Is history is more or less bunk?

The Economist magazine reports on a paper given at the recent Economic History Society's annual conference in the U.K. This work suggests history matters and matters for a long time. The paper looks at the effect of lynchings before 1930 in the U.S. on income distribution today.

The first, by Cornelius Christian of Oxford University, looks at the consequences of the lynching of black Americans between 1882 and 1930. Mr Christian found that this history of racial violence still echoes down the decades. He also found that the higher an area’s lynching rate before 1930, the wider the income gap between blacks and whites remained in 2008-12, even when adjusted for factors such as the education and employment levels of a local area. A high rate of lynching widens this gap by as much as 15% in some cases.

While an interesting empirical result, the question this raises is What is the mechanism that brings this effect about? Just how can something like lynchings 80-120 years ago be affecting income distribution today? It is not obvious what the link is. We need a theory to explain the data.