Ghanaian voters are twice as much likely to vote for parliamentary candidates who provide infrastructural development than those who promise financial support to individuals, a study by George K. Ofosu.  The Associate of the Ghana Center for Democratic Development (CDD-Ghana) found that in competitive constituencies, such a promise works whether or not a candidate and the voter belong to the same political party, whereas in non-competitive constituencies, this pledge only influences voters who belong to the candidate’s party. Equally important to voters is the commitment of candidates to regularly organize community meetings to listen to their concerns and debrief them on parliamentary debates. The research also shows that candidates who offer to attend or financially support social events (eg. funerals, religious events, and traditional festivals), or help deal with government bureaucracy or secure state employment are more likely to be voted for than those who do not. The candidates’ profession has no effect on their chances of being selected. Hailing and living in the constituency is more preferable and both female and male candidates are equally preferred. Conducted in November-December, 2018, the study provides a systematic analysis of which of the different types of constituency service influence voters’ choice and whether the effects differ by partisanship and electoral setting. Key findings Generally: - Citizens prefer candidates who will provide local public goods (infrastructure) almost twice as much as they like those who promise to give financial support to individuals. - Citizens were about 13 percentage points (pp) more likely to choose a candidate who pledged to spend the majority of the MPs Common Fund on public infrastructure compared to one who plans to use only a little (or none) of the fund for this purpose - In contrast, voters were only 7 percentage points more likely to pick a politician who committed to spending the majority of their funds to provide private benefits compared to one who plans to use only a little (or none) of the fund for this purpose (baseline) - Equally important to respondents is regularly organizing community meetings to listen to their concerns. Compared to the baseline (a candidate who will not organize community meetings), citizens are 13.5 percentage points more likely to vote for candidates who say they will organize regular community meetings every month, 13.4 percentage points more likely to vote for those who promise to organize meetings every three months, and 11.7 percentage points more likely to choose those who promise to hold meetings every six months - The probability that citizens select a candidate as MP, decreases to about 6 percentage points when the aspirant promises only yearly meetings - Candidates who promise to sometimes (half of the time) or always help constituents in dealing with the government bureaucracy or finding state employment are 7.2 and 10.2 percentage points more likely to be preferred compared to those who will hardly do so, on average, respectively. - Attending or financially supporting social events such as funerals, religious services, traditional festivals and naming ceremonies influences vote choice; compared to those who promise to hardly attend social events, those who offer to attend half of the time or always are 4.2 and 7.8 percentage points more likely to be selected, respectively. Other findings: Type of profession has no effect (although the educated may have a slight advantage) - Voters equally prefer female and male candidates - Candidate affiliated with the incumbent executive party (NPP) were 3.2 percentage points more preferred - Hailing and, importantly, living in the constituency is more preferable Do the effects vary by partisanship and the type of constituency? Here, the research focuses on the provision of public infrastructure and private financial transfers - In competitive electoral settings (irrespective of the level of segregation), the promise of public infrastructure is highly influential among a candidate’s co-partisans and non-copartisans - In non-competitive settings, the promise of public infrastructure is only effective among copartisans in segregated constituencies - In non-competitive, non-segregated constituencies, which constitute the majority of constituencies in the country, citizens are mostly influenced by regular (at least once every six months) meetings with constituents to listen to their concerns and brief them on debates in parliament. The effect ranges between 9 percentage points and 17 percentage points