Publications

Non-Standard Errors
Forthcoming in Journal of Finance
(with Albert Menkveld and 341 other coauthors).
SSRN | blog
Abstract: In statistics, samples are drawn from a population in a data-generating process (DGP). Standard errors measure the uncertainty in estimates of population parameters. In science, evidence is generated to test hypotheses in an evidence-generating process (EGP). We claim that EGP variation across researchers adds uncertainty: Non-standard errors (NSEs). We study NSEs by letting 164 teams test the same hypotheses on the same data. NSEs turn out to be sizable, but smaller for better reproducible or higher rated research. Adding peer-review stages reduces NSEs. We further find that this type of uncertainty is underestimated by participants.

Bias in the Effective Bid-Ask Spread
Published 2021 in Journal of Financial Economics, 142: 314-337.
Awarded the De la Vega Prize 2017
Published version (open access) | blog | SSRN  | internet appendix | code
Abstract: The effective bid-ask spread measured relative to the spread midpoint overstates the true effective bid-ask spread in markets with discrete prices and elastic liquidity demand. The average bias is 13-18% for S&P 500 stocks in general, depending on the estimator used as benchmark, and up to 97% for low-priced stocks. Cross-sectional bias variation across stocks, trading venues, and investor groups can influence research inference. The use of the midpoint also undermines liquidity timing and trading performance evaluations, and can lead non-sophisticated investors to overpay for liquidity. To overcome these problems, the paper proposes new estimators of the effective bid-ask spread.

Do Volatility Extensions Improve the Quality of Closing Call Auctions?
(with Ester Félez Viñas)
Published 2021 in Financial Review, 56(3): 385-406.
Published version (open access) | blog | SSRN | internet appendix
Abstract: To improve the efficiency of the closing price, many equity exchanges apply volatility extensions to their closing call auctions. If an imminent auction execution implies a large price change, the order submission period is extended to let traders reconsider their orders. This paper uses the introduction of closing auction volatility extensions at NASDAQ Nordic to provide the first analysis of the effects of such mechanisms. We find that the volatility extensions reduce transitory volatility and deter price manipulation at the close. Consistent with increased trust in the mechanism, the closing call auction attracts higher volumes after the change.

Information Revelation in Decentralized Markets
(with Albert Menkveld)
Published 2019 in Journal of Finance 74(6): 2751-2787. 
Published version | SSRN | slides | internet appendix | code
Abstract: How does information get revealed in decentralized markets? We test several hypotheses inspired by recent dealer-network theory. To do so we construct an empirical map of information revelation where two dealers are connected based on the synchronicity of their quote changes. The tests, based on EUR/CHF quote data including the 2015 crash, largely support theory: Strongly connected (i.e., central) dealers are more informed. Connections are weaker when there is less to be learned. The crash serves to identify how a network forms when dealers are transitioned from no-learning to learning, that is, from a fixed to a floating rate.

Risk and Return in High-Frequency Trading
(with Matthew Baron, Jonathan Brogaard, and Andrei Kirilenko).
Published 2019 in Journal of Financial and Quantitative Analysis 54(3): 993-1024.
Published version | SSRN | internet appendix
Abstract: We study performance and competition among high-frequency traders (HFTs). We construct measures of latency and find that differences in relative latency account for large differences in HFTs’ trading performance. HFTs that improve their latency rank due to colocation upgrades see improved trading performance. The stronger performance associated with speed comes through both the short-lived information channel and the risk management channel, and speed is useful for various strategies including market making and cross-market arbitrage. We find empirical support for many predictions regarding relative latency competition.

Components of the bid–ask spread and variance: A unified approach
(with Richard Henricson and Lars Nordén).
Published 2016 in Journal of Futures Markets 36(6): 545-563.
Published version | SSRN
Abstract: We develop a structural model for the price formation and liquidity supply of an asset. Our model facilitates decompositions of both the bid–ask spread and the return variance into components related to adverse selection, inventory, and order processing costs. Furthermore, the model shows how the fragmentation of trading volume across trading venues influences inventory pressure and price discovery. We use the model to analyze intraday price formation for gold futures traded at the Shanghai Futures Exchange. We find that order processing costs explain about 50% of the futures bid–ask spread, whereas the remaining 50% is equally due to asymmetric information and to inventory costs. About a third of the variance in futures returns is attributable to microstructure noise. Trading at the spot market has a significant influence on futures price discovery, but only a limited impact on the futures bid–ask spread.

Trading fast and slow: Colocation and market quality
(with Jonathan Brogaard, Lars Nordén, and Ryan Riordan).
Published 2015 in Review of Financial Studies 28(12): 3407-3443.
Published version | SSRN
Abstract: We exploit an optional colocation upgrade at NASDAQ OMX Stockholm to assess how speed affects market liquidity. Liquidity improves for the overall market and even for non-colocated trading entities. We find that the upgrade is pursued mainly by market-maker-type participants. Those that upgrade use their enhanced speed to reduce their exposure to adverse selection and to relax their inventory constraints. In particular, the upgraded trading entities remain competitive at the best bid and offer even when they are inventory constrained. Our results suggest that increasing the speed of market making participants can have benefits for market liquidity.

The aggressiveness of high-frequency traders
(with Lars Nordén and Dong Zhang).
Published 2014 in Financial Review 49(2), 395-419 (a special issue on HFT).
Published version | SSRN | internet appendix
Abstract: We study order aggressiveness of market-making high-frequency traders (HFTs), opportunistic HFTs, and non-HFTs. We find that market-making HFTs follow their own group’s previous order submissions more than they follow other traders’ orders. Opportunistic HFTs and non-HFTs tend to split market orders into small portions submitted in sequence. HFTs submit more (less) aggressive orders when the same-side (opposite-side) depth is large, and supply liquidity when the bid-ask spread is wide. Thus, HFTs adhere strongly to the trade-off between waiting cost and the cost of immediate execution. Non-HFTs care less about this trade-off, but react somewhat stronger than HFTs to volatility.

Closing call auctions at the index futures market
(with Lars Nordén).
Published 2014 in Journal of Futures Markets, 34(4), 299-319.
Published version | SSRN
Abstract: This paper investigates how the introduction of a closing call auction in the OMXS 30 index futures market influences market quality and price accuracy. Index futures markets are characterized by traders with no or little private information. Limit order book models where trader patience (rather than private information) determines trading strategies, predict that a closing call auction increases trader patience and hence improves closing price accuracy and end-of-day market liquidity. We analyze futures market liquidity in three dimensions: tightness, depth, and resiliency. Our empirical results show that the closing call auction indeed leads to increased trader patience and successfully improves the futures closing price accuracy. However, tightness and resiliency are unaffected by the regulatory change, and depth is decreasing. We hypothesize that the depth effect is due to an “order fishing” phenomenon, which is not considered in current theoretical models. When the potential of large market orders is high, opportunistic patient traders post limit orders in the depth of the order book to profit from impatient traders. In line with our hypothesis, order fishing activity increases sharply in the last minute of the trading day. When the closing call auction is introduced, and trader patience increases, the order fishing behavior vanishes.

The diversity of high-frequency traders
(with Lars Nordén).
Published 2013 in a special issue on HFT in Journal of Financial Markets, 16(4), 741-770.
Published version | SSRN
Abstract: The regulatory debate concerning high-frequency trading (HFT) emphasizes the importance of distinguishing different HFT strategies and their influence on market quality. Using data from NASDAQ-OMX Stockholm, we compare market-making HFTs to opportunistic HFTs. We find that market makers constitute the lion’s share of HFT trading volume (63–72%) and limit order traffic (81–86%). Furthermore, market makers have higher order-to-trade ratios and lower latency than opportunistic HFTs. In a natural experiment based on tick size changes, we find that the activity of market-making HFTs mitigates intraday price volatility.

The Components of the illiquidity premium: An empirical analysis of U.S. stocks 1927-2010
(with Björn Hansson and Birger Nilsson).
Published 2013 in Journal of Banking and Finance 37(11), 4476-4487.
Published version | SSRN
Abstract: This paper implements a conditional version of the liquidity adjusted CAPM (LCAPM). The conditional LCAPM allows for a time-varying decomposition of the total illiquidity premium into a level component and three risk components. The estimated average annual total illiquidity premium for US stocks 1927-2010 is 1.74%-2.08%, which is substantially lower than in most previous studies. The contributions from illiquidity level and illiquidity risk are 1.25%-1.28% and 0.46%-0.83%, respectively. Of the three illiquidity risk components, risk related to the hedging of wealth shocks is the most important, while commonality risk is the least important. The illiquidity premia are clearly time-varying, with peaks in downturns and crises, but with no general tendency to decrease over time. The level premium and the risk premium are significantly positively correlated around 0.35; indicating that in periods of turbulence both illiquidity cost and illiquidity risk premia tend to be high.

Alchemy in the 21st Century: Hedging with gold futures
(with Caihong Xu and Lars Nordén).
Published 2011 in Review of Futures Markets 19, 247–281.
Published version
Abstract: The Shanghai gold exchange is the largest spot gold exchange in the world, and the Shanghai gold futures contract, introduced in 2008, is already the fourth largest gold futures contract in the world. This paper is the first to study the Chinese gold market, analyzing hedging strategies utilizing the Shanghai gold futures. The results show that hedging with gold futures reduces the variance of an unhedged gold spot position by about 88% in its first two years of existence. During the second half of 2008, however, when the global financial crisis escalated, the variance reduction dropped to about 70%. Overall, the new Chinese gold futures prove to be attractive and well-needed hedging vehicles for domestic Chinese gold producers, refiners, consumers, and investors. Furthermore, it is found that the regression hedge outperforms bivariate GARCH hedging strategies out-of-sample, even though the latter are better suited to describe the return processes in-sample.

Causality in crude oil prices
(with Szymon Wlazlowski and Monica Giulietti).
Published 2011 in Applied Economics 43, 3337-3347.
Published version
Abstract: Crude oil markets witness growing disparity between the quality of crudes supplied and demanded in the market. The market share of low‐quality crudes is increasing due to the depletion of old fields and increasing demand. This is unnerving the practitioners and affecting the relevance of the traditional benchmark crudes due to the lack of lower quality benchmarks (Montepeque, 2005). In this article, we apply Granger causality tests to study the price dependence of 32 crudes in order to establish which crudes drive other prices and which ones simply follow general market trends. Our results indicate that some of the old benchmarks are still relevant while others can be disregarded. Our results also interestingly show that the low-quality Mediterranean Russian Urals crude, introduced in the late 1990s, has emerged recently as a significant driver of global prices.

Stock portfolio selection with full-scale optimization and differential evolution
(with Jane Binner).
Published 2009 in Applied Financial Economics 19, 1559–1571.
Published version
Abstract: Full-Scale Optimization (FSO) is a utility maximization approach to portfolio choice problems that has theoretical appeal but that suffers from computational burden in large scale problems. We apply the heuristic technique differential evolution to solve FSO-type asset selection problems of 97 assets under complex utility functions rendering rough utility search surfaces. We show that this problem is computationally feasible and that solutions retrieved with random starting values are converging to one optimum. Furthermore, the study constitutes the first FSO application to stock portfolio optimization. The results indicate that when investors are loss averse, FSO improves stock portfolio performance compared to Mean Variance (MV) portfolios. This finding widens the scope of applicability of FSO, but it is also stressed that out-of-sample success will always be dependent on the forecasting ability of the input return distributions.

Mean-variance vs. full-scale optimization: Broad evidence for the UK
(with Richard Anderson, Jane Binner, Thomas Elger, and Birger Nilsson).
Published 2008 in The Manchester School 76, 134–156.
Published version
Abstract: Portfolio choice by full-scale optimization applies the empirical return distribution to a parameterized utility function, and the maximum is found through numerical optimization. Using a portfolio choice setting of three UK equity indices we identify several utility functions featuring loss aversion and prospect theory, under which full-scale optimization is a substantially better approach than the mean–variance approach. As the equity indices have return distributions with small deviations from normality, the findings indicate much broader usefulness of full-scale optimization than has earlier been shown. The results hold in- and out-of-sample, and the performance improvements are given in terms of utility as well as certainty equivalents.