Are we pulling out of this recession? Economists are beginning to make comments to the effect of “the numbers we see indicate that things are improving”, but they do not show these numbers. These economists (and their interviewers) are content with merely reporting their interpretation. They are not interested in displaying the underlying reasons for their interpretation. I have been looking for simple to display and easy to understand graphs that would illustrate why these economists are saying what they’re saying. Most attempts have resulted in graphs that show the positive direction but these haven’t been very dramatic; the encouraging news is hard to detect. Consequently they’re not reproduced here, but I am still searching.
Recently though it’s come to my attention that the OECD (discussed briefly in an earlier article) has developed a means of showcasing the business cycle and the New York Times recently ran an article on their technique. It is the best example I’ve come across that shows that we’ve recently “turned a corner” in this economy. The graph, after all does actually show a turning of a corner and, both the NY Times article and the OECD data contain historical information that leads one to conclude that, if history is any guide, things will soon be getting better.
The image below comes from the OECD website and shows the “turning the corner” effect. Simple as this chart may appear it bears some explanation. The NY Times article also provides an explanation, so if this explanation leaves you cold you are encouraged to visit their site. Furthermore the NY Times article is has a great animation showing the historical significance of this recent corner that we just turned.
The chart uses an interesting technique to plot the data. To explain it, it may be useful to first use the same technique using more familiar data. For this we’ll look at how the length of the day varies across the year. You should know that the first day of summer is also the longest day of the year, the first day of winter is the shortest day of the year, and the first day of spring and fall have 12 hours of day and night. If we plot this with length of day on the Y-axis and the day of the year on the X-Axis we get the graph shown below. (day length courtesy of http://www.timeanddate.com/worldclock/astronomy.html centered on Seattle WA)
The graph above shows everything the graph below shows, but merely in a different format. If you look carefully you’ll notice that as the first day of summer approaches the days are getting longer but they are getting longer by less and less time. From March 30 to March 31 the day gets longer by about three and a half minutes. From June 19 to June 20 the day gets longer by about 4 seconds. But if we’re looking for signs of change it is useful to focus not only on how long the day is but on how fast it’s changing. The graph below puts the day length on the Y-Axis and the change from one day to the next on the X-Axis. In this graph Spring is when the days are longer than average (12 hours) and the days are getting longer. Summer is when the days are longer than average and the days are getting shorter.
This is essentially the same technique the OECD used to plot the business cycle. You may know that many economic indicators are standardized on a scale of 100, with 100 being normal. A consumer confidence number of 100 would mean that consumers were neither overly optimistic nor overly pessimistic about the economy. In this graph 100 is analogous to 12 hours of daylight in the graph above. You may also know that many of these economic indicators are updated and published monthly. Most news stories report on the value of the economic indicator and whether this month is higher or lower than last month. “Consumer confidence hit an all time low last month at blah, falling by blah percent from the previous month.” The OECD and NY Times charts are essentially reporting the same information.
If you’re just interested in knowing the value of the indicator you could plot that on a line, but most people also want to know if things are getting worse or getting better. “Even though the consumer confidence decreased last month it decreased by the smallest amount in 6 months.” Or, “Even though the consumer confidence is very low it is up slightly from last month.” As we’ve done on the day length graph, the X-Axis highlights the change from the last measurement. Thus if the index is improving the graph will have a positive X (or horizontal) component, and if things are getting better but are slowing down at getting better the X component will be positive but close to line. If the main indicator is above normal and is getting better then the graph is in the upper right, the “expansion” section. If the graph is above normal but getting worse, the graph is in the upper left, the “downturn” section.
Note, the OECD’s website allows one to interactively choose between 1 and 4 different economic indicators to plot using their handy graphing tool. (It also allows users to choose 1 or 2 countries to plot from about 40 different countries.) The chart above uses just 2 indicators but is easier to understand the techinque if you examine just one indicator at a time.
Even though we’re still in the worst section of the graph (lower left), things are below average and getting worse, the rate of getting worse has slowed. A “real” recovery is when things not only have stopped getting worse, but are actually getting better. This happens when the chart just enters the lower right section (from the lower left). The NYTimes article points out that this transition typically happens 6 months after the chart exhibits that turn-around behavior in the slowdown section. We are on the road to recovery but are not there yet.
The description of this graph both on the OECD website and in the NY Times article indicates that it is just a clever way to plot the information from one data set (eg monthly consumer confidence). However if this is the case then the “hair pin turn” the orange Business Confidence Index shows in the OECD graph shouldn’t be possible. I am not sure what explains this change, and I’ll issue an update if I figure this out. However I doubt this will affect the final conclusion.