Contributions in Merlin

Aggregators are facing increased scrutiny by regulatory authorities, suggesting these sites have considerable market power. On the other extreme, firms are bypassing aggregators, choosing instead to sell directly to consumers. This raises the question as to which party has more market power: the aggregator or the individual firm. Focusing on the airline industry, we investigate who benefits the most in the airline-aggregator relationship. Specifically, we ask what would happen to airline and aggregator site visits and purchases in the absence of a comprehensive aggregator. We first explore consumers’ search patterns on Southwest, an airline that has never been part of any aggregator. In a descriptive exercise, we find that consumers who book on Southwest are the least likely to visit aggregator sites. Second, we use the 2011 American dispute with Orbitz as an exogenous event, which led to American fares no longer being displayed on Orbitz for five months. We use this dispute to identify who was hurt the most – the aggregator or the airline - in the months following the dispute. Our findings indicate the aggregator loses the most when it is not comprehensive.

Digitization has vastly increased the amount of new music produced and available directly to consumers. While this has levelled the playing field between already-prominent and new artists, creators maynow be dependent on platform decisions about which songs and artist to promote. With Spotify emergingas a major interactive streaming platform, this paper explores the effect of Spotify’s playlists on boththe promotion of songs and the discovery of music by new artists, using four approaches. First, weexamine songs’ streaming volumes before and after their addition to, and removal from, major globalplaylists. Second, we compare streaming volumes for songs just on, and just off, algorithmic top 50 playlists. Third, we make use of cross-country differences in inclusion on New Music Friday lists,using song fixed effects to explain differences in streaming. Fourth, we develop an instrumental variables approach to explaining cross-country New Music Friday rank differentials based on home bias. Beingadded to Today’s Top Hits, a list with 18.5 million followers during the sample period, raises streams by almost 20 million and is worth between 116,000 Dollar and 163,000 Dollar. Inclusion on New Music Friday lists substantially raises the probability of song success, including for new artists.

The need to harmonise apparently irreconcilable arrangements in an ecosystem –nestedness andsegregation– has triggered so far different strategies. Methodological refinements, or the inclusion ofbehavioural preferences to the network dynamics offer a limited approach to the problem, since oneremains constrained within a 2-dimensional view of an ecosystem, i.e. a presence-absence matrix.Here we question this partial-view paradigm, and show that nestedness and segregation may coexistacross a varied range of scenarios. To do so, we rely on an upscaled representation of an ecologicalcommunity as ann-partite hypergraph, inspired by Hutchinson’s high-dimensional niche conceptand the latest trends on ecological multilayer networks. This yields an inclusive description of anecological system, for which we develop a natural extension of the definition of nestedness to largerdimensional spaces, revealing how competitive exclusion may operate regardless of a highly nestedbipartite system.

Cryptocurrencies are distributed systems that allow exchanges of native tokens among par-ticipants, or the exchange of such tokens for fiat currencies in markets external to these public ledgers. The availability of their complete historical bookkeeping opens up the possibility of understanding the relationship between aggregated users’ behaviour and the cryptocur-rency pricing in exchange markets. This paper analyses the properties of the transaction network of Bitcoin. We consider four different representations of it, over a period of nine years since the Bitcoin creation and involving 16 million users and 283 million transactions. By analysing these networks, we show the existence of causal relationships between Bitcoin price movements and changes of its transaction network topology. Our results reveal the in-terplay between structural quantities, indicative of the collective behaviour of Bitcoin users, and price movements, showing that, during price drops, the system is characterised by a larger heterogeneity of nodes activity.

null

Graph or network representations are an important foundation for data mining and machine learning tasks in relational data. Many tools of network analysis, like centrality measures, information ranking, or cluster detection rest on the assumption that links capture direct influence, and that paths represent possible indirect influence. This assumption is invalidated in time-stamped network data capturing, e.g., dynamic social networks, biological sequences or financial transactions. In such data, for two time-stamped links (A,B) and (B,C) the chronological ordering and timing determines whether a causal path from node A via B to C exists. A number of works has shown that for that reason network analysis cannot be directly applied to time-stamped network data. Existing methods to address this issue require statistics on causal paths, which is computationally challenging for big data sets. Addressing this problem, we develop an efficient algorithm to count causal paths in time-stamped network data. Applying it to empirical data, we show that our method is more efficient than a baseline method implemented in an OpenSource data analytics package. Our method works efficiently for different values of the maximum time difference between consecutive links of a causal path and supports streaming scenarios. With it, we are closing a gap that hinders an efficient analysis of big time series data on complex networks.

The unsupervised detection of anomalies in time series data has important applications, e.g., in user behavioural modelling, fraud detection, and cybersecurity. Anomaly detection has been extensively studied in categorical sequences, however we often have access to time series data that contain paths through networks. Examples include transaction sequences in financial networks, click streams of users in networks of cross-referenced documents, or travel itineraries in transportation networks. To reliably detect anomalies we must account for the fact that such data contain a large number of independent observations of short paths constrained by a graph topology. Moreover, the heterogeneity of real systems rules out frequency-based anomaly detection techniques, which do not account for highly skewed edge and degree statistics. To address this problem we introduce a novel framework for the unsupervised detection of anomalies in large corpora of variable-length temporal paths in a graph, which provides an efficient analytical method to detect paths with anomalous frequencies that result from nodes being traversed in unexpected chronological order.

We study the impact on the skill premium of increases in the quality of goods consumed by households(“trading up”). Our empirical work shows that high- quality goods are more intensive in skilled laborthan low-quality goods and that household spending on high-quality goods rises with income. Wepropose a model consistent with these facts. This model accounts for the past rise in the skill premiumwith more plausible rates of skill-biased technical change than those required by the canonical model.It also implies that an expansion of the skilled labor force reduces the skill premium by much lessthan in the canonical model.

This study empirically investigates whether the relationship between the fraction of filled apprenticeships in a particular occupation in the past and the fraction of prospective apprentices having very early intentions to train in this occupation has an impact on the decision to change the intended choice of occupation. We use a unique dataset from Switzerland containing detailed information on students’ early occupational ‘dreams’ (ages 13-14), before they undergo intensive career counselling, and combine it with information on their ultimate choice of occupation at the end of compulsory schooling (ages 15-16). The estimation results show that although the majority of students revise their initial intentions, those students who dreamed of learning an occupation with more training positions filled in previous years than peers interested in learning this occupation have a significantly higher probability of sticking to their initial dream occupation. Conversely, students who wished to train in an overly popular occupation have a higher probability of delaying the transition to upper-secondary education for at least one year, instead of switching to another occupation. In addition, we find on an aggregated level that a favourable situation on the apprenticeship market ultimately increases the premature contract termination rate due to a person-occupation-mismatch.

We review and condense the body of literature on the economic returns of public R&D on private R&D and find that: (i)private returns to R&D appear to be large and larger than the returns to alternative investments; (ii) private R&D and R&D subsidies are positively correlated and there is no evidence for crowding out; (iii)R&D cooperation increases private R&D; (iv) there appear to exist complementarities between alternative sources of funding; (v) the mobility of R&D workers, particularly of university scientists,is positively related to innovation; (vi) there are many university spin-offs but these are no more successful than non-university spin-offs; (vii) universities constitute important collaboration partnersand (viii)clusters enhance collaboration, patents and productivity. Key problems for economic policy advice are that the identification of causal effects is problematic in most studies and that little is known about the optimal design of policy measures.