Towards robust network based complex systems : from evolutionary cellular automata to biological models

Abstract Sitting between your past and your future doesn't mean you are in the present. Dakota Skye Complex systems science is an interdisciplinary field grouping under the same umbrella dynamical phenomena from social, natural or mathematical sciences. The emergence of a higher order organization or behavior, transcending that expected of the linear addition of the parts, is a key factor shared by all these systems. Most complex systems can be modeled as networks that represent the interactions amongst the system's components. In addition to the actual nature of the part's interactions, the intrinsic topological structure of underlying network is believed to play a crucial role in the remarkable emergent behaviors exhibited by the systems. Moreover, the topology is also a key a factor to explain the extraordinary flexibility and resilience to perturbations when applied to transmission and diffusion phenomena. In this work, we study the effect of different network structures on the performance and on the fault tolerance of systems in two different contexts. In the first part, we study cellular automata, which are a simple paradigm for distributed computation. Cellular automata are made of basic Boolean computational units, the cells; relying on simple rules and information from- the surrounding cells to perform a global task. The limited visibility of the cells can be modeled as a network, where interactions amongst cells are governed by an underlying structure, usually a regular one. In order to increase the performance of cellular automata, we chose to change its topology. We applied computational principles inspired by Darwinian evolution, called evolutionary algorithms, to alter the system's topological structure starting from either a regular or a random one. The outcome is remarkable, as the resulting topologies find themselves sharing properties of both regular and random network, and display similitudes Watts-Strogtz's small-world network found in social systems. Moreover, the performance and tolerance to probabilistic faults of our small-world like cellular automata surpasses that of regular ones. In the second part, we use the context of biological genetic regulatory networks and, in particular, Kauffman's random Boolean networks model. In some ways, this model is close to cellular automata, although is not expected to perform any task. Instead, it simulates the time-evolution of genetic regulation within living organisms under strict conditions. The original model, though very attractive by it's simplicity, suffered from important shortcomings unveiled by the recent advances in genetics and biology. We propose to use these new discoveries to improve the original model. Firstly, we have used artificial topologies believed to be closer to that of gene regulatory networks. We have also studied actual biological organisms, and used parts of their genetic regulatory networks in our models. Secondly, we have addressed the improbable full synchronicity of the event taking place on. Boolean networks and proposed a more biologically plausible cascading scheme. Finally, we tackled the actual Boolean functions of the model, i.e. the specifics of how genes activate according to the activity of upstream genes, and presented a new update function that takes into account the actual promoting and repressing effects of one gene on another. Our improved models demonstrate the expected, biologically sound, behavior of previous GRN model, yet with superior resistance to perturbations. We believe they are one step closer to the biological reality.

Explicit ruin formulas for models with dependence among risks

We show that a simple mixing idea allows one to establish a number of explicit formulas for ruin probabilities and related quantities in collective risk models with dependence among claim sizes and among claim inter-occurrence times. Examples include compound Poisson risk models with completely monotone marginal claim size distributions that are dependent according to Archimedean survival copulas as well as renewal risk models with dependent inter-occurrence times.

SESAM - a new framework integrating macroecological and species distribution models for predicting spatio-temporal patterns of species assemblages

Two different approaches currently prevail for predicting spatial patterns of species assemblages. The first approach (macroecological modelling, MEM) focuses directly on realised properties of species assemblages, whereas the second approach (stacked species distribution modelling, S-SDM) starts with constituent species to approximate assemblage properties. Here, we propose to unify the two approaches in a single 'spatially-explicit species assemblage modelling' (SESAM) framework. This framework uses relevant species source pool designations, macroecological factors, and ecological assembly rules to constrain predictions of the richness and composition of species assemblages obtained by stacking predictions of individual species distributions. We believe that such a framework could prove useful in many theoretical and applied disciplines of ecology and evolution, both for improving our basic understanding of species assembly across spatio-temporal scales and for anticipating expected consequences of local, regional or global environmental changes. In this paper, we propose such a framework and call for further developments and testing across a broad range of community types in a variety of environments.

Sibling rivalry and vocal negotiation in the barn owl Tyto alba

Chez les animaux, les jeunes dépendant des parents durant leur développement sont en compétition pour obtenir la nourriture, qu'ils quémandent par des cris et postures ostentatoires et se disputent physiquement. Les frères et soeurs n'ont pas la même compétitivité, en particulier s'ils diffèrent en âge, et leur niveau de faim fluctue dans le temps. Comme dans tout type de compétition, chacun doit ajuster son investissement aux rivaux, c'est à dire aux besoins et comportements de ses frères et soeurs. Dans le contexte de la famille, selon la théorie de sélection de parentèle, les jeunes bénéficient de leur survie mutuelle et donc de la propagation de la part de gènes qu'ils ont en commun. L'hypothèse de la « négociation frères-soeurs » prédit que, sous certaines conditions, les jeunes négocient entre eux la nourriture, ce qui réduit les coûts de compétition et permet de favoriser les frères et soeurs les plus affamés. La littérature actuelle se focalise sur les signaux de quémande entre enfants et parents et les interactions compétitives frères-soeurs sont étudiées principalement au sein de paires, alors que les nichées ou portées en comprennent souvent de nombreux. Cette thèse vise à mieux comprendre comment et jusqu'à quel point plusieurs jeunes ajustent mutuellement leurs signaux de besoin. C'est une question importante, étant donné que cela influence la répartition de nourriture entre eux, donc la résolution du conflit qui les oppose et à terme leur valeur évolutive. Le modèle d'étude est la chouette effraie (Tyto alba), chez laquelle jusqu'à neufs poussins émettent des milliers de cris chacun par nuit. Ils négocieraient entre eux la prochaine proie indivisible rapportée au nid avant que les parents ne reviennent : un poussin affamé crie plus qu'un autre moins affamé, ce qui dissuade ce dernier de crier en retour et par la suite de quémander la nourriture aux parents. L'investissement optimal correspondrait donc à écarter son frère en permanence vu que l'arrivée des parents est imprévisible, mais à moindre coût. Dans un premier axe, nous avons exploré au sein de dyades les mécanismes acoustiques permettant aux poussins de doser leur effort vocal durant les heures de compétition où ils sont laissés seuls au nid. Nous avons trouvé que les poussins évitent de crier simultanément, ce qui optimiserait la discrimination du nombre et de la durée de leurs cris, lesquels reflètent de façon honnête leur niveau de faim et donc leur motivation. L'alternance des cris paraît particulièrement adaptée au fait que les poussins se fient à des variations temporelles subtiles dans le rythme et la durée de leurs vocalisations pour prendre la parole. En particulier, allonger ses cris tout en criant moins dissuade efficacement le rival de répondre, ce qui permet de monopoliser la parole dans de longs « monologues ». Ces règles seraient universelles puisqu'elles ne dépendent pas de la séniorité, de la faim, ni de la parenté et les poussins répondent à un playback de façon similaire à un vrai frère. Tous ces résultats apportent la première preuve expérimentale que les juvéniles communiquent de façon honnête sur leurs besoins, ajustent activement le rythme de leurs cris et utilisent des composantes multiples de leurs vocalisations d'une façon qui réduit le coût de la compétition. De plus, il s'agit de la première démonstration que des règles de conversation régissent de longs échanges vocaux chez les animaux de façon comparable aux règles basiques observées chez l'Homme. Dans un second axe, nous avons exploré les stratégies comportementales que les poussins adoptent pour rivaliser avec plusieurs frères et soeurs, par le biais d'expériences de playback. Nous avons trouvé que les poussins mémorisent des asymétries de compétitivité entre deux individus qui dialoguent et répondent plus agressivement au moins compétitif une fois qu'ils sont confrontés à chacun isolément. Dans la même ligne, quand ils entendent un nombre variable d'individus criant à un taux variable, les poussins investissent le plus contre des rivaux moins nombreux et moins motivés. En accord avec les prédictions des modèles théoriques, les poussins de chouette effraie escaladent donc les conflits pour lesquels leur chance de gagner contrebalance le plus l'énergie dépensée. Nous révélons ainsi que 1) les jeunes frères et soeurs 'espionnent' les interactions de leurs rivaux pour évaluer leur compétitivité relative, ce qui est sans doute moins coûteux qu'une confrontation directe avec chacun, et 2) dosent leur investissement vocal en fonction du nombre de rivaux actuellement en compétition et de leur motivation de façon concomitante. Ces résultats montrent que les interactions entre frères et soeurs au nid reposent sur des mécanismes similaires à ceux observés, mais encore de façon anecdotique, chez les adultes non apparentés qui se disputent les territoires et partenaires sexuels. Cette thèse souligne donc combien il est crucial de considérer dorénavant la famille comme un réseau de communication à part entière pour mieux comprendre comment les jeunes résolvent les conflits autour du partage des ressources parentales. Plus généralement, elle révèle l'importance de la dynamique temporelle des vocalisations dans les conflits et la communication des animaux. A la lumière de nos résultats, la chouette effraie apparaît comme un modèle clé pour de futures recherches sur la résolution des conflits et la communication acoustique. - In species with parental care, offspring contest priority access to food by begging through conspicuous postures and vocalisations and by physically jockeying. Siblings differ in their competitiveness, especially in the case of age and size hierarchies, and their hunger level fluctuates in time. As in competition in general, each individual should adjust its investment to opponents that is to say to its siblings' needs and behaviours. In the particular context of family, according to kin selection theory, siblings derive extra fitness benefits from their mutual survival and hence the spreading of the genes they share. The "sibling negotiation" predicts that, under certain conditions, young would negotiate among them priority access to food, which reduces competition costs and enables promoting the most hungry siblings. To date, the literature focuses on signals of need between parents and offspring and competitive interactions (in particular among siblings) are mostly studied within pairwise interactions, yet they commonly involve more numerous rivals. This PhD aims at better understanding how and the extent to which several young siblings compete through signalling. This is important since this influences how food is allocated among them, thus the outcome of sibling rivalry and ultimately their fitness. I use the barn owl (Tyto alba) as a model, in which the one to nine nestlings emit a simple noisy call thousands of times per night. Thereby, they would negotiate among them priority access to the indivisible food next delivered prior to parents' feeding visits. A hungry nestling emits more calls than a less hungry sibling, which deters it to call in return and ultimately beg food at parents. The optimal investment thus corresponds to constantly deterring the rival to compete, given that parents' arrival is unpredictable, but at the lowest costs. In the first axis of my thesis, we explored within dyads the acoustic mechanisms by which owlets dose vocal effort when competing during the hours they are left alone. We found that owlets avoid overlapping each other's calls. This would enhance the discrimination of both call number and duration, which honestly reflect individuals' hunger level and hence motivation to compete. Such antiphony seems best adapted to the fact that siblings actually use subtle temporal variations in the rhythm and duration of their calls to take or give their turn. Owlets alternate monologs, in which lengthening calls efficiently deters the rival to respond while reducing call number. Such rules depend neither on seniority, hunger level nor kinship since nestlings responded similarly to a live sibling and an unrelated playback individual. Taken together, these findings provide the first experimental proof that dependent young honestly communicate about their need, actively adjust the timing of their calls and use multicomponent signals in a way that reduces vocal costs. Moreover, this is the first demonstration of conversational rules underlying animal long-lasting vocal exchanges comparable to the basic turn-taking signals observed in humans. In the second axis, we focused on the behavioural strategies owlets adopt to compete with more than one sibling, using playback experiments. We found that singleton bystanders memorised competitive asymmetries between two playback individuals dialoguing and responded more aggressively to the submissive one once they later faced each of both alone. Moreover, when hearing a varying number of nestlings calling at varying rates, owlets vocally invested the most towards fewer and less motivated rivals. In line with predictions from models on conflict settlement, barn owls thus escalate contests in which their chance of winning best counterbalances the energy spent. These results reveal that young socially eavesdrop on their siblings' interactions to assess their relative competitiveness at likely lower costs than direct confrontation, and dose vocal effort relative to both their number and motivation. This shows that young siblings' interactions imply mechanisms similar to those observed, yet still anecdotally, in unrelated adults that contest mates and territories. This PhD therefore highlights how crucial it is to further consider family as a communication network to better understand how siblings resolve conflicts over the share of parental resources. More generally, it provides important insights into the role of the temporal dynamics of signalling during animal contests and communication. In the light of our findings, the barn owl emerges as a key model for future research on conflict resolution and acoustic communication in animals.

Measurement with Some Theory: a New Approach to Evaluate Business Cycle Models

We propose a method to evaluate cyclical models which does not require knowledge of the DGP and the exact empirical specification of the aggregate decision rules. We derive robust restrictions in a class of models; use some to identify structural shocks and others to evaluate the model or contrast sub-models. The approach has good size and excellent power properties, even in small samples. We show how to examine the validity of a class of models, sort out the relevance of certain frictions, evaluate the importance of an added feature, and indirectly estimate structural parameters.

Méthodes sismiques actives adaptées à l'étude de milieux hétérogènes : application aux glissements de terrain

Résumé Les glissements de terrain représentent un des principaux risques naturels dans les régions montagneuses. En Suisse, chaque année les glissements de terrains causent des dégâts qui affectent les infrastructures et ont des coûts financiers importants. Une bonne compréhension des mécanismes des glissements peut permettre d'atténuer leur impact. Celle-ci passe notamment par la connaissance de la structure interne du glissement, la détermination de son volume et de son ou ses plans de glissement. Dans un glissement de terrain, la désorganisation et la présence de fractures dans le matériel déplacé engendre un changement des paramètres physiques et en particulier une diminution des vitesses de propagation des ondes sismiques ainsi que de la densité du matériel. Les méthodes sismiques sont de ce fait bien adaptées à l'étude des glissements de terrain. Parmi les méthodes sismiques, l'analyse de la dispersion des ondes de surface est une méthode simple à mettre en oeuvre. Elle présente l'avantage d'estimer les variations des vitesses de cisaillement avec la profondeur sans avoir spécifiquement recours à l'utilisation d'une source d'onde S et de géophones horizontaux. Sa mise en oeuvre en trois étapes implique la mesure de la dispersion des ondes de surface sur des réseaux étendus, la détermination des courbes de dispersion pour finir par l'inversion de ces courbes. Les modèles de vitesse obtenus à partir de cette procédure ne sont valides que lorsque les milieux explorés ne présentent pas de variations latérales. En pratique cette hypothèse est rarement vérifiée, notamment pour un glissement de terrain dans lequel les couches remaniées sont susceptibles de présenter de fortes hétérogénéités latérales. Pour évaluer la possibilité de déterminer des courbes de dispersion à partir de réseaux de faible extension des mesures testes ont été effectuées sur un site (Arnex, VD) équipé d'un forage. Un profil sismique de 190 m de long a été implanté dans une vallée creusée dans du calcaire et remplie par des dépôts glacio-lacustres d'une trentaine de mètres d'épaisseur. Les données acquises le long de ce profil ont confirmé que la présence de variations latérales sous le réseau de géophones affecte l'allure des courbes de dispersion jusqu'à parfois empêcher leur détermination. Pour utiliser l'analyse de la dispersion des ondes de surface sur des sites présentant des variations latérales, notre approche consiste à déterminer les courbes de dispersions pour une série de réseaux de faible extension, à inverser chacune des courbes et à interpoler les différents modèles de vitesse obtenus. Le choix de la position ainsi que de l'extension des différents réseaux de géophones est important. Il tient compte de la localisation des hétérogénéités détectées à partir de l'analyse de sismique réfraction, mais également d'anomalies d'amplitudes observées sur des cartes qui représentent dans le domaine position de tir - position du récepteur, l'amplitude mesurée pour différentes fréquences. La procédure proposée par Lin et Lin (2007) s'est avérée être une méthode efficace permettant de déterminer des courbes de dispersion à partir de réseaux de faible extension. Elle consiste à construire à partir d'un réseau de géophones et de plusieurs positions de tir un enregistrement temps-déports qui tient compte d'une large gamme de distances source-récepteur. Au moment d'assembler les différentes données une correction de phase est appliquée pour tenir compte des hétérogénéités situées entre les différents points de tir. Pour évaluer cette correction nous suggérons de calculer pour deux tir successif la densité spectrale croisée des traces de même offset: Sur le site d'Arnex, 22 courbes de dispersions ont été déterminées pour de réseaux de géophones de 10 m d'extension. Nous avons également profité du forage pour acquérir un profil de sismique verticale en ondes S. Le modèle de vitesse S déduit de l'interprétation du profil de sismique verticale est utilisé comme information à priori lors l'inversion des différentes courbes de dispersion. Finalement, le modèle en deux dimension qui a été établi grâce à l'analyse de la dispersion des ondes de surface met en évidence une structure tabulaire à trois couches dont les limites coïncident bien avec les limites lithologiques observées dans le forage. Dans celui-ci des argiles limoneuses associées à une vitesse de propagation des ondes S de l'ordre de 175 m/s surmontent vers 9 m de profondeur des dépôts de moraine argilo-sableuse caractérisés par des vitesses de propagation des ondes S de l'ordre de 300 m/s jusqu'à 14 m de profondeur et supérieur ou égal à 400 m/s entre 14 et 20 m de profondeur. Le glissement de la Grande Combe (Ballaigues, VD) se produit à l'intérieur du remplissage quaternaire d'une combe creusée dans des calcaires Portlandien. Comme dans le cas du site d'Arnex les dépôts quaternaires correspondent à des dépôts glacio-lacustres. Dans la partie supérieure la surface de glissement a été localisée à une vingtaine de mètres de profondeur au niveau de l'interface qui sépare des dépôts de moraine jurassienne et des dépôts glacio-lacustres. Au pied du glissement 14 courbes de dispersions ont été déterminées sur des réseaux de 10 m d'extension le long d'un profil de 144 m. Les courbes obtenues sont discontinues et définies pour un domaine de fréquence de 7 à 35 Hz. Grâce à l'utilisation de distances source-récepteur entre 8 et 72 m, 2 à 4 modes de propagation ont été identifiés pour chacune des courbes. Lors de l'inversion des courbes de dispersion la prise en compte des différents modes de propagation a permis d'étendre la profondeur d'investigation jusqu'à une vingtaine de mètres de profondeur. Le modèle en deux dimensions permet de distinguer 4 couches (Vs1 < 175 m/s, 175 m/s < Vs2 < 225 m/s, 225 m/s < Vs3 < 400 m/s et Vs4 >.400 m/s) qui présentent des variations d'épaisseur. Des profils de sismiques réflexion en ondes S acquis avec une source construite dans le cadre de ce travail, complètent et corroborent le modèle établi à partir de l'analyse de la dispersion des ondes de surface. Un réflecteur localisé entre 5 et 10 m de profondeur et associé à une vitesse de sommation de 180 m/s souligne notamment la géométrie de l'interface qui sépare la deuxième de la troisième couche du modèle établi à partir de l'analyse de la dispersion des ondes de surface. Abstract Landslides are one of the main natural hazards in mountainous regions. In Switzerland, landslides cause damages every year that impact infrastructures and have important financial costs. In depth understanding of sliding mechanisms may help limiting their impact. In particular, this can be achieved through a better knowledge of the internal structure of the landslide, the determination of its volume and its sliding surface or surfaces In a landslide, the disorganization and the presence of fractures in the displaced material generate a change of the physical parameters and in particular a decrease of the seismic velocities and of the material density. Therefoe, seismic methods are well adapted to the study of landslides. Among seismic methods, surface-wave dispersion analysis is a easy to implement. Through it, shearwave velocity variations with depth can be estimated without having to resort to an S-wave source and to horizontal geophones. Its 3-step implementation implies measurement of surface-wave dispersion with long arrays, determination of the dispersion curves and finally inversion of these curves. Velocity models obtained through this approach are only valid when the investigated medium does not include lateral variations. In practice, this assumption is seldom correct, in particular for landslides in which reshaped layers likely include strong lateral heterogeneities. To assess the possibility of determining dispersion curves from short array lengths we carried out tests measurements on a site (Arnex, VD) that includes a borehole. A 190 m long seismic profile was acquired in a valley carved into limestone and filled with 30 m of glacio-lacustrine sediments. The data acquired along this profile confirmed that the presence of lateral variations under the geophone array influences the dispersion-curve shape so much that it sometimes preventes the dispersion curves determination. Our approach to use the analysis of surface-wave dispersion on sites that include lateral variations consists in obtaining dispersion curves for a series of short length arrays; inverting each so obtained curve and interpolating the different obtained velocity model. The choice of the location as well as the geophone array length is important. It takes into account the location of the heterogeneities that are revealed by the seismic refraction interpretation of the data but also, the location of signal amplitude anomalies observed on maps that represent, for a given frequency, the measured amplitude in the shot position - receiver position domain. The procedure proposed by Lin and Lin (2007) turned out to be an efficient one to determine dispersion curves using short extension arrays. It consists in building a time-offset from an array of geophones with a wide offset range by gathering seismograms acquired with different source-to-receiver offsets. When assembling the different data, a phase correction is applied in order to reduce static phase error induced by lateral variation. To evaluate this correction, we suggest to calculate, for two successive shots, the cross power spectral density of common offset traces. On the Arnex site, 22 curves were determined with 10m in length geophone-arrays. We also took advantage of the borehole to acquire a S-wave vertical seismic profile. The S-wave velocity depth model derived from the vertical seismic profile interpretation is used as prior information in the inversion of the dispersion-curves. Finally a 2D velocity model was established from the analysis of the different dispersion curves. It reveals a 3-layer structure in good agreement with the observed lithologies in the borehole. In it a clay layer with a shear-wave of 175 m/s shear-wave velocity overlies a clayey-sandy till layer at 9 m depth that is characterized down to 14 m by a 300 m/s S-wave velocity; these deposits have a S-wave velocity of 400 m/s between depths of 14 to 20 m. The La Grand Combe landslide (Ballaigues, VD) occurs inside the Quaternary filling of a valley carved into Portlandien limestone. As at the Arnex site, the Quaternary deposits correspond to glaciolacustrine sediments. In the upper part of the landslide, the sliding surface is located at a depth of about 20 m that coincides with the discontinuity between Jurassian till and glacio-lacustrine deposits. At the toe of the landslide, we defined 14 dispersion curves along a 144 m long profile using 10 m long geophone arrays. The obtained curves are discontinuous and defined within a frequency range of 7 to 35 Hz. The use of a wide range of offsets (from 8 to 72 m) enabled us to determine 2 to 4 mode of propagation for each dispersion curve. Taking these higher modes into consideration for dispersion curve inversion allowed us to reach an investigation depth of about 20 m. A four layer 2D model was derived (Vs1< 175 m/s, 175 m/s <Vs2< 225 m/s, 225 m/s < Vs3 < 400 m/s, Vs4> 400 m/s) with variable layer thicknesses. S-wave seismic reflection profiles acquired with a source built as part of this work complete and the velocity model revealed by surface-wave analysis. In particular, reflector at a depth of 5 to 10 m associated with a 180 m/s stacking velocity image the geometry of the discontinuity between the second and third layer of the model derived from the surface-wave dispersion analysis.

Detection of Optimal Models in Parameter Space with Support Vector Machines

The paper proposes an approach aimed at detecting optimal model parameter combinations to achieve the most representative description of uncertainty in the model performance. A classification problem is posed to find the regions of good fitting models according to the values of a cost function. Support Vector Machine (SVM) classification in the parameter space is applied to decide if a forward model simulation is to be computed for a particular generated model. SVM is particularly designed to tackle classification problems in high-dimensional space in a non-parametric and non-linear way. SVM decision boundaries determine the regions that are subject to the largest uncertainty in the cost function classification, and, therefore, provide guidelines for further iterative exploration of the model space. The proposed approach is illustrated by a synthetic example of fluid flow through porous media, which features highly variable response due to the parameter values' combination.

Prime Property Institutions for a Subprime Era: Exploring Innovative Models of Residential Development and Finance

This paper breaks new ground toward contractual and institutional innovation in models of homeownership, equity building, and mortgage enforcement. Inspired by recent developments in the affordable housing sector and in other types of public financing schemes, this paper suggests extending institutional and financial strategies such as timeand place-based division of property rights, conditional subsidies, and credit mediation to alleviate the systemic risks of mortgage foreclosure. Alongside a for-profit shared equity scheme that would be led by local governments, we also outline a private market shared equity model, one of bootstrapping home buying with purchase options.

Ethics Outside, Within, or Beyond OR Models

Since ethical concerns are calling for more attention within Operational Research, we present three approaches to combine Operational Research models with ethics. Our intention is to clarify the trade-offs faced by the OR community, in particular the tension between the scientific legitimacy of OR models (ethics outside OR models) and the integration of ethics within models (ethics within OR models). Presenting and discussing an approach that combines OR models with the process of OR (ethics beyond OR models), we suggest rigorous ways to express the relation between ethics and OR models. As our work is exploratory, we are trying to avoid a dogmatic attitude and call for further research. We argue that there are interesting avenues for research at the theoretical, methodological and applied levels and that the OR community can contribute to an innovative, constructive and responsible social dialogue about its ethics.

Exploring models for the length of stay distribution.

Diagnosis Related Groups (DRG) are frequently used to standardize the comparison of consumption variables, such as length of stay (LOS). In order to be reliable, this comparison must control for the presence of outliers, i.e. values far removed from the pattern set by the majority of the data. Indeed, outliers can distort the usual statistical summaries, such as means and variances. A common practice is to trim LOS values according to various empirical rules, but there is little theoretical support for choosing between alternative procedures. This pilot study explores the possibility of describing LOS distributions with parametric models which provide the necessary framework for the use of robust methods.

Panel index VAR models: Specification, Estimation, Testing and Leading Indicators

This paper proposes a method to conduct inference in panel VAR models with cross unit interdependencies and time variations in the coefficients. The approach can be used to obtain multi-unit forecasts and leading indicators and to conduct policy analysis in a multiunit setups. The framework of analysis is Bayesian and MCMC methods are used to estimate the posterior distribution of the features of interest. The model is reparametrized to resemble an observable index model and specification searches are discussed. As an example, we construct leading indicators for inflation and GDP growth in the Euro area using G-7 information.

Communication, Coordination, and Efficiency in Evolutionary One-population Models

We analyze the role of commitment in pre-play communication for ensuring efficient evolutionarily stable outcomes in coordination games. All players are a priori identical as they are drawn from the same population. In games where efficient outcomes can be reached by players coordinating on the same action we find commitment to be necessary to enforce efficiency. In games where efficienct outcomes only result from play of different actions, communication without commitment is most effective although efficiency can no longer be guaranteed. Only when there are many messages then inefficient outcomes are negligible as their basins of attraction become very small.

Business Cycles, Unemployment Insurance, and the Calibration of Matching Models

This paper points out an empirical puzzle that arises when an RBC economy with a job matching function is used to model unemployment. The standard model can generate sufficiently large cyclical fluctuations in unemployment, or a sufficiently small response of unemployment to labor market policies, but it cannot do both. Variable search and separation, finite UI benefit duration, efficiency wages, and capital all fail to resolve this puzzle. However, both sticky wages and match-specific productivity shocks help the model reproduce the stylized facts: both make the firm's flow of surplus more procyclical, thus making hiring more procyclical too.