Paris December 2019

We exploit one of the largest reform of dividend taxation in France in 2012 to estimate the effect of an increase in dividend tax rate on corporate policies, using administrative data covering the universe of firms and employees. In a difference-in-difference setting, we find that exposed firms swiftly cut dividend payments, both at the extensive and intensive margin, with an implied elasticity of -0.6. We find that treated firms build up considerable amount of cash reserves after reducing their dividend payment but do not change their investment policies. We report a precisely estimated null effect of dividend taxes on investment rate, with an implied elasticity around 0.04.

This paper examines how firms adapt to climate-change risks resulting from their supply-chain networks. Combining a large sample of global supplier-customer relationships with granular data on local temperatures and flooding incidents, we first document that the occurrence of climate shocks at affected supplier firms has both a large direct and indirect negative effect on earnings and revenues of suppliers and their customers. Second, we show that customers are 10% to 20% more likely to terminate existing supplier-relationships when realized climate shocks at the supplier firms exceed ex-ante expected climate shocks. Further, customers subsequently switch to suppliers with lower heatwave and flooding exposure. Our results indicate that climate change affects the formation of global production networks.

Stock-option compensation is coherent with the principle of agency theory, by encouraging risk averse executives to take a more risk neu- tral stance. However, it is unclear whether the contract is actually more powerful than personal characteristics in shaping decision-making. We compare experimentally the risk aversion and prudence of 100 par- ticipants under a stock-option incentive contract and a classic equity granting one. We measure a number of personality variables, as well as cortisol and testosterone levels, which have been related to risk taking in previous studies. We rst underline that stock-options contract in- deed drives more risk neutral behaviour, both regarding variance and skewness. This eect is mainly driven by a shift in focus from losses to gains. Loss aversion and higher cortisol levels were linked to a prefer- ence for the safer equity contract. Second, we show that cortisol and testosterone levels are linked to preferences for variance and skewness, often in a quadratic fashion. Third, we show that this impact is actually stronger than the one of the contract, both in terms of predictive power and economic eect. Therefore, our study highlights the impact of per- sonality and even biological variables on risk taking, and underline the need for a behavioural approach of contract setting.

We document that both the dividend yield and earnings yield can predict the future inflation across advanced economies. The inflation predictability reinforces the return predictability and reduces the dividend growth predictability and we show both discount rates and cash flows play important roles in determining prices. We test three hypothesis related to future growth prospect, risk aversion and behavior bias to justify the positive correlation among inflation and dividend (earnings) yields. Evidence suggests that high expected inflation has tended to coincide with periods of lower real economic growth and higher discount rates which lead to the drop in today's prices. To rationalize the inflation predictability, we develop and estimate a long-run risk model featuring inflation non-neutrality. The estimated model can reproduce both the inflation predictability and the documented asset pricing facts.

Using a stochastic volatility affine term structure model, we explicitly consider the interrelation between yield curves and macro and volatility factors. We provide estimates of short rate expectations, term premium and convexity of nominal yields and for their real and inflation components for four different currency areas: US, Euro Area, UK, and Japan. We find that in all areas there are non-negligible convexity effects in correspondence with high volatility periods, and that term premium and convexity explain a significant proportion of the dynamics at the long end of the yield curve. Using panel regressions, we show that, overall, short rate expectations are procyclical while term premia exhibit a countercyclical behaviour and tend to increase with yield volatility. We also detect strong cross-country co-movements both in short rate expectations and term premia, with the degree of connectedness exhibiting significant time variation.

Small businesses are championed by politicians looking to win votes, but face significant credit constraints that limit their growth. We study how the competitiveness of congressional elections affects government subsidies to small business lending. To identify the causal impact of electoral competitiveness, we examine politically-motivated congressional redistricting ("gerrymandering") and exploit the discontinuity in post-redistricting electoral competitiveness between districts where redistricting party incumbents narrowly won and narrowly lost the pre-redistricting election. We find that districts with electorally vulnerable congressional representatives receive larger Small Business Administration (SBA) loan guarantees than districts with more entrenched representatives, and this leads to higher growth in local employment and wages in the short run. The gains in employment and wages disappear in the long run, and is accompanied by a decline in the number of business establishments. Overall, our results suggest that politically-motivated credit subsidies to small businesses can provide short-term economic benefits, but may distort local credit markets in the long run.

The short-selling of exchange-traded funds (ETFs) creates "phantom" ETF shares with cash flows rights but no associated voting rights. Both regular and phantom ETF shares trade at ETF market prices. However, while regular shares are backed by the underlying securities of the ETF and voted as directed by the sponsor, phantom shares are backed by collateral that is not voted. Introducing a novel measure of phantom shares both of the ETF and corresponding underlying securities, we find that increases in phantom shares are associated with (i) decreases in number of proxy votes cast (for and against), (ii) increases in broker non-votes, and (iii) increases in the vote premium over the voting record date for important votes for the underlying stocks of the ETF. Consistent with poor governance, firms with the highest proportion of phantom shares underperform.

Using a simple cheap-talk game, we theoretically demonstrate that corporate social responsibility (CSR) helps mitigate the CEO-board information asymmetry, leading to more informed advising and monitoring by the board. By optimally engaging in CSR, the board can take advantage of stakeholder-induced information revelation and reduce its informational dependence on the CEO, which enables the shareholders to choose an ex ante higher level of board independence. For a sample of U.S. firms between 1999 and 2013, we find strong support for this strategic complementarity between board in-dependence and the information value of CSR. Our results highlight a novel rationale for CSR – the information motive.

We study how markets solve the standard but computationally hard problem of maximizing utility subject to a budget constraint in the face of indivisibilities. In a first experiment with complete markets, we test whether prices fully reveal the solution. We find that complete markets are fairly ineffective, and that quality decreases with instance complexity using a metric suggested in the theory of computation. Still, participants use information in prices and volume to improve their choices. In a second experiment, inspired by the theory of oracles in computer science, we show that a reduction to only one carefully chosen traded asset dramatically improves price quality and information dispersion. Our analysis provides a formalization of Hayek's ideas about the transmission of knowledge in societies through markets (Hayek, 1945). Ours is fundamentally different from the noisy rational expectations asset pricing model proposed in Grossman (1976, 1981), i.e., it does not appeal to mathematical statistics, but to the theory of computation.

I investigate how the 2011 and 2014 EU stress tests affect the risk-taking of European banks. I document a non-monotonic relationship between banks' risk-shifting resulting from regulatory arbitrage and the tightness of their capital constraint (i.e., the distance between their ex-ante capital ratio and the regulatory level): banks with capital ratios marginally above the regulatory level do more regulatory arbitrage than banks with a level of capital ratio significantly below or above the regulatory level. I also study the indirect effect of the tests on the financing costs of banks which are excluded from the tests: their financing costs on the corporate bond market increase with the level of negative information released in the country in which they are located.

Do competition and incentives offered to designated market makers (DMMs) improve market liquidity? Using data from NYSE Euronext Paris, we show that an exogenous increase in competition among DMMs leads to a significant decrease in quoted and effective spreads, mainly through a reduction in adverse selection costs. In contrast, changes in incentives, through small changes in rebates and requirements for DMMs, do not have any tangible effect on market liquidity. Our results are of relevance for designing optimal contracts between exchanges and DMMs and for regulatory market oversight.

We conduct a spectral analysis of the relation between fund flows and mispricing. Hedge funds and mutual
funds both behave as low-pass filters, deploying high-frequency flows toward low-frequency mispricing.
But hedge funds attenuate high-frequency flows more than do mutual funds, thus improving market
efficiency 2 to 7 times more slowly than mutual funds worsen efficiency. Time-series and cross-sectional
tests indicate that risk, limited access to capital, and implementation costs explain why hedge funds behave
as low-pass filters. We propose a model to rationalize these results, which highlight the frequencydependent
effects of (especially arbitrage) capital on market efficiency.

Why do some firms enter a new sector by acquiring an existing company ("buy"), while others do so using their existing resources ("build")? Using a novel data set constructed by merging French employer payrolls with commercial M\&A data sets, we show that firms are more likely to buy when their existing workforce does not include skills needed in the sector of entry. This relationship is more pronounced when labor market frictions make it difficult to hire key workers. Using a novel data set constructed by merging French employer payrolls with commercial M\&A data sets, we study firms' decision to build or buy. Firms with an existing workforce that is not adapted to the sector of entry are more likely to buy, especially when it is difficult to hire key workers for the sector of entry. Firms that enter by building realize lower entry sales when their existing workforce is not adapted to the sector of entry, especially in the presence of labor market frictions. Our results suggest that firms buy to acquire their targets' human capital when adapting their existing workforce is too costly.

We study the introduction of a microwave link between Frankfurt and London and identify its effects on German stocks that trade in both locations using French firms that trade only in London as the control group. We find that the link improves the integration between the two markets as evidenced by a reduction in arbitrage opportunities. The effects on market quality differ across stocks as high-frequency traders strategically utilize their speed advantage to supply or consume liquidity. In London, mid cap stocks benefit from increased liquidity provision whereas order flow for large stocks becomes more toxic. London's increased competitiveness for mid cap stocks also translates into an increase in its information share, while the information shares of large cap stocks are unchanged. These results suggest that market integration differently affects investors based on their speed advantage, trading venues based on their level of competitiveness, and stocks based on their size.

We show that the decision to participate in the stock market depends on the ability of equities to hedge the individual permanent earnings shocks, consistent with implications of life-cycle models. Those households who refrain from stock investing display positive correlation between their own permanent income innovations and market returns. These results owe to a two-step empirical strategy. First, a minimum distance estimation disentangles the aggregate from the idiosyncratic permanent component of labor income risks. The second step reconstructs the individual life-cycle dynamics of persistent shocks through a Kalman filter applied to the estimated labor income process. We are thus able to obtain the full cross-sectional distribution of individual correlations between permanent shocks and market returns.

We propose a framework that prices assets with respect to other assets and treat the excess return of every stock as a potential risk factor. In addition to the market, size or profitability premia, individual excess returns also carry a premium for their similarity with their peers. The peer risk premium manifests itself as a markup to fundamental factor loadings. Neglecting peer effects results in mispricing and misinterpreting factor risk as idiosyncratic risk. We find statistical evidence that markets price firm similarities by scrutinizing the S&P100 constituents. The Fama-French residuals of the returns contain a non-idiosyncratic common factor, which vanishes upon inclusion of peer effects. The model also produces a peer factor that withstands the five Fama-French factors.

We argue that risk sharing motivates the bank-wide structure of bonus pay. In the presence of financial frictions that make external financing costly, the optimal contract between shareholders and employees involves some degree of risk sharing whereby bonus pay partially absorbs earnings shocks. Using payroll data for 1.26 million employee-years in all functional divisions of Austrian, German, and Swiss banks, we uncover several empirical patterns in bonus pay that are difficult to rationalize with incentive theories of bonus pay---but support an important risk sharing motive. In particular, bonuses respond to performance shocks that are outside the control of employees because they originate in other bank divisions or even outside the bank.

We provide evidence on how personal shocks that reduce a CEO’s career horizon, triggered by either the CEO’s diagnosis of a serious illness or an illness or death of a close relative, affects corporate policies. We validate our identification strategy by showing that these events are not predictable based on observable characteristics and that the CEOs exposed to such events experience greater turnover rates and lower residual time-in-office. Following the shock, and while the CEO is still in office, these firms moderate both R&D and capital expenditures and increase in cash distributions. While these results are consistent with greater short-term orientation in this setting, it is not necessarily detrimental to shareholders, as performance increases in the aftermath of the shock. Earnings management, CEO compensation, and the likelihood of being acquired remain unchanged, indicating that the improved performance comes from the implementation of relatively more efficient firm policies rather than from opportunistic behavior.

Using a sample of 3,977 individual buyout managers with more than 10,330 deal involvements in 5,030 unique buyout transactions, we investigate individual manager performance and its persistence. We find evidence for deal-level gross PME performance persistence at this level of analysis. In explaining the cross-section of deal returns, the individual is around four times more important than the PE organization. Interestingly, none of the typical human capital variables (age, PE tenure, MBA, Ivy League attendance) has explanatory power in the cross-section of deal returns. Instead, we document that investment performance increases with a manager’s industry expertise. As a consequence, we find that individual persistence only exists for managers who stay within a certain industry, whereas it disappears for manager who invest across different industries.

Recent studies show that loss probability (LP) is a decisive factor when people
evaluate risk of assets in laboratory experiments, suggesting a positive relationship
between LP and expected stock returns. This corresponds to the classical \Safety-
First" principle. We nd strong empirical support for this prediction in the U.S.
stock market. During our sample period, average risk-adjusted return dierences
between stocks in the two extreme LP deciles exceed 0.75% per month. The positive
LP eect, characterized by the intention of some investors to pay low prices for
high LP stocks, remains signicant after controlling for traditional downside risk
measures.

How does uncertainty of crypto-assets affect traditional asset classes? Using a vector autoregression (VAR) methodology, I answer this question by analyzing volatility spillovers between five asset classes (crypto-assets, stocks, bonds, fiat-currencies, and commodities). Given the vast heterogeneity within each asset class, my VAR specification accounts for cross sectional variation across and within each asset class. By transforming the VAR residuals into sectoral shocks, I am able to distinguish between volatility spillovers across, and volatility co-movements within asset classes. I find that on average volatility of crypto-assets accounts for 15% of the volatility contagion received by traditional asset classes. The directional spillovers from crypto-asset to bonds and to fiat-currencies are particularly strong, capturing the wealth channel and the remittance channel, respectively.

We study the supply side of the market for CEOs as outside board members in the context of increasing legal and shareholder pressure on CEOs to refrain from accepting outside directorships. Consistent with potential benefits to their employing firms, CEOs have been increasingly more likely to accept directorships at vertically-related firms, especially when information and learning about upstream/downstream industries is expected to be particularly valuable. CEO directorships at vertically-related firms have a significant positive effect on firm value and performance and help their firms handle industry shocks. Consistent with agency/entrenchment costs, CEO outside directorships at unrelated firms are more likely for sending firms with weaker governance and have a significant negative effect on firm value. Taken together, our results suggest that external pressure has reduced the number of outside directorships that CEOs take and has improved their overall quality, although some evidence of agency problems remains.

There are large cross-sectional differences in the probability and magnitudes of mispricing among stocks. The traditional explanation for mispricing is limits to arbitrage. We show that, regardless of arbitrage costs, mispricing persists when professional investors lack incentives to expend investigative resources to detect mispricing. In the cross-section, these incentives are driven by the expected dollar gains from arbitrage opportunities. Arbitrageurs rationally neglect to investigate stocks if the expected arbitrage revenue is small. For such stocks, information discovery by investors is slow and the mispricing is corrected mostly through mandatory financial disclosures by firms. Using measures of institutional attention and trading discreteness we confirm this mechanism empirically. The revenue potential channel explains persistent mispricing better than any classic arbitrage frictions.

In this paper we find that the decline in the momentum profitability is partly driven by option trading. Momentum profits arise from the short leg and therefore on barriers to short selling. We find strong evidence that the presence of stock options creates alternate avenues for short selling, augmenting the stock lending market, thus contributing to improved pricing efficiency. However, when option trading becomes expensive, there are more barriers to short selling and the short position offers lower returns. Our results are robust to exogenous changes in short selling due to Regulation SHO and are supported by other short-leg depended anomalies.