|Baseline and Rate of Change (BaR) Analysis Grid©|
The BaR Analysis Grid© goes well beyond simple trend analysis and adds several important dimensions to economic analysis. Most usefully, the BaR takes a jumble of economic indicators and arranges then into an understandable pattern. The BaR:
|The vertical axis shows whether the indicators are above or below the baseline. The horizontal axis specifies each indicators three-month rate of change, positive or negative.
|The economy is strongest when the majority of the indicators are well above the baseline, and more so when most are in the upper-right quadrant. Historically, when most indicators move to the left side of the grid, and drop towards or below the baseline, the economy falls into a recession. See a display of the business cycle here.|
Rate of Change
A comparison of methodologies has shown that using a three-month trend for the rate of change is the most accurate for portraying economic conditions. The rate of change is calculated as follows.
1. The data for most economic indicators are tracked monthly, using a three-month exponential moving average. The three-month EMA eliminates some statistical noise that is common with most economic measures. Corporate profits, which are reported quarterly, are not adjusted.
2. Using the moving averages, the three month trend for each indicator is calculated.
3. The trend is adjusted so that all indicators are proportional when they are plotted to the BaR Analsysi Grid©. Proportional adjustments are explained below.
Baselines are intended to show where an indicator is likely to be when a recession begins. Each indicator's baseline is determined by it unique characteristics and historical trends prior to recessions. Like the rate of change, the baseline percentages are also proportionally adjusted when plotted, as explained below. Charts of the economic indicators and their baselines can be seen here.
Based on each indicators unique range of change, all plotted coordinates are set to a base of 100, so that the percent changes for all indicators are of equal weight. This is necessary because the range of percent change for some measures is larger than others, yet their influence on the economy may not be any greater than indicators that have a smaller range of percent change. For example, since 2010, the highest annualized three-month trend for building permits was 90%; whereas, it was only 9% for retail sales. However, the smaller percent change for retail sales is as meaningful to that measure, and to the overall economy, as the larger percent change is for building permits. To equalize both measures, their unique ranges of change are used. Thus a 30% change in building permits, when divided by the 90% high, would equal 33%. In comparison, a 3% change in retail sales, when divided by the 9% high, would also equal 33%. By dividing each indicator's percent change by its unique range, the indicators share a common base of 100 (as 33% = 33 / 100 = portion divided by base). This allows each economic indicator to be proportionally plotted to the BaR Analysis Grid©. (This example simplifies how the adjusted proportional rate of change and baseline percentages are actually calculated, but it demonstrates the essence of the methodology. In the actual algorithm, when an indicator's range is calculated, outliers beyond the three standard deviations from the mean are removed.)
An argument can be made that some indicators, like corporate profits, should be weighed more heavily as their influence on the economy is greater. However, the BaR Analysis Grid© is designed to act in a similar fashion as the overall economy. In the real economy, when corporate profits decline significantly, over time, this pulls down other indicators. Hence, the influence of declining corporate profits works through the economy with some delay. This same pattern would occur on the BaR, making the grid an accurate representation of the economy.