<Ignore Depletion At Your Peril The Oil Drum Algeria, Land of Opportunity?>

30 comments on Consumer Confidence and Gas Prices

Comments can no longer be added to this story.

| Show without comments | PDF version

Oil CEO on November 30, 2005 - 6:46am Permalink | Subthread | Comments top
Inflation is very simple. Laziness cannot be used for an excuse.

Inflation is also incredibly important when dealing with price issues - although, admittedly, I'm not sure how important it is in this particular case.

For future reference - this is how you do it. I'm assuming you are using excel. Build yourself an inflation column, to be used whenever you need it. I'll send you mine, it just depends on whether you trust it or not.

In the meantime this is how you build one.

Two sites. One uses CPI, the other GDP deflator, or whatever. If they are the same, whoa did I make a mistake.

http://www.westegg.com/inflation/
http://www1.jsc.nasa.gov/bu2/inflateCPI.html

I just checked them, they are both CPI, but they are different(yet similar) - there must have been a reason I chose to do it this way at the time:) Doesn't matter, add a GDP one in for your own recipe.

Make an excel table. List your years down the left. Get the percentage change year to year. add these values.

The trick is to build the damn thing. I averaged two sources. Whether this makes mine less accurate or increase the chances of it being accurate in the first place is anybody's guess.

The farther we progress into the future from the past, the less important accuracy concerning the past is. As long as we are in the ballpark.

My favorite example of this is Michael Economides' chart in "The Color of Oil" showing the inflation adjusted price of a barrel in the 1880's as about $200. The book was written in 2000. In 2002 he took part in a debate with Deffeyes and a couple of other academics in Houston, I believe. I have read a summary of the debate. In it he claims that same price in the 1880's as $1000. Now that is a discrepancy that deserves an explanation.

So anyway, once you have your list of yearly percent increases you just cut and paste into any spreadsheet with nominal values.

My number for 1983 in 2004 dollars, for instance is, 87.85%. If you come up with 95% or 70% it doesn't really matter. Ballpark is key. We are just trying to get the graph to make all years appear closer to reality, i.e 2006, then they would be in nominal values. But you already know this.

Like I said, I'll send you my list. Laziness is not an option.

Comments can no longer be added to this story.
Stuart Staniford on November 30, 2005 - 6:57am Permalink | Subthread | Parent | Comments top
Yeah, but I have monthly data for CSI and - so using yearly inflation time series is going to be a so-so bad hack (it's not that I have any conceptual difficulty with it). The BLS (who actually produce the CPI) has monthly CPI indices for recently, but not over this range of time (at least not that I could find). I looked around for half an hour, couldn't find a long monthly series, and decided it didn't matter enough to stay up even later doing it - correlations if they exist are going to be at reasonably short range anyway, and there the inflation doesn't matter. There is no way inflation is going to make this r^2 any better. If you want to be less lazy than me and send me an Excel of a long monthly CPI deflator, by my guest. I'll replot the picture :-)
Comments can no longer be added to this story.
Oil CEO on November 30, 2005 - 7:07am Permalink | Subthread | Parent | Comments top
alright, I think I'm getting what you are saying. you really want monthly versus yearly. I missed that at first.

Is there anyway you can convert the Consumer confidence to yearly, or a 12-month moving average?

What about comparing percentage change rather than overlapping price and CCI numbers?

Maybe I'm the lazy one. But laziness cannot be an excuse, so I've got to find another excuse. Hahaha.

I really like the idea you've got here on this correlation, I'm just trying to think of an accurate way to measure it.

Comments can no longer be added to this story.
Oil CEO on November 30, 2005 - 7:13am Permalink | Subthread | Parent | Comments top
Sorry, yeah I know you did a percentage change month-to-month. I mean more of a long-term, 6-month moving average type comparison. I'm big on that. Short-term is very problematic in my view. Most people couldn't tell you what they paid for gas yesterday. And if you remember, you are confusing yourself with most people.
Comments can no longer be added to this story.
Stuart Staniford on November 30, 2005 - 7:16am Permalink | Subthread | Parent | Comments top
The particular correlation I measured is between the percentage CSI change from one month to the next, and the percentage gas price change from one month to the next. Those price swings are generally much larger than inflation from one month to the next any time recently (which is going to be much less than 1%), and therefore there is negligible error in my correlation measurement. The only thing that is not quite right is the visual impression of the gas line in the first graph. On the other hand, if I put in a CPI deflator, all the folks who believe the CPI is an evil conspiracy to hide the true level of inflation could pop up and argue it was way off anyway (the pre-Clinton CPI differs from the current one by several percentage points a year, apparently, which is going to compound to pretty big factors of difference over a few decades).
Comments can no longer be added to this story.
Oil CEO on November 30, 2005 - 7:48am Permalink | Subthread | Parent | Comments top
Inflation shouldn't compound. It only appears in the month or year it occured.

The Pre-Clinton CPI numbers are bigger because inflation was much bigger(Pre-Greenspan).

If inflation was 12%(I'm guessing) in 1978 versus %1.5-%3.0 in the 1990's, that 12 percent is not going to compound.

For the sake of argument,let's say inflation was 12% in 1978 and 1% every year after that. Well 1978 dollars would be converted to 1979 dollars(in 1979) by mutiplying them by 1.12. End of story.They are now 1979 dollars.

In 1980 we would be converting 1979 dollars to 1980 dollars by multiplying them by 1.01. The 12% is history.
If we revise that 12% to 15% later we only have to go back to the one calculation in 1979 to make the change. There is no compounding.

Comments can no longer be added to this story.
Oil CEO on November 30, 2005 - 8:31am Permalink | Subthread | Parent | Comments top
I ammend myself. You are correct,corrected inflation does compound - only not in the way that true compounding implies. It is heavily diluted over time.

For instance, if we start with a dollar in 1970, and call inflation 5% in 1970, and 2.5% every year afterwards, we get $1.31 in 1980.

If we go back and correct inflation by a huge 5% in 1970 to 10%(highly unlikely, worst case scenario), we get $1.37 in 1980. 6 cent difference. But 5 cents of that occurs in first year, the other cent takes 9 years to accumulate due to this compounding.

Assuming corrections happen in both directions, cancelling each other out over time, would it be reasonable to assume that that "ballpark" is again the key to dealing with inflation?

Comments can no longer be added to this story.
Oil CEO on February 10, 2006 - 11:17pm Permalink | Subthread | Parent | Comments top
Test Post

Comments can no longer be added to this story.
Oil CEO on February 10, 2006 - 11:42pm Permalink | Subthread | Parent | Comments top
Test Post

Comments can no longer be added to this story.
Oil CEO on February 11, 2006 - 12:25am Permalink | Subthread | Parent | Comments top
TEST POST

Comments can no longer be added to this story.
Oil CEO on February 11, 2006 - 1:01am Permalink | Subthread | Parent | Comments top
oilprod2005
Comments can no longer be added to this story.
Fernando on November 30, 2005 - 4:15pm Permalink | Subthread | Parent | Comments top
Stuart,

Thank you. I like very much the level of detail, effort and intelligence that you always put in your articles.

As I've said somewhere else before, what would happen if we charted oil prices in Euros instead of US dollars?

I know that oil prices are high in Euros too, but, have you considered that the US dollar is not as strong as a few years ago?

For instance, other commodities (grains, gold, silver, etc) are expensive, too.

As a local analyst said to me: "The only cheap thing is the US Dollar".

Fernando

Comments can no longer be added to this story.
Sunlight on December 1, 2005 - 12:45pm Permalink | Subthread | Parent | Comments top
Stuart, I think you are right to take the side of the null hypothesis (gasoline prices not explaining changes in consumer confidence). The key, however is to understand what consumer confidence is about? Is it just a feeling, or do consumers act on it? What is the correlation between consumer confidence and consumer spending, which I think is your real target?

Alan Greenspan, who is a deep student of economic indicators (no matter what your view is of his policy setting record) has little regard for consumer confidence, because he believes that consumers simply spend all they have, at least until you get to the upper income brackets. Therefore personal income is a better predictor of spending behavior. So far, the way we see consumers going mad for the Black Friday Christmas deals almost regardless of the economic outlook, would seem to prove him right. Most Americans seem to spend like crazy as long as they have a job.

Just as a side note: I happened to have spent last weekend in Detroit. The Detroit Free Press carried lots of coverage about Black Friday consumer madness, noted approvingly (at least it seemed to me) when it discussed families who had figured out how to scramble for the latest and greatest gadgets on their list. Just after GM announced the closure of 17 auto plants, which is not exactly good for Detroiters. Is this a sign of addiction, or what?

Comments can no longer be added to this story.

30 comments on Consumer Confidence and Gas Prices

Comments can no longer be added to this story.

| Show without comments | PDF version