Markov's Frog

Schroedinger's cat's next state was indeterminate. Pavlov's dog's next state was increasingly predictable. Markov's frog's next state depends only on his current state: and so, as a French philosopher of the XXth century pointed out, the question is "dans quel état j'erre?".

Location: Alsace, France

Friday, March 13, 2015

First-linkage of U.S. Census Data Households

A VoxEU note recently presented an apparently novel use of U.S. Census data: to measure racial segregation in American neighborhoods. Simply put, if almost all households of a particular declared race are adjacent to another household of the same race, there is clumping or clustering, hence segregation.

Whence the Claim : "Federal census enumerators would go door-to-door when enumerating the residents of a community. Consequently, next-door neighbours appear adjacent to one another on the census manuscript pages. With the availability of complete census manuscripts for 1880 and 1940, this allows us to identify the races of all household heads and of their next-door neighbours. "

There are some possible problems with this, but I don't know that they invalidate the conclusions one would (and the authors do) draw from using the measure as proposed. I have browsed census data from 1940, and a few others, in the course of genealogical curiosity.  Much of the data I've seen was already transcribed, the 1940 data was not.

  1. The 1940 data in the neighborhood I studied did not have all the households reported consecutively with their physically adjacent neighbors;  when nobody was home, there were gaps filled later on other sheets.
    • it may be that either the transcribers have re-sorted the data to put households back in physical adjacency order, or that it does not really matter (it might not, but I haven't thought it through so cannot say).  
    • Would the phenomenon have changed in its frequency or handling (by the census takers and processors) between the two censuses in a way that distorts the apparent trend? It might have; urbanisation had increased, perhaps household size had decreased, the latter of which might increase the number of first-pass gaps (nobody home when the enumerator knocked on the door). If list adjacency does not reliably indicate residential adjacency, the statistic is not what it is meant to be.
    • Does it matter whether the enumerators' routes crossed the street or turned the corner? It might, depending on the street to cross. A household on a corner would, I believe, feel the house around the corner to be closer than the house across the street, particularly if the street is one that serves as a sort of demarcation of neighborhood boundaries: the street on which side one resides matters.
  2. Household and dwelling place are not the same thing.  Many households are in multiple-household dwelling places, such as apartment buildings or  boarding houses (not only for singles), or residential hotels.  How then to meaningfully apply the concept of next-door neighbor?
    • I do not know how they handled this, I've only read the VoxEU piece, not the paper it summarizes and pitches. Perhaps I shall.
    • Given the intended use, I would expect that dwelling in a place where any of the other heads of household "under the same roof" is of a different declared race, whether or not they live on the same floor (who know? not the census) counts as having a different-race neighbor.
    • It is not so easy to decree that someone living in a multiple household dwelling place that is mono-racial (according to declarations of heads of household) next door to one that is mono-racial of a different race has a next-door neighbor of a different race.
I think the proposed statistic is interesting, and possibly useful.  But it needs some vetting.

Tuesday, December 16, 2014

Season's Attitudes

An enterprise called Statista has published a puzzling poll result concerning attitudes to Christmas in Northern Europe (and the UK) in the end of 2014, based on data gathered by YouGov.
Infographic: One European Country Is Not Excited About Christmas | Statista
You will find more statistics at Statista

This was brought to my attention by the Independent UK piece "France is the northern European country that cares the least about Christmas.

Monday, February 13, 2006



Thank you, Amazon! Searching for the book "Search Methodologies", I came across the book "The Craft of Probabilistic Modelling", which looks pretty interesting. However, this was on Amazon DE and I had trouble understanding; I moved to Amazon UK, hoping to find more about the content. Lo and behold, not only did I find that book, I found the proceeding from the "Process Algebra and Probabilistic Methods: Performance Modeling and Verification" conferences held in 2001 and 2002. Why had I not heard of this conference before? Will it be held again?

Searching Google with "PAPM-PROBMIV" (a rather unlikely string of characters in any other context, I expect) turned up 9660 matches. Unfortunately, this seems to be one of those conferences without a permanent institution or society, the type whose web site is re-invented each year by the organizing-hosting-venue owner: very tedious to find all the years.

2001 : First edition, Aachen, Germany, "Co-located with MMB and PNPM"
2002 : Second edition, associated with FLoC '02, affiliated with CAV 2002
2003 :
2004 : QEST 2004, Enschede, The Netherlands
2005 : joined International Conference on Modeling Techniques and Tools for Computer Performance Evaluation (TOOLS) and the International Workshop on Petri Nets and Performance Models (PNPM) to form QEST'05 - The International Conference on Quantitative Evaluation of SysTems
2006 : QEST 2006 , Riverside, California, September 11-14, 2006

The book that caught my eye, "The Craft of Probabilistic Modelling", is currently unavailable (dixit Springer-Verlag); Amazon lists a few used book sellers. I cannot find the table of contents. The German Amazon site has more info than the UK one, it lists the authors of the chapters (it is "a collection of personal accounts"), so I'll see what I can find about their specialities:
  • N. T. Bailey (The Mathematical Theory of Infectious Diseases and its Applications. Hafner Press, New York, 1975.)
  • J. W. Cohen (Extreme value distribution for the M/G/1 and the G/M/1 queueing systems. Annales de l'institut Henri Poincaré (B) Probabilités et Statistiques),(Random Walk with a Heavy-Tailed Jump Distribution. Queueing Syst. 40(1): 35-73 (2002))
  • W. J. Ewens (Aspects of Optimality Behavior in Population Genetics Theory. Evolution and Biocomputation 1995: 7-17)
  • E. J. Hannan (Applications of Time Series Analysis in Astronomy and Meteorology,Chapman & Hall, March 1997)
  • M. Iosifesu (Dependence with complete connections and its applications, FINITE MARKOV PROCESSES AND THEIR APPLICATIONS, RANDOM PROCESSES AND LEARNING)
  • J. Keilson (A note on the summability of the entropy series. Information and Control, 18(3):257-260, April 1971. Multi-Server Threshold Queues with Hysteresis. Perform. Eval. 21(3): 185-213 (1995))
  • D. G. Kendall
  • M. Kimura (research interests:
    * Mathematical methods for modeling and analysis of complex systems
    * Statistical analysis of World Wide Web
    * Nonlinear time series analysis based on dynamical systems theory
    * Machine learning
    * Neural computation )
  • M. F. Neuts (The fundamental period of the queue with Markov-modulated arrivals,), (Matrix-Geometric Solutions in Stochastic Models: An Algorithmic Approach,)
  • K. R. Parthasarathy (Probability Measures on Metric Spaces, 1967), (Positive definite kernels, continuous tensor products and central limit theorems of probability theory, 1972), (An Introduction to Quantum Stochastic Calculus, 1992), (The Mathematics of Error Correcting Quantum Codes)
  • N. U. Prabhu
  • H. Solomon
  • R. Syski
  • L. Takacs
  • R. L. Tweedie
  • D. Vere-Jones (Some applications of probability generating functionals to the study of input-output streams. J. Roy. Statist. Soc. Ser. B 30, 321--333. R-2)
  • G. S. Watson
  • P. Whittle
  • J. Gani (Herausgeber)

Friday, January 27, 2006

Will Increased Longevity be Short-lived?

Nice title, I think. But I am not (yet) ready to try to seriously address it.

The title came to me yesterday when thinking about some demographic questions. Walking in town I saw a poster for a musical group called "Kill the Young!". I recently read Michel Houellebecq's book, "Possibility of an Island", a thoroughly depressing consideration of the ageing (is there really an e in ageing?), the quest for eternal youth, and the place, or lack thereof, for the increasing number of "greys".

Most proximately, a few days ago I had read an article entitled "Doom and Demography" by an optimistic compensationist called Nicholas Eberstadt (at the American Enterprise Institute); his contention that there is an "incessant stream of dire--and consistently wrong--predictions of global demographic overshoot" I find particularly odious. He seems to be of those who do not recognize that spending one's heritage is not the same as spending one's earnings; that we have lived better using petroleum, high-tech fishing, and depletion of aquifers than we would have lived otherwise is not a demonstration that the planet's potential to support us is as high, or higher, than it was thirty years ago. He then concludes that "greying and population decline" are the next crises we'll be reading a lot about (hah hah, more population crises...), as with a tone suggesting we should not be any more worried or concerned than we should have been by the population overshoot that didn't happen!? And yet, he wrote in Foreign Affaires, Summer 1991:

Population Change and National Security
Nicholas Eberstadt
Summer 1991
Summary: It is not so much a question of how population growth threatens world security, so much as of how fertility differentials between rich and poor countries will threaten the developed world, by producing resource scarcities and irresistible migration pressures.[emphasis mine] The fastest-growing Third World areas are those least likely to share Western values, and could produce "a fractious, contentious and inhumane international order" rather more dangerous than the Cold War. Rather than pursue pro-natalist policies to bring Western birth rates up to close the gap, the West should consider ways to improve the export of Western values.

I guess he has the humility to admit he was wrong? I guess a former-doom-sayer has the right to criticize the trade. Now, back to my questions...

  1. If there is a stable (maximum) number of people possible, then greater longevity implies less frequent replacement (lower turnover): what are the consequences for resistance to diseases (virii) which continue to evolve at much higher rates? And for mutation and natural selection in general?
  2. There are some important social design questions (including manpower and educational planning, pension financing, social safety-net and income redistribution schemes, sustainable resource allocation, etc.); how can I (make a living and) work on them?
  3. How can one deflate the rate-based arguments that lead to the conclusion that population growth is the only solution to provide (materially and services) for an aging population? How to demonstrate that it is not only not the only solution, but that is an absurd and unviable one?

I found a couple of papers from the XXVth International Congress on Population, held in Tours, France, last summer, which help illuminate question 3:
Population trends, employment and labour migration in the European Union

Serguey Ivanov, United Nations

The size of the EU population of working age will start to shrink in a few years. By 2050, there will be fewer men and women of working age in the EU-25 than there have been in EU-15 before enlargement. This is likely to happen even though the projections have factored in the influx of 23 million working-age migrants. Increases in labour utilization could greatly offset some of the adverse demographics. Increasing immigration or fertility are other options. Combining United Nations demographic projections with alternative assumptions with respect to labour utilization illustrates the effects of different scenarios and highlights intra-EU diversity in terms of dynamics of labour supply and adjustment options.

See paper.

Building a simplified model to assess the impact of population ageing, employment trends and immigration levels on pension sustainability in the EU-25 member states

Fernando Gil Alonso, Universitat Autònoma de Barcelona

Although the problem of pensions funding cannot be reduced solely to the demographic factor, the construction of highly sophisticated models to analyse the impact of ageing on pension sustainability introduces problems of data availability and comparability between EU Member States owing to the different types of pension systems existing in each country. To overcome these obstacles, the approach used in this paper is different. A model has been defined using some simplified assumptions, which do not provide precise forecasts but orders of magnitude on the implications of ageing for the 25 EU countries over the period 2000-2030. This is done by using a range of scenarios of population and employment trends in order to assess the impact of a series of alternative measures –like the increase of the effective average age of labour exit or the growth of labour supply through immigration– to assure the sustainability of pension systems. The results emphasise the importance of employment growth for Europe and in particular the need to keep older workers at work, thereby contributing to both increased participation and delaying retirement.

See paper.

The first paper considers just how to provide enough labor-aged people in Europe, which seems to be "at least as many as today"; rather than considering any manpower-planning scenarios based on evolved composition of economic activity in Europe, the approach is purely macro. As the author notes, any growth would be entirely driven by productivity gains. The solution technique is to study scenarios, varying assumptions about fertility, labor utilisations rates, and net immigration. While interesting (fertility could bounce back almost enough, without reflecting unrealistic levels), it begs the underlying question: how many people, with what skills, will we need to support both themselves and the young, retired, and unemployed?

The latter paper, apparently intended as a sort of back-of-the-envelope baseline, is a horribly good example of what a policy analysis should not be. It sets the decisions in terms of number of people working, ratios of retired to working, and so on. The analysis considers inter-country variations in its chosen parameters and ratios (number of people working, effective retirement age, e.g.), but neglects some economic and policy issues of great importance:

  • Will per capita GDP (per person, not per worker) converge across the 25? There remain disparities both in productivity and in prices (and salaries), as the Bolkestein and longshoremen issues have highlighted.
  • Could another model better address the employment opportunities (or lack thereof), via manpower planning, input-output, CGE, to establish the credibility of a scenario with 98 million immigrants, 113% participation of working-age population (160% in Ireland) and 40% reduction in per capita benefits?

How can a policy roadmap intended to steer the next 45 years ingore the economic disparities today and the pressure to converge them? What will the economy of the future comprise? What natural resources will we transform, with what energy? What will we offer to exchange for imported materials and products? Will our offer be sufficiently attractive? I have nothing against working to a riper age, but wonder where these jobs will come from; last I heard, we do not currently employ everyone who wants to work: how shall we go from 173 million to 289 million jobs?

As I warned I would, I have strayed quite far from the topic of the title in this note. I'm old enough to be in favor of longevity, and hope it will prove sustainable for all of humanity. But as long as we answer the wrong questions (just because they are simple enough to answer?), we have little chance of finding the right answers. If we fail to negociate this unprecedented shift in economic demography, the longevity we expect today may not remain available very long.

Wednesday, January 25, 2006

Tailoring My Conference Experience

There are probably good reasons for organizing conferences (congresses, conventions) with only a small portion of the experience allotted to plenary sessions, and much to sessions in "tracks". Among them come to mind offering the possibility to present (and get feedback) to many more participants, and in effect organizing mini-conferences for specialists in the tracks which might well not be viable as independent conferences (economies of scale, low marginal cost of time devoted to tracks).

The down side for attendees who are not specialists (or sufficiently knowledgeable to be truly interested in some very technical topics) is that they will spend a non-negligeable amount of time in session which they may barely understand, and about which they are certainly unlikely to provide good criticism, feedback, and suggestions. What is more, they have to choose among the tracks the one with which they have the most affinity.

Perhaps the greatest challenge of this sort that I've faced was ISMP 2003. The parallel sessions, which occupied about 60% of the in-session time, required ten selections (one per session) among about twenty-five options each! The number of tracks is indicative of how specialized they were, by the way: academic Balkanization at work.
Session natureNumberTotal HoursOptionsSequences
Plenary 6 6:45 includes opening, announcement of prize-winners1
Semi-plenary 4 4:00 3 81
Parallel 10 15:00 up to 26
Breaks (coffee, lunch)159:45

One solution, often used by organizers of trade shows with "workshops" or "tutorials" (at extra cost) is to schedule the sessions with restricted attendance before (or after) the main conference, rather than interleaved. However, conferences such as ISMP are designed for five-legged sheep. If too little of the conference is interesting, one can choose not to attend, right?

In an ideal world, the program (schedule) for multi-track conferences would not be decided by the organizers, but by the attendees. I would imagine a system allowing registered attendees to vote for the titles they would like to hear, then the schedule would be calculated to enable the conference to provide the greatest total interest, avoiding (as much as possible) scheduling at the same time two or more sessions appealing to the same attendees, and concomitantly avoiding "deserts" without something for everyone.

In the absence of such a system, I've begun considering my options for ROADEF 2006.
Session natureNumberTotal HoursOptionsComplexification
Plenary 4 4:30 includes annual meeting of the association x1
Semi-plenary 0 0:00
Parallel 10 15:15 (77%) 7 x 282,475,249
Breaks (coffee, lunch)99:00plus pôt d'accueil, gala dinner

I could print out the fifteen pages (actually, I ought to be able to edit it to ten pages, one per session, with the seven tracks on one page, but it annoys me to have to do the word processing myself) and then eliminate until I have my program. But what I prefer to do is add some JavaScript to the schedule page allowing me to click to hide sessions I want to "eliminate" and then print just the final filtered list.

The JavaScript to do this is explained on the Microsoft site, and there are also examples on "real live" sites, like (who use it to allow readers to hide boxes of figures or other complementary info). In a nutshell,
  1. Enclose the text to hide or show in a div or span with a unique id attribute
  2. Define a function (called "toggle(e)", for instance), which either swaps the visibility property between visible and hidden, or swaps the style display property between "none" and ""
  3. Add an identified way to invoke the function, as the "onclick" property of a visible element, or an anchor (link) with "href=javascript:toggle(joe)"

In practice, this took me a lot longer to achieve than I expected. Fortunately, I found the lemonde code to use to debug my code (it worked, mine looked similar but simpler, why didn't mine work?). Apparently I don't do JavaScript nearly often enough; it took me much too long to recognize my problem case-sensitivity-- the difference between "GetElementByID" (wrong) and "GetElementById" (right)!!

My function:

<script language="javascript">
function toggle(e){
cible = document.getElementById(id_cible);
if ("none"){"";
} else { = "none";

An example of a block it hides/shows (showing the title in any case):
<p class="headbox2">
<a class="pnav" href="javascript:toggle('LuS2P1')"> Salle 2 – Bioinformatique I – Rumen Andonov</a>
<a name="LuS2S1"></a></p>

The toggle call has been set up in the session title; now to wrap the content seems necessary to set at least one property of the object, by the way, or else the script cannot test the display property (object has no properties error).
<div id="LuS2P1" style="visibility: visible;">
<p>* Identification de nouvelles protéines et de variants de protéines connues par méthode évolutionnaire, Jean-Charles Boisson, Laetitia Jourdan, El-Ghazali Talbi et Christian Rolando
* Suffix Tree ANalyser (STAN): Un outil de recherche de motifs nucléiques et peptidiques dans les chromosomes, Anne-Sophie Valin, Grégory Ranchy, Yoann Mescam, Patrick Durand, Sébastien Tempel, Jacques Nicolas


Monday, January 23, 2006

Workshop on Bayesian Cognition

I missed another possibly interesting workshop in Paris, France, January 16 - 18, 2006.

Organized by the bayesian-cognition community,
Scope and goal:
Animals and artificial systems alike are faced with the problem of making inferences about their environments and choosing appropriate responses based on incomplete, uncertain and noisy data.
Probabilistic models and algorithms are flourishing in both life sciences an information sciences as ways of understanding the behavior of subjects and the neural processing underlying this behavior, and building robots and artificial agents that can function effectively in such circumstances.
This workshop will gather life and information scientists to discuss the latest advances in this subject, specifically addressing the following topics:
  • Probability theory as an alternative to logic
  • Probabilistic models of neurons and assembly of neurons
  • Probabilistic models of CNS functionality
  • Stochastic synchronisation of neuronal assemblies
  • Probabilistic interpretation of psychological and psychophysical data
  • Probabilistic inference and learning algorithms
  • Probabilistic robotics

It did say "limited number of places", though. Unfortunately, there is no mention of proceedings to be published.

Despite the stated purpose, the topics look to be more those of information scientists than life scientists. I hope they had a good time!

Sunday, January 22, 2006

OR for Better Management of Sustainable Development

EURO XXI in Iceland, July 2-5, 2006

21st European Conference on Operational Research

I'm pleased to note that the conference does appear to really align its content with its theme. Semi-Plenary Sessions planned include only two that are "just math", plus another which addresses the means of solving mathematical programs:

  • Unified Approaches and Techniques for Approximation Algorithms to NP-hard Problems, Dorit Hochbaum, University of California, Berkeley, USA
  • How Good are Interior Point Methods? Tamás Terlaky, McMaster University, Canada
  • Global Optimization in Practice: Software and Applications, János D. Pintér, PCS, Inc and Dalhousie University, Canada

One which is about markets:

  • Pricing and Revenue Management with Endogenous Demand
    Ioana Popescu, INSEAD, France

Two which address methodological questions in a more general setting, and one which addresses methodology in the realm of sustainable development:

  • Social Choice and Computer Science
    Fred S. Roberts, DIMACS, and Rutgers University, New Brunswick, USA
  • Problem Structuring and Formulation in Decision Aiding
    Valerie Belton, University of Strathclyde, United Kingdom
  • Problem Structuring and Sustainable Development
    Jonathan Rosenhead, London School of Economy and Political Science, UK

Finally, four which address real-world domains:

  • Operational Research and Fisheries Management
    Ragnar Arnason, University of Iceland, Reykjavik, Iceland
  • Dynamics and Control of Epidemic Processes
    Gustav Feichtinger, Vienna University of Technology, Austria
  • Energy, Environmental Management, and Sustainable Development
    Klaus S. Lackner, The Earth Institute, Columbia University, USA
  • Design and Optimization of Emission Trading Markets and Sustainable Bargaining Systems
    Stefan W. Pickl, Universitaet der Bundeswehr, Munich, Germany

It would even be possible (albeit a luxury and not very reasonable for a non-presenter) to then attend ISMP 2006 in Rio de Janeiro from July 30 to August 4.

Workshop On Growth and Employment in Europe

I don't plan to hop over to Oxford for the day, particularly since the presentations announced appear more "reporting" than "recommending", and at thirty minutes per presentation--including questions and discussion--"workshop" seems to misrepresent the nature of the event.

DATE: 27 January 2006
VENUE: Oxford Brookes University Business School, Wheatley Campus, Wheatley, Oxford OX33 1HX

Title: Paradise Lost? The Fate of the Scandinavian Economies in
the 21st Century

DENMARK: Jesper Jesperson (
Title: The Danish Case: Flexicurity and Demand Management

GERMANY: Eckhard Hein ( and Achim Truger (
Title: Germany's Stagnation in the European Context

FRANCE: Pascal Petit ( /
Title: Thirty Years of Massive Unemployment: Its Impact on the
Wage Labour Nexus

ITALY: Elisabetta De Antoni
Title: Italy and the EMU: A Controversial Issue

UK: Philip Arestis ( and Malcolm Sawyer (
Title: Economic Policy in the UK Under New Labour: The End of Boom and Bust?

GREECE/IRELAND: Eleni Paliginis (
Title: The Impact of EMU on Greece and Ireland

SPAIN: Jesus Ferreiro (, Carmen Gomez ( and Felipe Serrano (
Title: Can Miracles Come to an End? Limitations of the Current Model of Growth in Spain

PORTUGAL: Isabel Salavisa (
Title: The End of Catching up in Portugal: Causes and Consequences

Tags : employment | Europe | economics |


Thursday, June 29 – Saturday, July 1, 2006
at the Saint-Petersburg State University,
Saint-Petersburg, Russia
For detailed information: EAEPE sympo06.doc.
I've been wanting to visit Saint Petersburg for a few years, to benefit by the big sprucing up for the tercentennial in 2003, now that the celebration is finished and the crowds are presumeably back to normal levels. However, looking at the conference info, I'm don't know that I'd be welcome: "... to bring together a limited number of researchers (maximum 50)..." This may well be one of those events where intellectual tourism is unwelcome; where presenting a paper is required for attendance. Indeed, "Everyone is urged to volunteer to serve as a session chair." "The symposium is open to everyone working on labor markets across the social sciences."

The first segment is focused on comparative research of labor market transformations. Globalization, internationalization and localization of the labor force make it important to examine inter-related labor markets. This work should take into account how labor markets have, on one hand, class, ethnic, and gender dimensions and, on the other hand, that there is movement between labor markets (of different temporal durations) and these movements are in themselves influenced by class, race and gender. We would particularly welcome papers on intra-European mobility and inter-European/Asian labor mobility, including the role that Chinese and Indian workers are playing in European labor markets.

The second segment is devoted to the multiple interpretations of labor and its historical transformation(s). In particular we want to bring together different social scientists including economic historians, sociologists and anthropologists who work on the conceptualization of labor, its relationship to production and distribution, and as an essential aspect of human nature and human behavior. We believe that enriched by recent improvements in the various social sciences a fuller concept of labor is possible. This conceptualization will be both more ontologically oriented and be more sophisticated than the conceptualization used in the 20th century. In addition, this conceptualization is influenced by changes in institutions and technologies. An important part of this re-conceptualization of labor is to look at employer-employee relations, contractual arrangements and conflicts, alienation and exploitation, and therefore the challenges for labor politics and trade union movements in the 21st century.

Assuming the Scientific Committee of eight attends, that only leaves about forty-two places...what would Douglas Adams think of that?
Oleg Ananyin (Moscow, Russia)
Grбinne Collins (Trinity College Dublin, Ireland)
John Groenewegen (Delft, The Netherlands)
Maria Lissowska (Warsaw, Poland)
Pascal Petit (Paris, France)
Irina Peaucelle (Paris, France)
Mikhail Sinyutin (St-Petersburg, Russia)
Yuri Veselov (St-Petersburg, Russia)

From the Heterodox Economics Newsletter-21,