Search Results (Subtype:"Working paper", isMemberOf:"UNU:985", Display Type:"Report", Author ID:"Verspagen, Bart", Author:"Silverberg, Gerald") - UNU Collections
http://collections.unu.edu/
United Nations UniversityenFez 2.1 RC3http://blogs.law.harvard.edu/tech/rssA Note on Michelacci and Zaffaroni, Long Memory, and Time Series of Economic Growth
http://collections.unu.edu/view/UNU:1081
In a recent paper in The Journal of Monetary Economics, Michelacci and Zaffaroni (2000)estimate long memory parameters for GDP per capita of 16 OECD countries. In this note weargue that these estimations are questionable for the purposes of clarifying the time seriesproperties of these data (presence of unit roots, mean reversion, long memory) because theauthors a) filter out a deterministic linear-in-logs trend instead of first-differencing in logs,and manipulate the data in other highly questionable ways, b) rely on the semiparametricGeweke and Porter-Hudak (GPH) method as modified by Robinson, which is known to behighly biased in small samples. We re-examine these results using Beran's nonparametricFGN estimator and Sowell's exact maximum likelihood ARFIMA estimator. These methodsavoid the small-sample bias and arbitrariness of the cut-off parameters of Robinson's methodand allow us to control for short memory effects, although the parametric ARFIMA estimatorintroduces specification problems of its own. We also look at the influence of the choice ofsub-periods on the results. Finally, we apply Robinson's method to our treatment of the dataand show that MZ's results no longer hold, nor are their cut-off parameter and filteringinsensitivity claims substantiated.2013-12-13T13:01:02Z
Silverberg, Gerald
og Verspagen, Bart
Self-organization of R&D search in complex technology spaces
http://collections.unu.edu/view/UNU:1188
We extend an earlier model of innovation dynamics based on invasive percolation by adding endogenous R&D search by economically motivated firms. The {0,1} seeding of the technol-ogy lattice is now replaced by draws from a lognormal distribution for technology 'difficulty'. Firms are rewarded for successful innovations by increases in their R&D budget. We compare two regimes. In the first, firms are fixed in a region of technology space. In the second, they can change their location by myopically comparing progress in their local neighborhoods and probabilistically moving to the region with the highest recent progress. We call this the mov-ing or self-organizational regime. We find that as the mean and standard deviation of the log-normal distribution are varied, the relative rates of aggregate innovation switches between the two regimes. The SO regime has higher innovation rates, other things being equal, for lower means or higher standard deviations of the lognormal distribution. This results holds for in-creasing size of the search radius. The clustering of firms in the SO regime grows rapidly and fluctuates in a complex way around a high value which increases with the search radius. We also investigate the size distributions of the innovations generated in each regime. In the fixed one, the distribution is approximately lognormal and certainly not fat tailed. In the SO regime, the distributions are radically different. They are much more highly right skewed and show scaling over at least two decades with a slope of almost exactly one, independently of parame-ter settings. Thus we argue that firm self-organization leads to self-organized criticality.2013-12-13T12:39:11Z
Silverberg, Gerald
og Verspagen, Bart
Long Memory in Time Series of Economic Growth and Convergence
http://collections.unu.edu/view/UNU:1052
2013-12-13T13:02:03Z
Silverberg, Gerald
og Verspagen, Bart
The size distribution of innovations revisited: an application of extreme value statistics to citation and value measures of patent significance
http://collections.unu.edu/view/UNU:1170
This paper focuses on the analysis of size distributions of innovations, which are known to be highly skewed. We use patent citations as one indicator of innovation significance, constructing two large datasets from the European and US Patent Offices at a high level of aggregation, and the Trajtenberg (1990) dataset on CT scanners at a very low one. We also study self-assessed reports of patented innovation values using two very recent patent valuation datasets from the Netherlands and the UK, as well as a small dataset of patent license revenues of Harvard University. Statistical methods are applied to analyse the properties of the empirical size distributions, where we put special emphasis on testing for the existence of 'heavy tails', i.e., whether or not the probability of very large innovations declines more slowly than exponentially. While overall the distributions appear to resemble a lognormal, we argue that the tails are indeed fat. We invoke some recent results from extreme value statistics and apply the Hill (1975) estimator with data-driven cut-offs to determine the tail index for the right tails of all datasets except the NL and UK patent valuations. On these latter datasets we use a maximum likelihood estimator for grouped data to estimate the Pareto exponent for varying definitions of the right tail. We find significantly and consistently lower tail estimates for the returns data than the citation data (around 0.7 vs. 3-5). The EPO and US patent citation tail indices are roughly constant over time (although the US one does grow somewhat in the last periods) but the latter estimates are significantly lower than the former. The heaviness of the tails, particularly as measured by financial indices, we argue, has significant implications for technology policy and growth theory, since the second and possibly even the first moments of these distributions may not exist.2013-12-13T12:39:57Z
Silverberg, Gerald
og Verspagen, Bart