Friday, December 30, 2011


Model structure--the relationship between model elements.

Calibration--using data-sets to ensure that model matches. Differences between these two constitutes specification error.

Validation --Tests to measure  models predictive capacity. Requires a second data-set.

Model Development Principles
•    Concentrate on transparency. Models should be communicable. This necessitates the development of nested modeling.
•    Limit interconnectivity between model elements to the minimum possible.
•    Model structure should be transparent. Relationships between model elements shoul be clear, and should (when possible) correspond to participants pre-existing mental categories.
•    Equations available, data-sources documented. When possible, a model should be packaged to contain it's own documentation.

Tuesday, December 27, 2011

Modeling Dynamic Systems with System Dynamics

The theory and process of system dynamics modeling requires specialized knowledge and mind-set which is time-consuming to develop.

With the notable exception major metropolitan areas, and certain innovative MPO's, most regions no longer contain the problem solving capacity 'in-house'. Existing in-house capacity is primarily institutionalized in a 'system maintenance' role. The capacity to solve problems has been outsourced to consultants.  Consultants maintain the capacity to offer on-demand expertise and solutions based on comparable experience. Also maintain specialized of occaisonally needed technical know-how. Effecively, act as  intermediaries between decision makers and technical information and interpreters of that information.

System Dynamics does represent a 'new scientific revolution', in the sense that it is replacing the linear and atomistic representations of the world. However, the educational project should remain separate from system dynamic consulting. The structure of scientific revolutions dictates that the theory will be more widely adopted as more practical applications for the theory are found and demonstrated.

Friday, December 23, 2011


'Mega-technic' method of economic growth--increasing economic activity through increases in scope, rather than in efficiency. Doing so makes sense when increasing production is a capital intense endeavor, benefitting existing firms and institutions. As things change, vast capital stocks actually problematic. No longer located in an expanding market, but instead competing for limited resources. In such a context, a vast stock of capital can actually be a competitive weakeness. Developed over time, it naturally includes older, less productive technology. (Institutions also).

Wednesday, December 21, 2011

Dangers of Linear Regression

Linear regression is adequate for short-term forecasts, but dangerous over long time periods. Any given regression provides a snap-shot of current conditions. As conditions change over time, the predictive capacity of a regression declines.

Maintaining the predictive capacity requires repeating the process of data collection, cleaning, and regression.  This runs up agains the limitations of data: Statistical methods require statistically significant data samples to function. Long-term data series requires that data collections methods, geography, and metrics remain constant over time, with  'ideal' data-sets are collected at a single point in time. These limitations on the availability of data limit what can be regressed. 

More dangerously, there is an increasing reliance on automatically calculated statistical analytics  as measures of formal statistical validity, without the recognition that these measures are not 'absolute', but rather innovative methods developed to detect known errors in the application of other statistical methods. Successfully applying and interpreting these results requires a separate body of knowledge to identify and explain anomalies.

To improve model quality, there is a strong desire to reduce the number of variables present in a regression analysis. When faced with two highly correlated variables, only one may be included. This becomes extremely problematic if two highly correlated variables diverge over time,  it becomes an open question about which variable actually possessed predictive capacity. Or if either variable did, and the validity of a model actually resulted from a the two variables linkage to a third, unregressed variable. Regression models are only capable of showing correlation between different variables, rather than causal relationships. Without that explicit linkage, it becomes possible to draw conclusions that are statistically valid, but that have limited utility.


Urban growth is a reflection of economic growth. Economic activity takes place in space.
As the amount of economic activity taking place increases, so does the amount of space needed to 'house' that activity. The amount of space needed can be provided by making either making more intense use of existing urbanized area, or by making additional urban land available.

As a thought experiment, consider a 'bounded city', which can no longer expand in space because of of a 'hard boundary' (Political, legal, geographic, regulatory, etc.). This city is thus limited in its capacity to expand its urban area. But because economic activity takes place in space, contined expansion of economic activity requires additional space. Without the capacity to add additional urban space, existing urban space must be used more intensively. A combination of three factors makes this difficult.

Theoretically, urban land is developed in a manner characteristic with its highest and best use, so that in a context of limited land, urban land would convert to high value uses. In reality, development occurrs in accordance with the highest and best use at the time of development. Changes in the amenity of a parcel over time means that the highest and best use can also change. However, the capacity of the parcel to adapt to these changes is minimal. Once developed, a parcel's land use is fixed by the structure of the developed building. Many structures are so specialized as to be unsuitable for other uses, regardless of changes in the highest and best use. Urban land use thus remains fixed over long periods of time.

Secondly, urban land suffers from fragmentation. Large parcels are partially developed, or broken into smaller parcels. Over time, this process generates progressively smaller and less coherent parcels for development, which are progressively more difficult to develop. The impact of this dynamic is compounded over time. Because of differences in construction and maintenance, different structures depreciate at different rates and are available for redevelopment at different times. This makes it difficult to recombine smaller parcels into larger parcels.

Redevelopment occurrs when the income generated by new development is sufficient to cover the cost of clearing old structures, erecting new ones, and covering the resultant loss of income from the destruction of old structures. As urban land becomes scarcer, the value of urban land and the resultant rents that can be charged rises, making it more difficult to find a replacement use that will provide sufficient income to be worth redeveloping.

Given the constraints posed by re-use, urbanized area tends to expand in response to economic development. But urban expansion does not occurr in a random manner, but in  a pattern dictated by the function of urban land markets. Because the value generated by undeveloped parcels on the urban fringe (greenfields) is extremely minimal, they are developed in preference to redeveloping existing sites.

Employment centers occupy the most central locations—not out of a direct desire for centrality, but because of their primacy in the metropolitan development process. They come first, and the rest of the metropolis orients itself around them. Second most-centrally located are retail uses. While they follow residential in the development process, the competitive advantage represented by a more central location ensures willingness to 'outbid' residential users. Residential uses located at the least central locations, where land values are lowest.

Centrality should be understood in a network sense, rather than in a geographic sense. While their has historically been a correlation between the geographic center and centrality, it is not a causal linkage. Historically, the center of a city occupied the most 'central' location, because it was located en-route to the largest number of destinations. Development of limited access transportation networks such as subways and freeways changes this dynamics, so that proximity to transportation network access points becomes the best measure of centrality.

Land Value

Land values are lowest at the edge of the urban area because amenity value is at it's lowest.  A use located on the urban fringe has half the amenity exposure of a more central location, simply because land beyond the urban fringe remains undeveloped.

Post-Graduate Education

Predict the future demand for skills, requires the capacity to predict the future. Faced with limited ability to do so, ongoing education is critical. Life-long learning not as an 'enrichment' exercise (History of Jazz, Appreciating Wine), or an opportunity to explore the basics of other fields, but as an ongoing excercise is professional development. There are a record number of students returning for graduate degrees and doctoral degrees, in an effort to improve their earnings and gain competitive advantage. Traditional degre programs are unsuitable for this.

The modern university education includes both technical and liberal arts elements, designed to develope well-developed people with a range of capabilities that will be useful over the course of a lifespan, as well as technical and professional training that will be useful immediately upon graduation. That does not reflect the demands of modern society.

The University needs to grab a growing share of the 'adult education' market, which will require changes in its pedagogical method. Classes structured for working professionals with very different needs, capacities and requirements then traditional students. I advocate for the expansion of certificate-scope programs. I qualify it with the use 'scope', because the historical role of certificate development has been one of curricular development--first a certificate, then a minor, before becoming the focus of an entire degree, post-graduate or doctoral degree.

Tuesday, December 20, 2011

Urban Services

Enormous economic prosperity made available by urbanization, characterized by the development of services. (Services differ from goods in that services require proximity to deliver.) As a result, their exists a population density threshold for which it is economically viable to deliver services.


Large corporations possess the spare capacity to generate necessary infrastructure (physical or institutional) for the expansion of economic activity, thus adding elements to a system necessary for system growth or maintenance.