Contributions in Merlin

The Internet has enabled a new paradigm of economic production, where individual users per- form work for others, often in small units, for short periods of time, and without formal contracts or monetary payments. These distributed work systems can arise in many places, for example in peer-to-peer (P2P) ¯le-sharing networks, in ad-hoc wireless routing networks, or in 3G content distribution systems. The particular challenge is to incentivize users to perform work for others, even though all interactions are bilateral and monitoring is not possible. In this paper, we formal- ize the problem of designing incentive-compatible work accounting mechanisms that measure the net contributions of users, despite relying on voluntary reports. We describe a fully distributed accounting mechanism called BarterCast and introduce the Drop-Edge extension which re- moves any incentive for a user to make misreports about its own interactions. We prove that the information loss necessary to achieve this incentive compatibility is small and vanishes in the limit as the number of users grows. In some domains, users may be able to cheaply create fake identities (i.e, sybils) and use those to manipulate the accounting mechanism. A striking negative result is that no sybil-proof accounting mechanism exists if one requires responsiveness to a single positive report. To evaluate the welfare properties of BarterCast+DropEdge, we ¯rst present results from a discrete, round-based simulation, showing that the mechanism achieves very high e±ciency. We have also implemented the mechanism in Tribler, a BitTorrent software client, that is al- ready deployed in the real world and has thousands of users. Experimental results using Tribler demonstrate that the mechanism successfully prevents free-riding in P2P-¯le-sharing systems, and achieves better e±ciency than the standard BitTorrent protocol.

We present a new market design for a peer-to-peer (p2p) backup application and provide a theoretical and experimental analysis. In domains such as p2p backup, where many non-expert users find markets unnatural or unexpected, it is often a pragmatic requisite to remove or hide the market’s complexities. To this end, we introduce a new design paradigm which we call “hidden market design,” and show, how a market and its user interface (UI) can be designed to hide the underlying complexities, while maintaining the market’s functionality. We enable the p2p backup market using a virtual currency only, and we develop a novel market UI that makes the interaction for the users as seamless as possible. The UI hides or simplifies many aspects of the market, including combinatorial resource constraints, prices, account balances and payments. In a real p2p backup system, we can expect users to update their settings with a delay upon price changes. Therefore, the market is designed to work well even out of equilibrium, by maximizing the buffer between demand and supply. The main theoretical result is an existence and uniqueness theorem, which also holds if a certain percentage of the user population is price-insensitive or even adversarial. However, we also show that the more freedom we give the users, the less robust the system becomes against adversarial attacks. Furthermore, the buffer size has limited controllability via price changes alone and we show how to address this. We introduce a price update algorithm that uses daily aggregate supply and demand data to move prices towards the equilibrium, and we prove that the algorithm converges quickly towards the equilibrium. Finally, we present results from a formative usability study of the market UI, where we found encouraging results regarding the hidden markets paradigm. The market design presented here is implemented as part of a Microsoft research project and an alpha version of the software has been successfully tested.

Static and dynamic games are important tools for the analysis of strategic interactions among economic agents and have found many applications in economics. In many games equilibria can be described as solutions of polynomial equations. In this paper we describe state-of-the-art techniques for finding all solutions of polynomial systems of equations and illustrate these techniques by computing all equilibria of both static and dynamic games with continuous strategies. We compute the equilibrium manifold for a Bertrand pricing game in which the number of equilibria changes with the market size. Moreover, we apply these techniques to two stochastic dynamic games of industry competition and check for equilibrium uniqueness.

This paper analyzes optimal sampling and pricing of paid content for publishers of news websites. Publishers offer free content samples both to disclose journalistic quality to consumers and to generate online advertising revenues. We examine sampling where the publisher sets the number of free sample articles and consumers select the articles of their choice. Consumerslearn from the free samples in a Bayesian fashion and base their subscription decisions on posterior quality expectations. We show thatsampling enhances subscription demand only if consumers have low quality expectations in relation to actual quality. Taking advertising and subscription revenues into account, we find that the publisher should employ either a paid-only or a free content strategy when consumers have high quality expectations. When consumers have low quality expectations, employing a sampling strategy may be optimal for the publisher.

We consider a simple CAPM with heterogenous expectations on assets mean returns while keeping the assumption of homogenous expectations on the covariance of returns. Our first result derives the security market line as an aggregation result without using the two-fund-separation property. In particular every investor can hold optimal portfolios that are underdiversified.In our model alpha-opportunities can be explained as a feature of financial market equilibria and we can show that alpha-opportunities erode with the assets under management and that the hunt for alphaopportunities is a zero-sum game. Then we endogenize the agents information by allowing them to be either passive, in which case they invest according to the average expectation embodied in the market returns, or to be active, in which case they can acquire superior information at some cost. It is shown that expecting a positive alpha is not necessarily a good criterion for becoming active. Moreover, the less risk averse investors are more inclined to be active and delegating active investment to portfolio managers only makes sense if the performance fee increases with the skill of the portfolio manager. Finally, in our model it turns out that only a market in which all investors are passive and share the same correct belief is stable with respect to information acquisition. Hence the standard CAPM with homogenous beliefs can be seen as the long run outcome of our model.

Structured financial products have gained more and more popularity in recent years, but nevertheless has their success so far not thoroughly been analyzed. In this article we develop a theoretical framework for the design of optimal structured products and analyze the maximal utility gain for an investor that can be achieved by introducing structured products. We demonstrate that most successful structured products are not optimal for a perfectly rational investor and investigate the reasons that make them nevertheless look so attractive for many investors

We present results from the first large-scale international survey on time discounting, conducted in 45 countries. Cross-country variation cannot simply be explained by economic variables such as interest rates or inflation. In particular, we find strong evidence for cultural differences, as measured by the Hofstede cultural dimensions. For example, high levels of Uncertainty Avoidance or Individualism are both associated with strong hyperbolic discounting. Moreover, as application of our data, we find evidence for an impact of time preferences on the capability of technological innovations in a country and on environmental protection.

n many countries structured investment products are popular among retail investors. We explain the demand for these products using unique field data where we let subjects freely design their "favorite" structured product. Results suggest that the supply with capital protected products (guarantee certificates) might indeed be demand-driven. This does not seem to be the case for other product categories where marketing and sales practices might play a more important role. In a survey among financial practitioners we find furthermore that a demand for capital protected products can be explained by loss aversion and saving motifs, e.g. for buying a house.

null

Contrary to the common wisdom that asset prices are hardly possible to forecast, we show that high and low prices of equity shares are largely predictable. We propose to model them using a simple implementation of a fractional vector autoregressive model with error correction (FVECM). This model captures two fundamental patterns of high and low prices: their cointegrating relationship and the long memory of their difference (i.e. the range), which is a measure of realized volatility. Investment strategies based on FVECM predictions of high/low US equity prices