Authoritative economic growth forecasts are often optimistically biased. Negatively skewed variation--negative shocks being larger than positive shocks--could contribute to bias by making long-run average growth smaller than typical-year (median) growth. This positively biases forecasts based on typical years. We compare medians and means in real per-capita GDP growth across countries, regions, and time windows from 1820-2016. Over decadal periods, we find mean growth rates <1%/y smaller than median growth rates in most countries and regions (median 0.23%/y across countries). Surprisingly, we find both large- and medium-magnitude shocks contribute to these differences, rather than only large ‘black swan’ events. We find negative skewness correlated with high levels and slow growth of per-capita GDP and population, and high per-capita GDP growth volatility, building on previous studies. We find negative skewness alone insufficient to explain recent growth over-projections by the International Monetary Fund (IMF) and the U.S. Congressional Budget Office (CBO).