A bimodal distribution is a distribution with two clear peaks, which usually means the data are being generated by two different underlying groups rather than one single average pattern.
Why economists care
Economists often summarize data with a mean, median, and variance. That can hide important structure. If wages, firm productivity, or test scores are bimodal, the issue is often not “more spread” but “two populations in one sample.”
Common examples include:
- wages split between low-skill and high-skill workers,
- firm productivity split between informal and formal producers,
- inflation expectations split between well-anchored and poorly anchored households.
How to read it
Two humps in a histogram suggest bimodality, but economists still have to ask whether the pattern is real:
- Is the sample large enough?
- Are the peaks stable across years or subsamples?
- Is the apparent bimodality just noise from bin choice?
If the pattern is real, the next step is usually to model the separate groups directly instead of forcing one average response on everyone.
Economic interpretation
Bimodality often points to segmentation, threshold effects, or regime differences. In labor economics, that may mean two skill tiers. In development economics, it may reflect a poverty trap where some households stay near a low-income equilibrium while others move to a higher one.
That is why bimodality matters for policy. A single policy aimed at the average observation may fit neither group well.