The DOM index

There’s no better evidence that the stock market has no connection to the lives of Americans than the fact that the Dow has been tanking of late.  As of the close on 9/9/11 it’s down about 14% over  2 months.  But in that time it’s gone down, then up, and down, and down some more, and in all likelihood has a ways to go before bottoming out.

Has the actual real-world, day-to-day well-being of any group of people except a small number of traders gone down by 14% in the same time span?  Probably not.  Is the market responding to the long-term stagnation of middle class incomes, or to unemployment, or to food prices, or to life expectancy?  Not that I know of.  It’s been responding to abstractions, largely related to the debt of European nations and banks.

In a previous post I started discussing the idea of the DOM(estic) index: an attempt at a less-bad daily measure of the health of the United States.  It’s a daunting task, one with no clear right or wrong answers.  We can consider many different realms—sub-indexes of the main index that cover areas such as energy, health, employment, food, etc.  These sub-indexes would then be combined to produce a single index.  Because each topic is large in and of itself, I’d like to start today with energy, and then consider the others in subsequent posts.

Not to give away the plan early, but our goal is to set up a database and a ticker on this site that can be used by anyone on any other site to display the day’s DOM index (and maybe its sub-indexes).  But first we need to develop the concept a bit more fully.  Adam and I have considered 30-40 metrics (everything from the price of oil to weekly unemployment claims to pesticide use to median air quality) that might be important to include in the index.  Combining them meaningfully will be difficult, but we need to start somewhere.

First we should consider the goals that I spelled out before:

  1. To sever the false sense of connection between the stock market and general well-being.
  2. To provide a tight feedback loop between the metric(s) used and the levers used to change the system.

I’d argue that goal 1 is on its way to being achieved simply due to the financial crisis and its aftermath.  To address goal 2, we will aim, when we have a choice, to select frequently-updated data sources over infrequent (if less noisy) alternatives.

I should state an assumption we make: we assume we’re at the end of economic growth as it is currently measured.  (There is substantial evidence for this, as both the data and the reasoning indicate.)  That means that while we hope to back-calculate the DOM for dates in the past (really as far back as we can get data), I can imagine that it won’t be meaningful in those different eras.  Times like the late 1990s seemed great in many ways: low unemployment, strong economic growth, little military conflict, etc.  Yet in many ways it was setting us up for hard times to come: carbon emissions were rising at record rates, millions were without health care, the last plentiful oil fields (e.g. the North Sea) were near peak, etc.  In designing the index, we need to balance metrics that measure present well-being with those that are indicative of future well being.  That is, we need to stop discounting the future.  This adds a complicating factor: some (many?) short-term negatives may be long-term positives.  For example, decreasing oil consumption in a recession is due to a short-term economic negative (in conventional thinking), not due to any desire to save energy; however, this enables a more graceful descent (if the goal isn’t to ramp up growth immediately after the recession).

So to consider energy, we can look at a few different types of metrics: total consumption of various forms of energy, total available energy (either as reserves for fossil fuel sources or capacity for alternative sources), energy return on investment (EROI), side effects of that energy production/consumption, future estimates for energy availability, and meta-metrics such as the rates of change or per-capita values for these.  One difficult that arises immediately, that we found is the case for many of the metrics we’d like to use, is that the data is reported sporadically and with high variability, making it difficult to easily reflect changes in a daily metric.  The straightforward approach that we plan on using initially is to compute moving averages of the metrics over some period of time, as appropriate for the data source.

Here are several metrics we have considered including in the energy sub-index (with links to data sources for the ones for which we have good sources):

In this, I view high prices as a negative (because it impacts people who, like it or not, still use energy) but view decreased consumption that results from those high prices as a positive (because it is more sustainable in the long term).  In this, the goal is to let the data determine what’s positive and negative rather than trying to reason about it directly.

That rounds up the energy sub-index metrics.  I’d be interested in hearing about any others that would be valuable to include (and how you envision including them).  I’ll continue with several other of the sub-indexes in a subsequent post.

Leave a Reply

(required)

There aren't any comments at the moment, be the first to start the discussion!