Manifestos lay out parties’ foundational plans for the upcoming legislative term. However, many Germans – among them people with special needs, reading disabilities or migration backgrounds – struggle to understand the complex language in those manifestos. To involve these people equally in the political process, parties’ provide easy read manifestos [ERMs] written in easy language. The limited existing research approaches ERMs from a linguistic angle. This study opens up a more political lane of research, arguing that accurate translation of parties’ issue emphasis and positions from regular manifestos is crucial to achieve the goal of equal information and participation. However, parties could – strategically or inadvertently due to the translation process – emphasize topics of higher salience in the ERMs’ target audience like welfare or equality while de-emphasizing less salient topics like the economy. Building on German data, we use supervised machine learning and topic modeling to compare the issue emphasis and positions parties’ take in their regular and easy read manifestos. Our contribution is twofold: First, we can help to improve the ERM translation practices and foster more equal participation for all citizens. Second, we are breaking new ground by exploring the suitability of NLP methods on this unique form of language.