Site Network:

ICT4Law Regione Piemonte

WP09 Controbutions from DipECO, Dep. of Economics and Public finance (former name)

WP09 - DipEco, Dipartimento di Scienze economico-sociali e matematico-statistiche (Department of economics and statistics), formerly Dipartimento di Scienze economiche e finanziarie G.Prato (Department of Economics and Public Finance)

Goals and results

The stating point was that to investigate complex systems dynamic, as in the case of the interactions between law systems and economic sectors, the use of traditional mathematical models had to be improved, introducing the new tools of the simulation applied to agent based systems.
The shell that we have now built (current version of SLAPP, see references below), can indeed integrate behavioral evaluations, actual data, rule influence, general knowledge of the economic and social environment in a unique framework, to forecast dynamic changes, also in a “what” if perspective.
The starting point has been the knowledge on enterprise and organization simulation, wher we hade developed a tools such as jES (the Java Enterprise Simulator), introduced at http://web.econ.unito.it/terna/jes
Starting from jES, we have integrated cognitive and social components in the model to obtain a complete simulator, improving tools such as SLAPP (Swarm-Like Agent Protocol in Python) where Swarm is at http://www.swarm.org, SLAPP at http://eco83.econ.unito.it/terna/slapp and Python at http://www.python.org.
The simulator, formerly jES and now AESOP (Agents and Emergencies for Simulating Organizations in Python) uses action, actors (acting and deciding peoples) and scheduling of events into an agent based framework. Actors can use fixed rules or can learn to improve rules in a changing environment and finally can be aware of the consequences of their behavior.
The kernel of the simulator is the capability of reproducing in a detailed way the decision making processes of collectivity of agents, but always on an individual basis. We have both no minded agents - as things to be done (orders) or units able to work with them - and minded - as the agents who have to express decisions within the model.
Simulating a single enterprise or organization or a system of enterprises or organizations (i.e. within a district or within a virtual enterprise system or in a regional or wider system) we can use in a direct way the “what if” analysis introducing changes into the simulation, while fully preserving the complexity of our context. Only in a true agent based context, with independent pieces of software expressing the different behavior of all the components of our environment, we can obtain unlimited granularity in the description of the investigated world. We must also interact in a correct way with actual enterprise’s data and for that we have to develop easy to use interfaces, i.e. based on the spreadsheet formalism.
The development of this tool is now satisficing, but still with the necessity of a complete revision to search for part that can be managed in a easier way and closer to jES initial philosophy, but now simplified for the user. On this task is working Gian Paolo Jesi, in cooperation with the Department of Management Science of the University of Bologna.

Within the WP09 and in parallel with the creation of the shell, several research activity are under development or concluded.

1. ICT4LAW member: Matteo Morini, work in progress
Rules affecting strategic behavior; Reflections on Complexity, or How Spatial Interactions in Evolutionary Games Matter.
In the previous work by Beraldo and Morini (see below), cooperative behaviors were envisioned to emerge from interacting agents without prior knowledge of the world. Rules were supposed to be learnt the hard way, by repeated encounters in an extended prisoners' dilemma game in which the lower payoffs led the worst performers to extinction. Defections (eventually, justified defections) were supposed to be observed and sanctioned by peers. In an intentionally oversimplified setting (randomly matched agents in an adimensional space), agents were endowed with a sophisticated learning machinery, CT-ANN (Terna 2000), linked to the simulator described above and improved within the project.
This side-project stems from a reflection on the relevance of the spatial dimension of the model, as put forward in Nowak works on spatial games, Novak et al. (1994) in particular.
The idea is to set aside for a moment the complexity inherent in sophisticated agents, while concentrating on the environmental side of the model, the dynamics of interaction in particular. Which allows to neglect unrealistic assumptions on the elaborating capabilities of agents which can't obviously keep in the simplest incarnations (bacteria, unicellular organisms, which, albeit exhibiting cooperative behaviors, needn't developing strategies, keeping memory of past encounters, or discounting future payoffs).
Departing from randomly-matching pairs of agents, we try to compare different spatial configurations (topologies) and introducing spatial irregularities, and measuring whether continuous versus discrete time can make a difference.
It appears that agents interacting on square lattices (with all the 8 neighbors or the 4 orthogonal ones only), with finite boundaries or toroidal closed spaces, hexagonal lattices, irregular spaces result in different outcomes.
Nowak M.A., Bonhoeffer S., May R.M. (1994), ”Spatial Games and the Maintenance of Cooperation”, Proc. Natl. Acad. Sci. USA, Vol. 91
Terna P. (2000), Economic Experiments in Swarm: a Neural Network Approach to the Self-Developement of Consistency in Agents’ Behaviour, in Economic Simulations in Swarm: Agent-Based Modelling and Object-Oriented Programming, Luna and Stefansson eds., Kluver Academic, Dordrecht

2. ICT4LAW member: Matteo Morini, work in progress
Sergio Beraldo* and Matteo Morini§
SwarmFest 2011 - The Emergence and Inevitability (or not) of Cooperation.
Indirect reciprocity is generally considered one of the leading mechanisms to explain how cooperation may emerge by natural selection. The basic intuition is that establishing a reputation of being a helpful individual increases the probability of being in turn helped.
Two models have been proposed to describe how indirect reciprocity may work: the standing model (Sugden 1986/2004) and the image-scoring model (Nowak and Sigmund 1998a and 1998b). Although there is evidence that the former model would perform better under a wide set of circumstances, it is often maintained that it requires individuals with an implausibly large capacity of processing recursive information. In Beraldo (2011) both models are put under scrutiny and shown to be ultimately equivalent (the recursive information process requirement is shown not to hold), but for the fact that the identity of the receiver remains unknown in the image-scoring model.
We’d love to see whether any of the two mechanisms (or both) are inevitable and instrumental in putting indirect reciprocity at work, and in this vein, we try to make cooperation emerge spontaneously, starting with a constant and finite population of agents, and endowing them with the least possible predefined knowledge: zero. This leads to bootstrapping our model with ‘dumb’ yet ‘adaptive’ agents, in the sense proposed in Terna (2000). Each artificial individual “must develop and adapt her capability of evaluating, in a coherent way, (1) what she has to do in order to obtain a specific result and (2) how to foresee the consequences of her actions” (Beltratti et al., 1996), survival being their specific result, by means of the cross-target method applied to neural networks.
The expected outcome is an evolving population of individuals (the survival of the fittest applies) adopting the most appropriate set of rules: ‘blind’ cooperation, cooperation with agents in ‘good standing’ and defection with the others, or no cooperation at all.
Mistakes in gathering and elaborating informations, and in acting, resulting in accidental defections and collaborations, have drawn much interest (prominently, Wu and Axelrod 1995); they are supposed to emerge in our model (and perhaps be corrected, in the long run) spontaneously. The dynamics of this and other processes will be very interesting to observe and stylize.
Beltratti, Margarita and Terna 1996, Neural Networks for Economic and Financial Modelling, International Thompson Computer Press, London
Beraldo 2011, Good Standing and Cooperation, International Review of Economics, 58:79-90
Nowak and Sigmund 1998a, The evolution of indirect reciprocity by image scoring. Nature 393:573-577
Nowak and Sigmund 1998b, The dynamics of indirect reciprocity. Journal of Theoretical Biology 194:561-574
Sugden 1986, The Economics of Rights, Cooperation and Welfare, Blackwell, Oxford. (Second edition 2004, Palgrave-Macmillan, Basingstoke)
Terna 2000, Economic Experiments in Swarm: a Neural Network Approach to the Self-Developement of Consistency in Agents’ Behaviour, in Economic Simulations in Swarm: Agent-Based Modelling and Object- Oriented Programming, Luna and Stefansson eds., Kluver Academic, Dordrecht
Wu and Axelrod 1995, How to Cope with Noise in the Iterated Prisoner’s Dilemma, Journal of Conflict Resolution, 39:183-189

3. ICT4LAW member: Michele Sonnessa, paper in publication, available on request
Autonomic Logistics Systems as system of rules: an economic perspective.
Autonomic logistics systems (ALS) are supply chain models mostly adopted in air force military systems. They are mainly characterized by special military technologies and equipments enabling planes to self forecast failures and communicate them before landing. By forecasting failures the maintenance can be set-up in short time, in respect to the traditional approach. If this technology is not available a failure can only be discovered and notified during the de- briefing activities in hangars. The preemptive capabilities impact on the logistics organization as well, with a better management, reduction of scheduled maintenance and better organization of skilled personnel activities.
A novel organizational approach has been introduced as a mix of technological and organizational logistic systems. The so called Autonomic Logistic System (ALS) is considered the most promising approach to improve military logistic systems supporting aircraft maintenance and failure fixing.
ALS can be designed as a computer network, through which some sensor can broadcast events. According to those events the logistics infrastructure can react and setup maintenance teams in time, providing the right person, the right tool and the right replacement kit. This information should provide greater insight into the cause-and-effect relationship between resources and readiness. Such coordination among people is mainly delegated to the computer network, reducing errors of coordination.

4. ICT4LAW member: Michele Sonnessa, work in progress
Data management to analyze the effect of hypotheses in simulation models: Model-Data-Observer for ICT4LAW.
Agent based modeling applied to policy making implies high flexibility in data exploration. Often research questions are improved or modified by the observation of simulation experiment outcomes. Thus, we investigate new approaches in collection and storage of data generated by agent based models. A new paradigm called Model-Data-Observer (MDO) is introduced. It benefits from the NoSQL emerging storage technologies, which well fit when data structures are not a priori defined or are characterized by high rate of change. MDO allow scientists to collect all information produced by a simulation experiment without requiring a definition of the analysis strategy. Data and phenomena investigation is, in fact, can be carried on after one or more experiment sessions are completed. Then, we describe an example of MDO implementation applied to the 'TrustMe' model, a simulation of competitive mechanisms investigating if trust relationship can spontaneously emerge in a simple styilized market. MDO provides different general improvement in agent based model design: a standard approach to collect data from an ABM; the off-line model outcomes analysis; a strong separation between model and presentation; run-to-run overall comparison.

5. ICT4LAW member: Lia Pacelli, paper in publication, Marco Malgarini, Massimo Mancini and Lia Pacelli, Temporary hires and innovative investments
Applied Economics, 2012, ??, 1–10 (on line in March)
Rules, Labor And Firm Performances.
The flexicurity approach claims a positive effect of flexible labour on firm performance, also through an increased ability to innovate. Critics consider it a deregulation of the labour market, decreasing investment in human capital and innovation. We contribute to this broad debate providing an estimate of the relationships linking innovative investment, substitution investment, permanent hires and temporary hires. In particular, we aim at affirming or denying that innovative investments are accompanied by a specific kind of workforce, being it stable or flexible. In doing so, we contribute to bridge the gap among two quite separate strands of literature, as existing literature usually analyses capital and labour separately.
Estimating a non linear recursive equation system we highlight a significant increase in the likelihood of hiring on a permanent base when the firm innovates; this holds till 2008. Afterward, during the crisis, innovating firms are more likely to hire using temporary contracts instead, a possible signal of a cost saving strategy adopted in a loose labour market by firms still able to innovate. Furthermore, both permanent and temporary hires never depend on increases in labour costs; however, substitution investment increases when labour cost increases, maybe in an attempt to increase labour productivity through a more efficient capital equipment.

6. ICT4LAW member: Lia Pacelli, work in progress
Employment protection and training: law and custom at the firm level.
Lia Pacelli, Giulia Maria Cavaletto
The economic pressure of globalization and the cultural pressure of liberism led most industrialized countries over the last two decades to deregulate the labor market, under the stated assumption that it was rigid and that it was responsible for high unemployment rates and low firm performance. This path generated segmented labor markets and decreased incentives to invest in workers’ human capital, making cost reduction strategies more appealing. Italy is a paradigmatic example of this evolution, coupling high formal protection of incumbent to null protection of entrants, high perceived employment rigidity to huge workers’ flows, high and increasing profit shares to low demand for graduate and skilled workers.
We interviewed human resources managers and entrepreneurs, labor consultants, workers’ and firms’ unions, lawyers and a judge of the labor court. We have analyzed the data both in the economic perspective of adjustment costs and with the sociological method of distributional statistics of lexical co-occurrences. First, we could compute directly the cost faced by a firm when hiring and training workers and when dismissing them: e.g. we compute a maximum of dismissal costs, in case of firing without just cause, of about 23 monthly wages, one third going to third parties. Second, we highlight customary patterns put in place at the firm level to manage the workforce within the legal setup: the common strategy pursued by the interviewed firms might be summarized in: “maximum reliability required to the worker, minimum commitment on the side of the firm, unstructured on-the-job training”.

7. ICT4LAW member: Riccardo Boero, work in progress, introductory paper available on request
Public policy, rules and local dynamics.
A paper titled “Local Economic Dynamics and Public Policy Design: an Input-output and Agent-based Approach” is in preparation. So far, it has been presented in invited talks at the Decision Applications Division of Los Alamos National Laboratory (USA) and at the Innovation Research Unit of University College Dublin (Ireland), and some of its conclusions have been presented in an article titled “Più politiche di sistema” appeared on the Italian business daily “Il Sole 24 Ore Nord Ovest”. The main accomplishment of this research is to provide policy makers an agent-based model to evaluate different configurations of government spending. In particular, the research points out how much incoordination and lack of shared strategies negatively affect policies and new laws employed by local governments.

8. ICT4LAW member: Riccardo Boero, paper forthcoming on Social Networks, available on request
Relationships and networks with laws and social norms.
The paper “Trust and Partner Selection in Social Networks: An Experimentally Grounded Model”, forthcoming on Social Networks, contributes to the debate on the design of cooperation supporting institutions such as laws and social norms. The experimentally grounded agent-based model presented in the paper is a software platform where different configurations of norms can be explored aiding law makers to improve the prediction of the consequences of new laws.

9. ICT4LAW member: Gilberto Turati Riccardo Boero, paper Sergio Beraldo & Stefania Ottone & Gilberto Turati, 2011. "Are genetically modified foods bad for my health?" Individuals' valuation and the choice among different information sources ICER Working Papers 05-2011, ICER - International Centre for Economic Research.
Investigation of information effect on consumer behavior.
We investigate the role of information on consumers’ valuation for food products containing genetically modified organisms (GMOs), using data from a specifically designed survey. We provide three main results. First, we show that introducing mandatory labels to identify whether or not a food product contains GMOs, significantly reduces consumers’ valuation. Second, adding to the label additional information on GMOs significantly affects valuation. Third, no matter the sign of the information previously received, consumers are more willing to trust General Practitioners (GPs), the information source they prefer most. Overall, these results indicate that the crucial issue is not the presence of the label per se, but the availability of the necessary information to make good use of the label content to assess potential health risks deriving from GM foods. In particular, our findings suggest that this can be achieved by properly informing (and convincing) GPs and other health professionals that risks for human health are minimal.

10. ICT4LAW member: Irakli Gogatishvili, text available on request
Stock market rules and agents’ reactions.
Irakli Gogatishvili thesis, used to create a basic framework to explore the effect of new rules into the stock market compared with actual people actions, reactions, adaptations.
The work presents an agent-based framework for the multiple learning in the con- tinuous double-auction (ABCDA) financial market. On the first stage the two types of trading agents are presented: (1) Zero-Intelligence (i.e. absolutely random and without learning ability) and (2) Zero-Intelligence Plus (ZIP) agents, with profit margin updating. Different scenarios are simulated, with different compositions of traders. Initially they are endowed with certain amounts of assets and money, which are randomly distributed among them. In the course of trade the complex dynamics is induced and important stylized facts are observed, that are persistent volatilities, fat tails in return distribution and volatility clustering and. The script for simulation is developed in Python using SLAPP (Swarm-Like Agent Protocol in Python). For data analysis and visualization we chose R libraries (through RPy2, i.e. Python interface to the R Programming Language). Next step is made by introducing of Markovian dynamics to the market. ZIP Agents are endowed with the q-learning ability via artificial neural network. The architecture and pseudocodes describing the behavior are presented. On the final stage the human-like approximate reasoning mechanism is introduced. The neuro-fuzzy controller mechanism is used to model fuzzy q-learning and corresponding pseudocodes are incorporated.

PREVIOUS WORKS, in last year report

11. ICT4LAW members: Massimiliano Piacenza e Gilberto Turati, paper Does Fiscal Discipline towards Sub-national Governments Affect Citizens’ Well-being? Evidence on Health, available on request
Second order fiscal rules and citizens’ behavior.
This paper aims at assessing the impact on citizens’ well-being of fiscal discipline imposed by Central Government to sub-national governments. Since health care policies involve strategic interactions between different layers of governments in many different countries, we focus on a particular dimension of well-being, namely citizens’ health. We model fiscal discipline by considering sub-national governments expectations of future deficit bailouts from the Central Government. We then study how these bailout expectations affect the expenditure for health care policies carried out by decentralized governments. To investigate this issue, we separate efficient health spending from inefficiencies by estimating an input requirement frontier. This allow us to assess the effects of bailout expectations on both the structural component of health expenditure and its deviations from the ‘best practice’. The evidence from the 15 Italian Ordinary Statute Regions (observed from 1993 to 2006) points out that bailout expectations do not significantly influence the position of the frontier, thus do not affect citizens’ health. However, they appear to exert a remarkable impact on excess spending.

12. ICT4LAW members: Massimiliano Piacenza e Gilberto Turati, paper Decentralization and Local Governments’ Performance: How Does Fiscal Autonomy Affect Spending Efficiency? (Authors M.Piacenza, G.Turati and L.Boetti), forthcoming on FinanzArchiv/Public Finance Analysis, available on request
Fiscal rules and local governments.
In Italy, as in other countries around the world, recent reforms share the goal of increasing the fiscal autonomy of lower tiers of governments, from Regions to Municipalities, in order to align spending with funding responsibilities and increase the efficiency in the provision of essential public services. The purpose of this paper is to assess spending efficiency of local governments and to investigate the effects of tax decentralization, focusing on the role played by incumbent politicians’ accountability. The analysis relies on a sample of Italian municipalities and exploits both parametric (SFA) and nonparametric (DEA) techniques to study spending inefficiency and its main determinants. Consistently with modern fiscal federalism theories, our results show that more fiscally autonomous municipalities exhibit less inefficient behaviours. We also find that the shorter is the distance from new elections, the higher is excess spending, thus giving further support to the traditional “electoral budget cycle” argument. Other political features of governing coalition, such as age and gender of the mayor, do not seem to exert any significant impact on inefficiency levels.

13. ICT4LAW member: Concetta Sorropago, text available on request
How evaluation rules affect management: Computing Congruence of Performance Evaluation Measures: A Simulation Approach.
Distortion and noisiness are two key parameters to verify the congruence of the performance measures adopted in an incentive schema to avoid their gaming and manipulation. Even if we have had many studies on this topic the theory is not able to explain/predict the real world companies’ behavior . The data necessary to study this phenomenon, are generally internal data that can be obtained only through a specific case-study. In this paper, we use a simulation approach to generate some of these firms’ data to analyze these issues. we show that building a toy model of a company that produces project turn-key allows us experimentally to compute by using a simulation approach the distortion and the noisiness of a set of performance measures adoptable for an incentive schema. This type of firm had also to cope with the dynamical scheduling of resources on multiple projects. This problem, which does not have closed mathematical solutions, is worked out by a binary maximization algorithm. Eventually, pro and cons of this approach are analyzed and further research development is suggested.

13. ICT4LAW member: Davide Vannoni, paper Increasing market interconnection: An analysis of the Italian electricity spot marketFederico Boffa, Viswanath Pingali, Davide Vannoni, International Journal of Industrial Organization, 28, 2010, 311-322
Market and network rules and economic emergences.
We estimate the benefits (in terms of savings to end-users) resulting from an improved interconnectivity in the Italian electricity spot market. The market is currently divided into two geographic zones – North and South – with limited inter-zonal transmission capacity that often induces congestion, and hence potential inefficiency. By simulating a fully interconnected market, we predict that the total spot market expenditure would reduce substantially. Moreover, since savings do not increase linearly with the size of new transmission capacity, even a slight increment to transmission capacity is found to substantially reduce end- users' expenditures. Finally, our analysis shows that the (partly State owned) dominant firm in the market is not maximizing short-term profits.