What Is the Base Effect?

The base effect is the effect that choosing a different reference point for a comparison between two data points can have on the result of the comparison. This often involves the use of some kind of ratio or index value between two points in a time-series data set, but can also apply to cross-sectional or other types of data.

Thinking about the base effect in comparing different numbers or pieces of data means considering the question, "Compared to what?" The choice of the basis for comparison can have a large effect on the apparent result of a comparison. If ignored or misunderstood, the base effect can lead to a major distortion and possibly mistaken conclusions. However, if considered carefully, it can be leveraged to improve an analyst's understanding of the data and the underlying processes that generate them.

Key Takeaways

  • The base effect refers to the effect that the choice of a basis of comparison or reference can have on the result of the comparison between data points.
  • Using a different reference or base for comparison can lead to a large variation in ratio or percentage comparisons between data points.
  • The base effect can lead to distortion in comparisons and deceptive results, or, if well understood and accounted for, can be used to improve our understanding of data and the underlying processes that generate them.

Understanding the Base Effect

The base effect occurs whenever two data points are compared as a ratio where the current data point or point of interest is divided or expressed as a percentage of another data point, the base or point of comparison. Because the base number makes up the denominator in the comparison, comparisons using different base values can yield widely varying results. If the base has an abnormally high or low value it can greatly distort the ratio, resulting in a potentially deceptive comparison.

The base effect is most commonly pointed out when discussing comparisons using time-series data where the raw data value at one point in time is being compared to another chosen point. It can occur whether there is a constant index base to which many values in the series are being compared, or when doing a moving period-to-period comparison.

The base effect can work for or against you. Choosing an inappropriate basis for comparison or ignoring the base effect in a time index can lead to a distorted perception of the magnitude or rate of change of the current point in a data series. This is related to the idea of garbage-in-garbage-out; if the value of the denominator in a comparison is uncharacteristic or unrepresentative of the overall data trend then the comparison will likewise be unrepresentative of the relationship between the current data point and the data series as a whole, and whatever process generate those data.

For example, the base effect can lead to an apparent under- or overstatement of figures such as inflation rates or economic growth rates if the point chosen for comparison has an unusually high or low value relative to the current period or the overall data.

On the other hand, understanding the base effect and choosing appropriate bases for the comparison you want to make (or at least accounting for the base effect in your comparison) can lead to a better understanding of the data or even the underlying process. For example, comparing monthly data points to their previous values 12 months prior can help filter out seasonal effects. Alternatively, comparing a data point to a long-run moving average of its own values can help reveal if the current datum shows an anomalously high or low value.

Example of the Base Effect

Inflation is often expressed as a month-over-month figure or a year-over-year figure. Typically, economists and consumers want to know how much higher or lower prices are today than they were one year ago. But a month in which inflation spikes may produce the opposite effect a year later, essentially creating the impression that inflation has slowed.

The distortion in a monthly inflation figure that results from abnormally high or low levels of inflation in the year-ago month is an example of the base effect. A base effect can make it difficult to accurately assess inflation levels over time. It diminishes over time if inflation levels are relatively constant, without strong outlier values.

Inflation is calculated based on price levels that are summarized in an index. The index may spike in June, for example, perhaps due to a surge in gasoline prices. Over the following 11 months, the month-over-month changes may return to normal, but when June arrives again the following year its price level will be compared to those of a year earlier when the index reflected a one-time spike in gasoline prices.

In that case, because the index for that month was high, the price change this June will be less, implying that inflation has become subdued when, in fact, the small change in the index is just a reflection of the base effect—the result of the higher price index value a year earlier.