This paper investigates the impact of Field Reservation Policy that paid farmers to not farm in 1970s Finland. The study uses a rich dataset, including over 290 000 newly digitized Agricultural Census documents from 1969, to analyze the long-term effects of the policy on farmers and their children. The findings reveal that eligible farmers in low-productivity regions were more likely to cease farming, although their income remained unchanged a decade later. However, their children experienced a 2.5% increase in annual earnings, were more likely to live in larger towns, and received slightly more education. The paper further explores the mechanisms behind these outcomes, showing that occupational choices account for nearly 60% of the earnings effect, while location and education play a smaller role. Children of lower cognitive ability drive the positive income effect. Despite the policy's economic benefits, it faced significant political backlash, contributing to the rise of the populist Finnish Rural Party in the 1970 parliamentary elections. This study contributes to the literature on the effects of moving to opportunity, the urban-rural income gap, and the political implications of structural transformation.

Analyzing archival data is of increased interest with digitization being a common bottleneck. This project presents a new deep learning method to automate data digitization focusing on retrieving data from regular forms with high accuracy rates. The novel method is to use two layered models: first to learn the cells in a form or table and then to transcribe the contents of each cell.