Let's assume that somebody knows what to do when it comes to the
and European economic and banking situation.
And let's assume that they'll come back from their extended vacation (or
coma, as the case may be) at some point in the next few months and put things
in order. Further, let's assume that all
the destructive decisions that were cemented into the foundation of our current
crisis in the fourth quarter of 2001 can be jack-hammered into tiny bits and
used as aggregate for paving the path out of the mess we're in. Assuming all of this (aided by some mythical
unicorn tears, pixie dust, and the whisker of a saber tooth tiger - yes, it's
one of those potions) and you don't need to know what heteroscedasticity is and you can just
return to your nostalgic Father's Day festivities. However, assuming that any of the
aforementioned fail to materialize, read on.
As with most of the statistics we use to make sense out of our world, a plethora of assumptions - both implicit and explicit - undergird our consensus 'knowledge'. I have yet to meet any individual expert or group of scientists in any field who have the academic or social integrity to actually state the assumptions that they have opted to deploy untested and have the further decency to critique their own conclusions based on the failure for those assumptions to hold. I marveled at the digital tautology infused in the papers presented at a conference from which I recently returned. The scientific method, if invoked, meant that whatever statement or conclusion was to follow must enjoy some hegemonic priority in the minds of the listeners. Let there be statistical significance in a regressed set of variables and, voilà, we've got truth. InvertedAlchemy readers are acutely aware of my critique of our untested assumptions in general and their tragic behavioral and policy consequences. However, it's worth noting that the current economic interventions being proposed in the G-20 are actually doing grave damage to our ability to even sustain our statistical myth. And that, is the subject of this week's observations.
Now some of you are more familiar - because you were unwilling to sit through one of my statistics lectures at the University - with the concept of dispersion than you are with the term heteroscedasticity (and yes, I'll forgive that indiscretion). And, for you, allow me to explain the following. Let's assume that you make a series of observations that seem to work most of the time. For example, let's say that we survey the world's middle class and find that, with additional money, we find that people report a better quality of life. Using regression, we conclude that, with more money, quality of life increases. However, when we extrapolate that observation to the whole population, we find that the relationship not only doesn't hold but we find that some people at very high levels of monetary wealth are miserable and some people at very low levels of monetary wealth are quite self-actualized. Rather than rejecting our correlation 'truth', we explain the information that challenges the correctness of our hypothesis as 'outliers' or 'unexplained' rather than holding the possibility that we had the wrong hypothesis for which we applied the wrong metrics to confirm an imposed outcome. And we do this because, in the name of convenience, we need to understand the world efficiently. I don't want to ask everyone from every culture to respond to my attempt to understand the world. I want to ask a few people a few questions and draw sweeping conclusions therefrom.
And here's the problem. When the variables I think I'm assessing or measuring have dispersions creating heteroscedasticity (frequently a function of metrics at the measurement extremes of data), to gain confidence in my observations and the resultant conclusions, I limit the data that I gather to insure: a) self-fulfilling hypothesis retention; and, b) consensus among my scientifically-minded colleagues who, like me, want reductionism over complexity. Ironically, the most damaging effect of heteroscedastic variables is not in their essence per se but rather in the 'error' or 'unexplained variance' that they represent to the generalizability of the model and its conclusions.
Now, take a deep breath, grab a nutritious snack (possibly some nerve conducting friendly egg yolks for your brain) and strap in for the reason why I've used James Carville's 1992 presidential campaign slogan for this post. The world - sorry to all of you intelligent designers out there - is heteroscedastic. And that's the case for the part we think we understand. That's bad news for all you adherents to the method advanced in the 10th century by Ibn al-Haytham and modernized in the 16th century by the likes of Galileo and Kepler. Most of what we confidently know that we know, we don't know. But here's what's worse. What's happening in
capitals across the G-20 is that we are adding scale to variables we do not
understand in the first place. We
already don't know how monetary supply behaves in manipulated interventions so
we put MORE money in. We don't know how
real property deflation (a risk of currency inflation) will impact our
long-standing social obligations so we create currency supply bubbles of
gargantuan proportions. We know that 'sovereign
debt' has the full faith and confidence of governments at a time when no one
has faith and confidence in governments so we develop schemes to issue more
sovereign debt! In other words, we are
increasing the dispersion in a variable set that we've already evidenced a
complete absence of mastery around and we somehow wait to see it show up in a
model that was wrong in its creation. At
the apex of this irony is the fact that the most wealthy - yes the uber-1% -
are clamoring for returns to invest their ill-gotten spoils and they're stuck
with, you guessed it, currency and sovereign debt. It must be a bummer to steal all the jewels
only to find out that the jewels are just paving stones!
So here comes the punchline. What we need is to shrink. We actually need to have the courage to strive for a more elegant less. Now the cool reality is that we'll get there one way or another. We can either take the elixir of living within our suitable means or we can be served the ipecac from less charitable hands. Starting with a deflation of our egoic, dominion-infused certainty of control, we need to accept that, in our finest moments, we describe, not explain (and certainly never predict). Rather than seeking to control, we are more suited to steward that with which we are entrusted. This does not suggest an aversion to accretive impulses. It does, however suggest that we need to increase the heterogeneity of the variables we measure while decreasing the expectation of successful imposition of conclusions on disregarded populations. Realizing that unexplained variance is more likely a reflection of the sum of our projected social monotony plus observational sloth rather than an unfathomable mystery, we need to confront the reality that more input into a broken model actually speeds the propagation of the problem rather than introducing any remedy. Heteroscedasticity is not our enemy but our ill-conceived piling into its maelstrom will be our undoing.
In Integral Accounting parlance, we can seek phase and state coherence where all of the utility we desire from a system is achieved leaving the system with as much retained optionality as possible. Simply put, we must engage in a process of removing ourselves from the end of consumption and see ourselves in participation with a cycling and recycling of matter and energy. In a bizarre paradox, we may find ourselves DOING MORE with LESS.