The marginal benefit of diversification in commercial real estate portfolios

The number of properties to hold to achieve a well-diversified real estate property portfolio presents a puzzle, as the estimated number is considerably higher than that seen in actual portfolios. However, Statman (1987) argues that investors should only increase the number of holdings as long as the marginal benefits of diversification exceed their costs. Using this idea we find that the marginal benefits of diversification in real estate portfolios are so small that investors are probably rational in holding small portfolios, at least as far as the reduction in standard deviation is concerned.

Location and collocation advantages in international innovation

Purpose – This paper examines the role of location-specific (L) advantages in the spatial distribution of multinational enterprise (MNE) R&D activity. The meaning of L advantages is revisited. In addition to L advantages that are industry-specific, the paper emphasises that there is an important category of L advantages, referred to as collocation advantages. Design/methodology/approach – Using the OLI framework, this paper highlights that the innovation activities of MNEs are about interaction of these variables, and the essential process of internalising L advantages to enhance and create firm-specific advantages. Findings – Collocation advantages derive from spatial proximity to specific unaffiliated firms, which may be suppliers, competitors, or customers. It is also argued that L advantages are not always public goods, because they may not be available to all firms at a similar or marginal cost. These costs are associated with access and internalisation of L advantages, and – especially in the case of R&D – are attendant with the complexities of embeddedness. Originality/value – The centralisation/decentralisation, spatial separation/collocation debates in R&D location have been mistakenly viewed as a paradox facing firms, instead of as a trade-off that firms must make.

The distribution of London residential property prices and the role of spatial lock-in

Much of mainstream economic analysis assumes that markets adjust smoothly, through prices, to changes in economic conditions. However, this is not necessarily the case for local housing markets, whose spatial structures may exhibit persistence, so that conditions may not be those most suited to the requirements of modern-day living. Persistence can arise from the existence of transaction costs. The paper tests the proposition that housing markets in Inner London exhibit a degree of path dependence, through the construction of a three-equation model, and examines the impact of variables constructed for the 19th and early 20th centuries on modern house prices. These include 19th-century social structures, slum clearance programmes and the 1908 underground network. Each is found to be significant. The tests require the construction of novel historical datasets, which are also described in the paper.

Non-uniform data distribution for communication-efficient parallel clustering

Global communicationrequirements andloadimbalanceof someparalleldataminingalgorithms arethe major obstacles to exploitthe computational power of large-scale systems. This work investigates how non-uniform data distributions can be exploited to remove the global communication requirement and to reduce the communication costin parallel data mining algorithms and, in particular, in the k-means algorithm for cluster analysis. In the straightforward parallel formulation of the k-means algorithm, data and computation loads are uniformly distributed over the processing nodes. This approach has excellent load balancing characteristics that may suggest it could scale up to large and extreme-scale parallel computing systems. However, at each iteration step the algorithm requires a global reduction operationwhichhinders thescalabilityoftheapproach.Thisworkstudiesadifferentparallelformulation of the algorithm where the requirement of global communication is removed, while maintaining the same deterministic nature ofthe centralised algorithm. The proposed approach exploits a non-uniform data distribution which can be either found in real-world distributed applications or can be induced by means ofmulti-dimensional binary searchtrees. The approachcanalso be extended to accommodate an approximation error which allows a further reduction ofthe communication costs. The effectiveness of the exact and approximate methods has been tested in a parallel computing system with 64 processors and in simulations with 1024 processing element