National Issues - connecting the factors that affect us (Addended)

Sun	Mon	Tue	Wed	Thu	Fri	Sat
1	2	3	4	5	6	7
8	9	10	11	12	13	14
15	16	17	18	19	20	21
22	23	24	25	26	27	28
29	30

George Rebane

When thinking about the upcoming election in the context of national issues and all that is going on which affects America, things get complicated pretty quickly. An established tool for representing, sorting out, and communicating such factors and their interaction is called the Influence Factors Diagram (IFD). An IFD is usually involved in the first steps of laying out the design of a system or process. It is conceptually extremely simple, even corporate management and (a few) politicians can deal with it. Lawyers hate them for their uncompromising clarity. For the intelligent reader – the ones I fantasize who frequent these pages – reading an IFD is a slam dunk. (I first introduced IFDs here.)

Below is an IFD that relates a number of factors comprising national issues that are critical to understanding the campaign arguments in the upcoming election. Each factor is named and represents some measure of its level or intensity. The arrows represent the influences that the factors have on each other. Each arrow shows the direction of influence from influencer to the influenced. The +/- signs indicate the type of influence. A plus (+) sign says that the influencer contributes a tendency for the influenced factor to follow the direction of the influencing factor – increase-to-increase and decrease-to-decrease – when everything else is kept the same. For example, ‘US National Debt’ influences ‘US Debt Service Costs’, and the plus indicates that when the debt goes up so do the country’s debt service costs, and vice-versa.

Conversely, a minus (-) sign indicates that the influenced factor tends to react in the opposite direction from that of its influencer, all things being equal. Look at ‘Regulatory Burdens’ influence on ‘America’s capacity to generate wealth’; when regulatory burdens increase, wealth generating capacity decreases, and vice versa.

The actual reaction of a factor depends on the sum total of the strengths of the influences it receives from all the factors that input to it. In some cases a negative influence is swamped by some positive influences, and, of course, the other way around can also occur. The main thing here is that the whole ensemble and their influences are laid out so that they can be analyzed and discussed in the context of all the significant interactions. This helps to remove the blinders that frequently cause a debate to become focused on a single one-dimensional influence of one factor upon its neighbor, missing the forest for the two trees. In a correctly drawn IFD, it may turn out that the two factors are not even neighbors, but have one or more mediating factors between them which are affected by even more factors. An IFD makes this clear.

Finally, an important aspect to discover from an IFD is what factors form loops, and what kind of loops are these – reinforcing or cancelling. This can also be thought of in terms of feedback loops, both positive feedback or negative feedback. Sometimes these are also called ‘virtuous’ and ‘vicious’ loops, but that depends on whether you consider the feedback itself to support or inhibit the factor you are focused on.

I have illustrated a positive feedback loop with green arrows. You can start with any factor in that loop and assume that its level is increasing/decreasing. Then trace the influences around the loop using the +/- sign conventions, and you should return to your starting factor being influenced to increase/decrease as you originally assumed. If the loop influences dominate, then there is a tendency for one or more involved factors to experience a ‘run away’ effect where their levels keep increasing/decreasing until something ‘breaks’ or intervenes to ‘brake’ the runaway influence.

In the obverse, I have also illustrated a negative feedback loop with red arrows. If you start with any factor whose influencing level is increased/decreased and trace it around the loop, you’ll come back to your original factor with an influence that opposes your assumed starting influence. Negative feedback loops have a self-correcting tendency on their member factors.

Discovering such loops (and even naming them) allows the analyst to focus on any such loop for, say, making policy or creating a simulation model, and in the process recognizing its effect which may need to be reinforced or minimized by doing something to another factor in the loop over which you have more control.

A fast way to determine if a loop is positive or negative feedback is to count the number of minus (-) signs in it. An even number of minus signs – 0, 2, 4, … - indicates that it is a positive (runaway) feedback loop. An odd number of minus signs – 1, 3, 5, … - indicates that it is a negative (self correcting) feedback loop.

I hope this little tutorial will make the displayed IFD useful in our discussion of national issues during the election campaign. I will be referring to it from time to time as needed to make or counter an argument - you are invited to do the same. I invite the reader to examine the IFD and ask any questions about it, or give me feedback on any disagreements that you may have with its design.

[Addendum] Upon perusal of the above IFD, another positive feedback loop pops up that is truly 'virtuous'. Consider 'America's capacity to generate wealth' going up. That will increase 'US GDP growth rate', which in turn will increase 'Spendable Government Revenues', even under the existing tax burden. Growing government revenues tends not only to reduce the existing 'Tax Burden', most certainly not to increase taxables and taxes rates. Finally, lessening the 'Tax Burden' has the salutary influence of increasing 'America's capacity to generate wealth', therefore confirming the virtuous nature of that feedback loop in our economy.

This is one more illustration of how the IFD has gained its place in the analysis, understanding, and communication of how complex combinations (i.e. systems) of interrelated factors behave.

[Technical Note] The techies will recognize that the IFD is really the first level design of a simulation model with the arrows indicating the partial derivatives of the influenced factor (variable) with respect to the influencing factor (variable), and the sign being interpreted in the usual way. The arrows may alternatively be interpreted as probabilistic influences in a Markov process, Bayes net, or even a causal net. They can then be viewed as ‘conditional probability tables’ which are really tensor constructs.

For those wanting a more thorough grounding in IFD, I recommend John Sterman's Business Dynamics - Systems Thinking and Modeling for a Complex World. Sterman is with MIT's Sloan School of Management, and has also developed software to support the building of computable IFDs for serious analysis and problem solving. However, the $164 price of the book and its accompanying software CD indicates that it remains the definitive resource for learning and using this powerful tool for modeling very complex systems, both in technology development, planning, and public policy.

Comments

Have a musical IFD, complements of Stephen Malinowski's software:

http://www.youtube.com/watch?v=gBaAh9oWIZ0

Posted by: TomKenworth | 13 August 2012 at 11:15 AM

And by the way, other than taking roll, I also had early versions of this diagramming software donated, and thus gave my comp lit kids (who often became my programming kids in 10th grade) a head start in such concepts.

http://www.inspiration.com/

Posted by: TomKenworth | 13 August 2012 at 11:21 AM

I thought by now someone else would ask? What software did you use to create this graphic?

Posted by: TomKenworth | 14 August 2012 at 12:41 AM