relevant findings regarding the literature of political connection and policy uncertainty, and for the Venezuelan economy overall. The political connection matters in periods of high uncertainty but until one point. From our results, we find that politically connected banks seem pro-risk as they give more credit when there is more policy uncertainty. However, on another level, it appears that they do receive privilege information form the government (bad news about the future) that makes them risk-averse at the same time as they also reserve more provisions. Additionally, the result of the relationship between uncertainty and profitability indicators, like ROA, indicates that being politically connected might not be extremely helpful to banks if they only benefit from having more information and do not receive any tangible benefit from the government. For further studies, it would be interesting to analyze how economic agents respond to policy uncertainty depending on the type of benefit they receive from being politically connected to some institutions.

About the Barcelona GSE Master’s Program in Economics of Public Policy

The Impact of Non-contributory Pensions. A Case Study for Costa Rica

Economics of Public Policy master project by Hazel Elizondo, Sandra Flores and Alicia de Quinto ’18

SUPEN

Editor’s note: This post is part of a series showcasing Barcelona GSE master projects by students in the Class of 2018. The project is a required component of every master program.


Authors:

Hazel Elizondo, Sandra Flores and Alicia de Quinto

Master’s Program:

Economics of Public Policy

Paper Abstract:

Even when only 20% of the elder population in the world receives pension coverage (Pallares-Miralles, Romero and Whitehouse, 2012) which in addition is not always adequate according to ILO, non-contributory pensions are present only in a handful of developing countries. Moreover, the elderly population is currently growing as individuals tend to live longer, what further evidences the imperative need to apply this type of programs. Costa Rica implemented a non-contributory pension policy in 1975 to ensure the livelihood of those in economic need that were not able to save provisionary funds to confront the old age risks, so that elders aged above 65 living in extreme poverty are eligible for coverage. Additionally, Costa Rica adopted a 186% increase on the pension amount in 2007 in order to mitigate poverty. This study aims to provide further empirical evidence of the indirect effects of the non-contributory pensions in Latin America, through a study case for Costa Rica that explores the impact of this pension on employment and schooling, household composition, and changes in well-being for the period from 2001 to 2009.

Figure: Poverty rate for different age ranges in Costa Rica

The methodology applied includes a first difference-in-differences specification (DD) as a general model, which compares the group of receivers before and after 2007 with a control group aged above 65 years old. Secondly, we exploit the discontinuity on the treatment assignment regarding the age of the oldest household member to define a Regression Discontinuity Fuzzy Design (RD). This local analysis only identifies the effects of receiving the pension, so that we move towards a third Difference in Discontinuity Design (diff-in-disc) that combines the previous models, quantifying the impacts of the pension increase as well. The RD and diff-in-disc settings include an alternative sample where the treated are households with a member between 65 and 69 years, while the counterpart is aged between 61 and 64.

Conclusions and key results:

Our results show a generally positive picture of the Costa Rican non-contributory pension, if we consider that the policy was designed to provide an allowance to elder that never contributed to the formal system, allowing them to retire at age 65. However, conditional income transfers sometimes involve unintended consequences that characterize the policy as defective.

Table: Main Results

In the case of the DD sample, where the family structures are characterized by households with senior members and households where the recipient is father or mother of the household head, the results show major spillover effects on the remaining members, especially in terms of labor- related reactions. Indeed, the estimations show that those households that benefit from the non- contributory pension reduce significantly by 0.179 the number of individuals in the labor force, compared to non-beneficiaries. Individuals in the treated households work 1.747 hours less than their counterparts and receive a labor income 61.9 USD lower than those households that do not receive the pension. Given that the Costa Rican non-contributory pension policy requires leaving the labor market as a necessary condition for receiving the grant, we might relate the reduction in labor participation to perverse incentives, as the remaining household members might take advantage of this transfer to change their time allocation preferences between work and leisure.

Nonetheless, the results obtained in the RD and diff-in-diff models rule out our preliminary interpretation. Both estimates reveal no significant reactions at the household level for any of the outcomes analyzed, what means that households do not change their employment-related decisions in the short-run, even when the recipient must leave the labor market. In this case, the households with senior members predominate over other type of family structures, hence we would have expected a significant decreasing effect for labor force participation. Probably this is because unemployment and job instability hit the most vulnerable population groups, so that individuals with uncertain job prospects see in the non-contributory pension an opportunity to receive a steady income. Moreover, we do not find evidence neither for the incentive for other young members of the family to move in with the elderly participant, nor for the recipient to move out and live on her own.

Download the full paper [pdf]


More about the Economics of Public Policy Program at the Barcelona Graduate School of Economics

Investigation of Sentiment Importance on Intraday Stock Returns

Data Science master project by Michele Costa, Alessandro De Sanctis, Laurits Marschall and S. Hamed Mirsadeghi ’18

Investigation of Sentiment Importance on Intraday Stock Returns

Editor’s note: This post is part of a series showcasing Barcelona GSE master projects by students in the Class of 2018. The project is a required component of every master program.


Authors:

Michele CostaAlessandro De SanctisLaurits Marschall and S. Hamed Mirsadeghi

Master’s Program:

Data Science

Paper Abstract:

The main goal of our Master Project is to predict intraday stock market movements using two different kinds of input features: financial indicators and sentiments from news and tweets. While the former are part of the common technical analysis of financial econometric models, the extracted sentiment of news articles and tweets from Twitter are also proven to correlate with stock markets movements. Our paper aims at contributing to the existing academic and professional knowledge in two main directions. First, we evaluate three different approaches to extract the sentiment from both social and mass media based on its forecasting power. Second, we deploy a battery of engineered features based on the sentiment, together with the financial indicators, in a machine learning model for a fine-grained minute-level forecasting exercise. In the end, two different classes of models are fitted to test the forecasting power of the combined input features. We estimated a classical ARIMA-model, and an XGBoost-model as machine learning algorithm. We collected data on the companies Apple, JPMorgan Chase, Exxon Mobil, and Boeing.

Figure: Exxon Mobil
The picture shows how sentiments towards Exxon Mobil moved over time. The two lines refers to two different methodologies: Loughran-McDonald is based on a financial dictionary while SentiStrength was trained on social media such as MySpace.


More about the Data Science Program at the Barcelona Graduate School of Economics